diff --git a/CMakeLists.txt b/CMakeLists.txt index fd68df87e..12addd473 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -454,16 +454,24 @@ check_cxx_source_compiles(" int main() {}" MP_VARIADIC_TEMPLATES) # Option BUILD_EXAMPLES -option(BUILD_EXAMPLES "Build nl-example" ON) +option(BUILD_EXAMPLES "Build examples: nl reader, nl writer" ON) +# Build NLReader example if (MP_VARIADIC_TEMPLATES AND BUILD_EXAMPLES) message(STATUS - "Building nl-example. Set -DBUILD_EXAMPLES=off to switch off.") - add_executable(nl-example src/nl-example.cc) - add_to_folder(${MP_FOLDER_PREFIX}util nl-example) - target_link_libraries(nl-example mp) + "Building nl-reader-example. Set -DBUILD_EXAMPLES=off to switch off.") + add_executable(nl-reader-example examples/nl-reader-example.cc) + add_to_folder(${MP_FOLDER_PREFIX}util nl-reader-example) + target_link_libraries(nl-reader-example mp) endif () +# Build NLWriter2 example +if (BUILD_EXAMPLES) + message(STATUS + "Building NLWriter2 example. Set -DBUILD_EXAMPLES=off to switch off.") + add_subdirectory(nl-writer2) +endif() + option(BUILD_DOC "Build documentation" ON) if(BUILD_DOC) add_subdirectory(doc) diff --git a/doc/source/components.rst b/doc/source/components.rst index c5f0e11f8..af72ba57b 100644 --- a/doc/source/components.rst +++ b/doc/source/components.rst @@ -12,13 +12,12 @@ on :ref:`driver-setups`. .. _NL-SOL-files: -NL and SOL files ----------------- +Reading NL and writing SOL files +----------------------------------- Any AMPL solver driver currenlty needs to input an NL file and report results in a SOL file. -AMPL might add in-memory model communication for efficiency, -which would be provided via the +The details are taken care of in the :ref:`recommended driver setup `. However, a :ref:`minimal driver setup ` might address the corresponding APIs directly, as described below. @@ -131,6 +130,25 @@ A standalone .sol file writer could be implemented by parameterizing the templates by minimal implementations of the `mp::BasicSolver` and `mp::ProblemBuilder` interfaces. + +.. _write-nl-read-sol: + +Writing NL and reading SOL files +------------------------------------ + +For modeling systems and applications using AMPL solvers, +MP provides a library to write NL and read SOL files. +The library is zero-overhead: it does not store the model, +nor does it require any intermediate objects to represent +model information. + +The API is provided by classes `mp::NLSOL`, `mp::NLWriter2`, +`mp::NLHandler2`, `mp::SOLReader2`, `mp::SOLHandler2`. +See +`example `_ +solving a small non-linear model. + + .. _recommended-driver-logic: Recommended driver logic diff --git a/doc/source/index.rst b/doc/source/index.rst index 88ecde1f9..0846657b5 100644 --- a/doc/source/index.rst +++ b/doc/source/index.rst @@ -7,6 +7,9 @@ new AMPL solver drivers. It provides type-safe and flexible interfaces suitable for linear and mixed-integer, non-linear, and Constraint Programming solvers. +Moreover, MP library provides +a :ref:`lightweight NL writer `. + MP replaces the previous `AMPL Solver Library`__ for solvers not requiring automatic differentiation. It is implemented in C++, parts of the API are exported in C. @@ -46,7 +49,8 @@ The documentation is divided in two sections: * :ref:`library-intro` shows instructions for downloading, installing and building the library * :ref:`howto` illustrates how to get familiar with the framework and how to quickly obtain the boilerplate code necessary to develop a new solver driver - * :ref:`components` shows the main components of the framework + * :ref:`components` shows the main components of the framework, + including NL reader and writer API * :ref:`howto-test` illustrates how to execute tests during development * :ref:`cppreference` contains the reference to all the classes in the framework diff --git a/src/nl-example.cc b/examples/nl-reader-example.cc similarity index 100% rename from src/nl-example.cc rename to examples/nl-reader-example.cc diff --git a/nl-writer2/CMakeLists.txt b/nl-writer2/CMakeLists.txt new file mode 100644 index 000000000..c13b21cf9 --- /dev/null +++ b/nl-writer2/CMakeLists.txt @@ -0,0 +1,32 @@ +cmake_minimum_required(VERSION 3.10) + +project(NLWriter2) + + +set(CMAKE_CXX_STANDARD 11) +set(CMAKE_CXX_STANDARD_REQUIRED True) + + +set(NLW2_DIR ${CMAKE_CURRENT_SOURCE_DIR}) +set(NLW2_LIB_FILES + ${NLW2_DIR}/src/nl-writer2.cc + ${NLW2_DIR}/src/nl-utils2.cc + ${NLW2_DIR}/src/dtoa.c) + +# Create library +set(NLW2_LIB_NAME "nlw2") +add_library(${NLW2_LIB_NAME} STATIC ${NLW2_LIB_FILES}) +target_include_directories( + ${NLW2_LIB_NAME} PUBLIC ${NLW2_DIR}/include) + +if (MSVC) ## Set the same as for your parent project +## target_compile_options( +## ${NLW2_LIB_NAME} PRIVATE ${AMPL_MSVC_COMPILE_OPTIONS}) +endif() + +# NLW2 example +add_executable(nl-writer-example + ${NLW2_DIR}/examples/nlsol_ex.cc) +target_include_directories(nl-writer-example PUBLIC + ${NLW2_DIR}/include) +target_link_libraries(nl-writer-example ${NLW2_LIB_NAME}) diff --git a/nl-writer2/README.md b/nl-writer2/README.md new file mode 100644 index 000000000..5db7dad59 --- /dev/null +++ b/nl-writer2/README.md @@ -0,0 +1,3 @@ +A lightweight AMPL NL file writer and SOL file reader. + +See mp.ampl.com. diff --git a/nl-writer2/examples/nlsol_ex.cc b/nl-writer2/examples/nlsol_ex.cc new file mode 100644 index 000000000..57320ef51 --- /dev/null +++ b/nl-writer2/examples/nlsol_ex.cc @@ -0,0 +1,55 @@ +/** + * Example using NLWriter2 and SOLReader2 + * to write NL file, execute solver, and read SOL file. + * + */ + +#include + +#include "nlsol_ex_mdl.h" +#include "nlsol_ex_nl.h" +#include "nlsol_ex_sol.h" + +#include "mp/nlsol.h" +#include "mp/nl-writer2.h" +#include "mp/nl-writer2.hpp" +#include "mp/sol-reader2.h" +#include "mp/sol-reader2.hpp" + +/// Invoke: +/// ipopt ["outlev=1 timelim=500" [text [filestub]]] +int main(int argc, const char* const* argv) { + if (argc<2) { + printf("%s\n", + "AMPL NL writer example.\n" + "Usage:\n" + " [\"\" [binary/text []]],\n\n" + "where is ipopt, gurobi, minos, ...;\n" + "binary/text is the NL format (default: binary.)\n" + "Examples:\n" + " ipopt \"\" text /tmp/stub\n" + " gurobi \"nonconvex=2 funcpieces=-2 funcpieceratio=1e-4\""); + exit(0); + } + std::string solver = (argc>1) ? argv[1] : "minos"; + std::string sopts = (argc>2) ? argv[2] : ""; + bool binary = (argc<=3) || std::strcmp("text", argv[3]); + std::string stub = (argc>4) ? argv[4] : "stub"; + + ExampleModel emdl; + ExampleNLFeeder2 nlf(emdl, binary); + ExampleSOLHandler2 esolh(emdl); + mp::NLUtils utils; + + mp::NLSOL + nlsol(nlf, esolh, utils); + nlsol.SetSolver(solver); + nlsol.SetSolverOptions(sopts); + if (!nlsol.Solve(stub)) { + printf("%s\n", nlsol.GetErrorMessage()); + exit(EXIT_FAILURE); + } + esolh.PrintSolution(stub); + + return 0; +} diff --git a/nl-writer2/examples/nlsol_ex_mdl.h b/nl-writer2/examples/nlsol_ex_mdl.h new file mode 100644 index 000000000..e1ee33390 --- /dev/null +++ b/nl-writer2/examples/nlsol_ex_mdl.h @@ -0,0 +1,421 @@ +#ifndef NLSOL_EX_MDL_H +#define NLSOL_EX_MDL_H + +#include +#include +#include +#include +#include + +#include "mp/nl-header.h" + +/** + * @brief The ExampleModel struct. + * + * A non-linear model with defined variables. + * Illustrates variable/constraint order, + * handling of nonlinearities, + * order of defined variables and their splitting + * into linear and non-linear parts. + * + +## eXample for NLWriter2/SOLReader2. + +## Variable bounds. +## Defined variabes. +## Pure linear constraint. +## Quadratic constraint. +## General nonlinear constraint. +## Initial guess (useful for local solvers.) +## Suffix. +## +## To write NL file and name files in AMPL, use commands +## ampl: option nl_comments 1; +## ampl: option auxfiles rc; +## ampl: write gmodel; +## +## Full documentation on NL format: +## technical report "Writing .nl Files" +## (https://ampl.github.io/nlwrite.pdf.) + +var x >=1.4, <= 30; +var y >=0; + +var t1 = x + 2*y + y^2; +var t2 = 3*x - 5*y; ## linear dvar +var t3 = 4*y + abs(y-2) + 6.38; + +minimize Obj_t2t3: 5*x + 3*t2 + 8*t3 + 15; + +s.t. C0_lin: 300 >= 5*x - 17*y >= -4; ## range constraint +s.t. C1_t2: x + 12*y - 18*t2 + 5*t2^2 <= 119; ## quadratic constraint +s.t. C2_t1t2t3: -38.2*y^2 +109*sin(t2) * (t1 + t3^1.5) >= -30; + +let x := 1.5; +let y := 0.11; + +suffix zork; +let C1_t2.zork := 5; + * + */ +struct ExampleModel { + /// Return NL header. + mp::NLHeader Header() const { + mp::NLHeader hdr; + + hdr.num_vars = n_var; + hdr.num_algebraic_cons = n_con; + hdr.num_objs = n_obj; + hdr.num_ranges = 1; + hdr.num_eqns = 0; + hdr.num_logical_cons = 0; + + /** Total number of nonlinear constraints. */ + hdr.num_nl_cons = 2; + hdr.num_nl_objs = 1; + hdr.num_compl_conds = 0; + hdr.num_nl_compl_conds = 0; + hdr.num_compl_dbl_ineqs = 0; + hdr.num_compl_vars_with_nz_lb = 0; + + /** Number of nonlinear network constraints. */ + hdr.num_nl_net_cons = 0; + hdr.num_linear_net_cons = 0; + + /** + Number of nonlinear variables in constraints including nonlinear + variables in both constraints and objectives. + */ + hdr.num_nl_vars_in_cons = 2; + + /** + Number of nonlinear variables in objectives including nonlinear + variables in both constraints and objectives. + */ + hdr.num_nl_vars_in_objs = 1; + + /** Number of nonlinear variables in both constraints and objectives. */ + hdr.num_nl_vars_in_both = 1; + + // Miscellaneous + // ------------- + + /** Number of linear network variables (arcs). */ + hdr.num_linear_net_vars = 0; + + /** Number of functions. */ + hdr.num_funcs = 0; + + // Information about discrete variables + // ------------------------------------ + + /** Number of linear binary variables. */ + hdr.num_linear_binary_vars = 0; + + /** Number of linear non-binary integer variables. */ + hdr.num_linear_integer_vars = 0; + + /** + Number of integer nonlinear variables in both constraints and objectives. + */ + hdr.num_nl_integer_vars_in_both = 0; + + /** Number of integer nonlinear variables just in constraints. */ + hdr.num_nl_integer_vars_in_cons = 0; + + /** Number of integer nonlinear variables just in objectives. */ + hdr.num_nl_integer_vars_in_objs = 0; + + // Information about nonzeros + // -------------------------- + + /** Number of nonzeros in constraints' Jacobian. */ + hdr.num_con_nonzeros = n_con_nz; + + /** Number of nonzeros in all objective gradients. */ + hdr.num_obj_nonzeros = n_obj_nz; + + // Information about names + // ----------------------- + + /** Length of longest con/obj name if names are available. */ + hdr.max_con_name_len = 0; // no need to set + + /** Length of longest variable name if names are available. */ + hdr.max_var_name_len = 0; // no need to set + + // Information about common expressions + // ------------------------------------ + + /** + Number of common expressions that appear both in constraints + and objectives. + */ + hdr.num_common_exprs_in_both = 1; + + /** + Number of common expressions that appear in multiple constraints + and don't appear in objectives. + */ + hdr.num_common_exprs_in_cons = 1; + + /** + Number of common expressions that appear in multiple objectives + and don't appear in constraints. + */ + hdr.num_common_exprs_in_objs = 0; + + /** + Number of common expressions that only appear in a single constraint + and don't appear in objectives. + */ + hdr.num_common_exprs_in_single_cons = 2; + + /** + Number of common expressions that only appear in a single objective + and don't appear in constraints. + */ + hdr.num_common_exprs_in_single_objs = 0; + + hdr.prob_name = "example_model"; + + return hdr; + } + + static constexpr int n_var = 2; + static constexpr int n_con = 3; + static constexpr int n_obj = 1; + + ///////////////////////////////////////////////// + //////////// Bounds and linear parts //////////// + ///////////////////////////////////////////////// + + /// Variables. + /// Put y first because it participates non-linearly + /// both in constraints and the objective. + /// This might be assumed by some NL solvers. + double var_lb[n_var] = { 0, 1.4}; + double var_ub[n_var] = {INFINITY, 30}; + const char* var_name[n_var] = {"y", "x"}; + + /// Defined variables. + /// It proves essential (IPOPT, MINOS) to split + /// those defined variables, whose linear parts + /// can be substituted into the constraints. + /// In this example, this is done for t3: the non-linear part + /// is extracted into a new defined variable nl(t3). + /// Note that x, y are swapped! + static constexpr int n_dvar = 4; + double dvar_linpart[n_dvar][n_var] = { + {0, 0}, {-5, 3}, {4, 0}, {2, 1} + }; + const char* dvar_name[n_dvar] + = {"nl(t3)", "t2", "t3", "t1"}; + + /// For some non-linear solvers (MINOS) it is necessary + /// to provide defined variables, which are used + /// in a single obj/con, just before that. + /// This info is returned here: + /// @param i: i<0 for obj -i-1, + /// i>0 for con i-1, + /// i=0 for all other defvars. + /// @return list of defined variables for \a i. + std::vector DefVarsInItem(int i) const { + switch (i) { + case -1: return {}; + case 0: return {0, 1}; + case 1: + case 3: return {}; + case 2: return {2, 3}; + default: + assert(0); + return {}; + } + } + + /// Algebraic constraint bounds. + /// Constraints are reordered: nonlinear first. + /// If we had logical constraints, they'd go last + /// (represented by expressions only.) + double con_lb[n_con] = {-INFINITY, -30, -4}; + double con_ub[n_con] = { 119, INFINITY, 300}; + const char* con_name[n_con] + = {"C1_t2", "C2_t1t2t3", "C0_lin"}; + + /// Typedef SparseVec<> + template + using SparseVec = std::vector< std::pair >; + + /// Algebraic constraints: linear parts. + /// + /// Linear parts (columns of the Jacobian) + /// are presented sparse, but should include entries + /// which can become nonzero in the Jacobian. + /// Thus, for C2_t1t2t3 we transmit the 0 elements. + /// + /// In the linear part of C1_t2 we substitute t2. + /// We could have put it into the nonlinear expression. + SparseVec con_linpart[n_con] = { + { {0, 102}, {1, -53} }, + { {0, 0}, {1, 0} }, + { {0, -17}, {1, 5} } + }; + /// Sizes of all Jacobian columns except the last + int col_sizes[n_var-1] = {3}; + int n_con_nz = 6; + + /// Obj sense: min + int obj_sense = 0; + /// Objective: linear part. + /// Similar to the Jacobian, + /// need all elements which can become nonzero. + /// + /// We substitute t2 and the linear part of t3 here. + SparseVec obj_linpart + = { {0, 17}, {1, 14} }; + const char* obj_name = "Obj_t2t3"; + int n_obj_nz = 2; + + ///////////////////////////////////////////////// + ////// Non-linear parts: expression trees /////// + ///////////////////////////////////////////////// + + // Here we write expression trees "on the fly". + // In the case when we'd write expressions from + // an intermediately stored tree, + // we might use ExprWriter::EPut(expr_node) + // to process subtrees recursively. + + /// Write dvar expressions. + template + void WriteDVarExpr(int i, EWriter& ew) const { + switch (i) { + case 0: // nl(t3) + { + auto ew01 = ew.OPut1(15, "abs"); + { + auto ew0101 = ew01.OPut2(0, "+"); + ew0101.NPut(-2); + ew0101.VPut(0, "y"); + } + } break; + case 1: // t2 + ew.NPut(0); // no non-linear part + break; + case 2: // t3 + { + auto ew01 = ew.OPut2(0, "+"); + ew01.VPut(2, "nl(t3)"); + ew01.NPut(6.38); + } break; + case 3: // t1 + { + auto ew01 = ew.OPut2(5, "^"); // binary opcode + ew01.VPut(0, "y"); // 1st arg + ew01.NPut(2); // 2nd arg + } break; + default: + assert(false); + {} + } + } + + /// Write constraint non-linear expressions. + template + void WriteConExpr(int i, EWriter& ew) const { + switch (i) { + case 0: // C1_t2 + { + auto ew01 = ew.OPut2(2, "*"); + ew01.NPut(5); + { + auto ew0101 = ew01.OPut2(5, "^"); + ew0101.VPut(3, "t2"); + ew0101.NPut(2); + } + } break; + case 1: // C2_t1t2t3 + { + auto ew01 = ew.OPut2(0, "+"); + { + auto ew0101 = ew01.OPut2(2, "*"); + ew0101.NPut(-38.2); + { + auto ew010101 = ew0101.OPut2(5, "^"); + ew010101.VPut(0, "y"); + ew010101.NPut(2); + } + } + { + auto ew0102 = ew01.OPut2(2, "*"); + { + auto ew010201 = ew0102.OPut2(2, "*"); + ew010201.NPut(109); + { + auto ew01020101 = ew010201.OPut1(41, "sin"); + ew01020101.VPut(3, "t2"); + } + } + { + auto ew010202 = ew0102.OPut2(0, "+"); + ew010202.VPut(5, "t1"); + { + auto ew01020201 = ew010202.OPut2(5, "^"); + ew01020201.VPut(4, "t3"); + ew01020201.NPut(1.5); + } + } + } + } break; + case 2: // C0_lin + ew.NPut(0); // no non-linear part + break; + default: + assert(false); + {} + } + } + + template + void WriteObjExpr(EWriter& ew) const { + auto ew01 = ew.OPut2(0, "+"); + { + auto ew0101 = ew01.OPut2(2, "*"); + ew0101.NPut(8); + ew0101.VPut(2, "nl(t3)"); + } + ew01.NPut(66.04); + } + + /// Primal initial guess. + double ini_x[n_var] = {0.11, 1.5}; + + /// Suffixes. + struct Suffix { + const char* name_; + int kind_; + std::vector values_; // store double's always + }; + static constexpr int n_suf = 1; + Suffix suf[n_suf] = { { + "zork", + 1, // constraints, integer + {5, 0, 0} + } }; + + + ///////////////////////////////////////////////// + /////////////////// Solution //////////////////// + ///////////////////////////////////////////////// + + std::vector sol_y_; + std::vector sol_x_; + int objno_ = -2; + int solve_result_ = -1; + std::map< + std::pair< std::string, int >, + std::vector< double > + > suf_out_; +}; + +#endif // NLSOL_EX_MDL_H diff --git a/nl-writer2/examples/nlsol_ex_nl.h b/nl-writer2/examples/nlsol_ex_nl.h new file mode 100644 index 000000000..6e8e8df51 --- /dev/null +++ b/nl-writer2/examples/nlsol_ex_nl.h @@ -0,0 +1,198 @@ +/** + * NLFeeder2 implementation for the nlsol_ex example. + * + */ + +#ifndef NLSOL_EX_NL_H +#define NLSOL_EX_NL_H + +#include "mp/nl-feeder2.h" +#include "nlsol_ex_mdl.h" + +class ExampleNLFeeder2 + : public mp::NLFeeder2 { +public: + /// Construct + ExampleNLFeeder2(const ExampleModel& mdl, bool binary) + : mdl_(mdl), binary_(binary) { } + + /** Provide NLHeader. + * + * This method is called first. + * + * NLHeader summarizes the model and provides some + * technical parameters, + * such as text/binary NL format. */ + mp::NLHeader Header() { + auto hdr = Model().Header(); + + hdr.format = binary_ + ? mp::NLHeader::BINARY : mp::NLHeader::TEXT; + + return hdr; + } + + /// NL comments? + bool WantNLComments() const { return true; } + + const char* ObjDescription(int i) { return "Obj_t2t3"; } + + int ObjType(int ) { return Model().obj_sense; } + + template + void FeedObjGradient(int i, ObjGradWriter& gw) + { WriteSparseVec(gw, Model().obj_linpart); } + + template + void FeedObjExpression(int i, ObjExprWriter& ew) + { Model().WriteObjExpr(ew); } + + + template + void FeedDefinedVariables(int i, DefVarWriterFactory& dvw) { + for (int dvar0: Model().DefVarsInItem(i)) { + int lin_nnz = CountNNZ(Model().dvar_linpart[dvar0]); + auto dv = dvw.StartDefVar( + dvar0 + Model().n_var, // index after direct vars + lin_nnz, Model().dvar_name[dvar0]); + /////////// Write the linear part: + auto linw = dv.GetLinExprWriter(); + WriteDense2Sparse_Pure(linw, + Model().dvar_linpart[dvar0]); + /////////// Write the expression tree: + auto ew = dv.GetExprWriter(); + Model().WriteDVarExpr(dvar0, ew); + } + } + + template + void FeedVarBounds(VarBoundsWriter& vbw) { + for (int i = 0; i < Model().n_var; i++) + vbw.WriteLbUb(Model().var_lb[i], Model().var_ub[i]); + } + + template + void FeedConBounds(ConBoundsWriter& cbw) { + for (int j=0; j + void FeedLinearConExpr(int i, ConLinearExprWriter& clw) + { WriteSparseVec(clw, Model().con_linpart[i]); } + + template + void FeedConExpression(int i, ConExprWriter& ew) + { Model().WriteConExpr(i, ew); } + + template + void FeedColumnSizes(ColSizeWriter& csw) { + if (WantColumnSizes()) + for (int i=0; i < Model().n_var-1; ++i) + csw.Write(Model().col_sizes[i]); + } + + template + void FeedInitialGuesses(IGWriter& igw) + { WriteDense2Sparse(igw, Model().ini_x); } + + template + void FeedSuffixes(SuffixWriterFactory& swf) { + for (int i=0; i + void FeedRowAndObjNames(RowObjNameWriter& wrt) { + if (wrt) { // && output_desired + for (int i=0; i + void FeedColNames(ColNameWriter& wrt) { + if (wrt) { // && output_desired + for (int i=0; i + void WriteSparseVec(GradWriterFactory& gw, const SparseVec& vec) { + auto gvw = gw.MakeVectorWriter(vec.size()); + for (auto v: vec) + gvw.Write(v.first, v.second); + } + + /// Write dense vector as sparse, + /// using provided writer factory. + template + void WriteDense2Sparse(GradWriterFactory& gw, const DenseVec& vec) { + auto gvw = gw.MakeVectorWriter(CountNNZ(vec)); + WriteDense2Sparse_Pure(gvw, vec); + } + + /// Count nnz + template + int CountNNZ(const DenseVec& vec) { + int n=0; + for (auto it=std::begin(vec); + std::end(vec) != it; ++it) + n += (*it != 0); + return n; + } + + /// Write dense vector as sparse, + /// given pure vector writer. + template + void WriteDense2Sparse_Pure(GradWriter& gw, const DenseVec& vec) { + for (int i=0; i + void WriteDense2Sparse_Pure(GradWriter& gw, const double (&vec) [n]) { + int sz = sizeof(vec) / sizeof(vec[0]); + for (int i = 0; i < sz; ++i) + if (vec[i]) + gw.Write(i, vec[i]); + } + + +private: + const ExampleModel& mdl_; + bool binary_ {true}; +}; + +#endif // NLSOL_EX_NL_H diff --git a/nl-writer2/examples/nlsol_ex_sol.h b/nl-writer2/examples/nlsol_ex_sol.h new file mode 100644 index 000000000..d4be7c491 --- /dev/null +++ b/nl-writer2/examples/nlsol_ex_sol.h @@ -0,0 +1,200 @@ +/** + * SOLHanlder2 implementation for the nlsol_ex example. + * + */ + +#ifndef NLSOL_EX_SOL_H +#define NLSOL_EX_SOL_H + +#include +#include + +#include "mp/sol-handler2.h" +#include "nlsol_ex_mdl.h" + +class ExampleSOLHandler2 + : public mp::SOLHandler2 { +public: + /// Construct + ExampleSOLHandler2(ExampleModel& mdl) + : mdl_(mdl) { } + + /** The NLHeader used to write the NL file. */ + mp::NLHeader Header() const { return Model().Header(); } + + /** Receive solve message. + * The message always ends with '\n'. + * + * @param nbs: number of backspaces + * in the original solve message. + * So many characters should be skipped + * from the message if printed straightaway. + * AMPL solver drivers can supply the message + * with initial backspaces to indicate + * that so many characters should be skipped + * when printing. For example, if the driver prints + * MINOS 5.51: + * and exits, and the message starts with that again, + * this part should be skipped just now. + */ + void OnSolveMessage(const char* s, int nbs) { + auto n = std::strlen(s); + if (nbs) { + if (nbs > n) + nbs = n; + s += nbs; + n -= nbs; + } + if (n > 0) { + std::printf("%s", s); + std::fflush(stdout); + } + } + + /** + * Can be ignored by external systems. + * @return non-zero to stop solution input. + */ + int OnAMPLOptions(const AMPLOptions& ) { return 0; } + + /** + * Dual values for algebraic constraints, + * if provided in the solution. + * Number of values <= NumAlgCons(). + * Implementation: + * + * duals.reserve(rd.Size()); + * while (rd.Size()) + * duals.push_back(rd.ReadNext()); + */ + template + void OnDualSolution(VecReader& rd) { + Model().sol_y_.reserve(rd.Size()); + for (int i=0; rd.Size(); ++i) + Model().sol_y_.push_back( rd.ReadNext() ); + } + + /** + * Variable values, if provided. + * Number of values <= NumVars(). + */ + template + void OnPrimalSolution(VecReader& rd) { + Model().sol_x_.reserve(rd.Size()); + for (int i=0; rd.Size(); ++i) + Model().sol_x_.push_back( rd.ReadNext() ); + } + + /** + * Receive notification of the objective index + * used by the driver (solver option 'objno'). + */ + void OnObjno(int on) { Model().objno_=on; } + + /** + * Receive notification of the solve code. + */ + void OnSolveCode(int sr) { Model().solve_result_ = sr; } + + /** + * OnIntSuffix(). + * + * For constraints, can include values for + * logical constraints (after algebraic.) + * Sparse representation - can be empty + * (i.e., all values zero.) + * + * const auto& si = sr.SufInfo(); + * int kind = si.Kind(); + * int nmax = nitems_max[kind & 3]; + * const std::string& name = si.Name(); + * const std::string& table = si.Table(); + * while (sr.Size()) { + * std::pair val = sr.ReadNext(); + * if (val.first<0 || val.first>=nmax) { + * sr.SetError(mp::SOL_Read_Bad_Suffix, + * "bad suffix element index"); + * return; + * } + * suf[val.first] = val.second; + * } + * if (mp::SOL_Read_OK == sr.ReadResult()) // Can check + * RegisterSuffix(kind, name, table, suf); + */ + template + void OnIntSuffix(SuffixReader& sr) { + const auto& si = sr.SufInfo(); + int kind = si.Kind(); + const std::string& name = si.Name(); + assert(0 == (kind & 4)); + auto& suf_val = Model().suf_out_[std::make_pair(name, kind)]; + while (sr.Size()) { + std::pair val = sr.ReadNext(); + if (suf_val.size() <= val.first) + suf_val.resize(val.first+1); + suf_val[val.first] = val.second; + } + } + + /** + * Same as OnIntSuffix(), but + * sr.ReadNext() returns pair + */ + template + void OnDblSuffix(SuffixReader& sr) { + const auto& si = sr.SufInfo(); + int kind = si.Kind(); + const std::string& name = si.Name(); + assert(0 != (kind & 4)); + auto& suf_val = Model().suf_out_[std::make_pair(name, kind)]; + while (sr.Size()) { + std::pair val = sr.ReadNext(); + if (suf_val.size() <= val.first) + suf_val.resize(val.first+1); + suf_val[val.first] = val.second; + } + } + + + ////////////////////////////////////////////// + /// Print solution. + /// @param stub: stub filename used. + void PrintSolution(const std::string& stub); + + +protected: + const ExampleModel& Model() const { return mdl_; } + ExampleModel& Model() { return mdl_; } + +private: + ExampleModel& mdl_; +}; + + +void ExampleSOLHandler2::PrintSolution( + const std::string& stub) { + printf( + "\n ********** SOLUTION (%s.sol) ***********\n", + stub.c_str()); + printf("%s\n", "Duals:"); + for (auto y: Model().sol_y_) + printf("%.17g\t", y); + printf("\n%s\n", "Primals:"); + for (auto x: Model().sol_x_) + printf("%.17g\t", x); + + printf("\nObjno used: %d, solve_result_num: %d\n", + Model().objno_, Model().solve_result_); + + printf("\n%s\n", "Suffixes:"); + for (auto suf: Model().suf_out_) { + printf(" %s [kind=%d]: ", + suf.first.first.c_str(), suf.first.second); + for (auto v: suf.second) + printf("\t%.17g", v); + printf("\n"); + } + printf("%s\n", ""); +} + +#endif // NLSOL_EX_SOL_H diff --git a/nl-writer2/include/mp/nl-feeder2.h b/nl-writer2/include/mp/nl-feeder2.h new file mode 100644 index 000000000..acb1059a1 --- /dev/null +++ b/nl-writer2/include/mp/nl-feeder2.h @@ -0,0 +1,534 @@ +/** + mp::NLFeeder2. + Interface for a model feeder into mp::NLWriter2. + mp::NLWriter2 is a zero-overhead NL model writer + implemented with inline template code. In particular, + it does not store any intermediate model representation. + + NL is a format for representing optimization problems such as linear, + quadratic, nonlinear, complementarity and constraint programming problems + in discrete or continuous variables. It is described in the technical report + "Writing .nl Files" (http://www.cs.sandia.gov/~dmgay/nlwrite.pdf). + + Usage: recommended via mp::NLSOL class. + + If you need the NL writing part only, proceed as follows: + + MyNLFeeder feeder; + mp::NLUtils nlutils; + auto result = mp::WriteNLFile("model", feeder, nlutils); + if (mp::WriteNL_OK != result.first) { + ... + } + + where feeder is an object that provides information on model + components. Below is an interface for such a feeder. + + See also mp::NLReader and mp::NLHandler classes. + + Copyright (C) 2023 AMPL Optimization, Inc. + + Permission to use, copy, modify, and distribute this software and its + documentation for any purpose and without fee is hereby granted, + provided that the above copyright notice appear in all copies and that + both that the copyright notice and this permission notice and warranty + disclaimer appear in supporting documentation. + + The author and AMPL Optimization Inc. disclaim all warranties with + regard to this software, including all implied warranties of + merchantability and fitness. In no event shall the author be liable + for any special, indirect or consequential damages or any damages + whatsoever resulting from loss of use, data or profits, whether in an + action of contract, negligence or other tortious action, arising out + of or in connection with the use or performance of this software. + + Author: Gleb Belov + */ + +#ifndef NLFEEDER2_H +#define NLFEEDER2_H + +#include +#include + +#include "mp/nl-header.h" + + +namespace mp { + +/** + \rst + NLFeeder2: writes model details on request + via provided callback objects. + + For the NL format, variables and constraints must have certain order. + + **Variable ordering:** first continuous, then integer. + Some solvers might require more elaborate ordering, see NLHeader. + + **Constraint ordering:** + first algebraic (including complementarity), then logical. + Some solvers might require nonlinear constraints first. + + `~mp::NLFeeder2` can be used as a base class for other feeders, + or just be an interface example. + + @param: *Impl* is a type derived from `~mp::NLFeeder`. + + @param: *ExprType* is a type storing expressions from + methods such as `~mp::NLFeeder2::FeedExpr`. If not used, + it can be any default-constructible type. + \endrst + */ +template +class NLFeeder2 { +public: + /** Destroys the object. */ + virtual ~NLFeeder2() {} + + /** The expression type. */ + typedef ExprType Expr; + + + ///////////////////// 1. NL HEADER AND OPTIONS ///////////////// + /** Provide NLHeader. + * + * This method is called first. + * + * NLHeader summarizes the model and provides some + * technical parameters, + * such as text/binary NL format. */ + NLHeader Header() { assert(0); return {}; } + + /// NL comments? + bool WantNLComments() const { return false; } + + /// The maximum number of significant digits written. + /// The default value requests full precision, which + /// might be the shortest representation that, when + /// converted to binary and properly rounded, will + /// give exactly the binary value stored in the computer. + int OutputPrecision() const { return 0; } + + /// Write bounds first? + /// The default is yes in AMPL, controlled by + /// (the value of option nl_permute) & 32 + /// (the bit is 0 for yes). + /// Changing this option is deprecated, see + /// https://netlib.org/ampl/changes. + bool WantBoundsFirst() const { return true; } + + /// Want Jacobian column sizes? + /// Required by some nonlinear solvers. + /// Options: 0 - none, 1 - cumulative, + /// 2 - non-cumulative. + /// This option controls how ColSizeWriter + /// writes the provided sizes (which should be + /// non-cumulative). + int WantColumnSizes() const { return 1; } + + + ///////////////////// 2. OBJECTIVES ///////////////////// + /** Description for objective function \a i + * (\a i in 0..num_objs-1). + * With WantNLComments()==true, this is + * written to text-format NL as a comment. */ + const char* ObjDescription(int i) { return ""; } + + /** Provide type of objective \a i. + * 0 - minimization; + * 1 - maximization. */ + int ObjType(int i) { return {}; } + + /** Feed gradient for objective \a i. + * Should include entries for all potentially + * nonzero elements (sparsity pattern). + * + * Implementation skeleton: + * if (obj_grad[i].size()) { + * auto sv = svw.MakeVectorWriter(obj_grad[i].size()); + * for (size_t j=0; j + void FeedObjGradient(int i, ObjGradWriterFactory& ) { } + + /** Feed nonlinear expression of objective \a i. + * + * Implementation example: + * ew.EPut(obj_root_expr[i]); + * + * Details of ObjExprWriter: see NLWriter2. */ + template + void FeedObjExpression(int i, ObjExprWriter& ) { } + + + ///////////////////// 3. DEFINED VARIABLES ///////////////////// + /** Defined variables. + * + * Classical NL writes first the defined variables + * which are used in several places (constraints and/or + * objectives). Defined variables used in a single place + * (1 constraint, or 1 objective), are written + * just before the expression tree of their usage. + * + * For most solvers, this requirement can be ignored + * and this method can return all defined variables + * in the first group (for \a i=0). + * + * The method is guaranteed to be called in the following order: + * 1. For \a i=0; + * 2. For \a i>0, increasing, before constraint \a (i-1)'s expression; + * 3. For \a i<0, decreasing, before objective \a (-i-1)'s expression. + * + * @param i: + * - For \a i=0, feed a sequence of defined variables + * used in several constraints and/or objectives. + * - For \a i>0, feed the defined variables used solely + * in constraint \a i-1. + * - For \a i<0, feed the defined variables used solely + * in objective \a -i-1. + * + * Implementation skeleton: + * // dvar_index in num_vars..num_vars+num_defvars-1. + * for (int dvar_index: dvar_indexes[i]) { + * auto dv = dvw.StartDefVar(dvar_index, lin_nnz, name_or_comment); + * /////////// Write the linear part: + * auto linw = dv.GetLinExprWriter(); + * for (int i=0; i + void FeedDefinedVariables(int i, DefVarWriterFactory& ) { } + + + ///////////////////// 4. VARIABLE BOUNDS ///////////////////// + /** Bounds for variables (except defined variables). + * Use +-inf for missing lower and/or upper bounds. + * Note that variable type is given by variable ordering, + * see NLHeader. + * + * Implementation skeleton: + * for (int i = 0; i < hdr.num_vars; i++) + * vbw.WriteLbUb(lb[i], ub[i]); + */ + template + void FeedVarBounds(VarBoundsWriter& ) { } + + + ///////////////// 5. CONSTRAINT BOUNDS & COMPLEMENTARITY /////// + /// \rst + /// Algebraic constraint bounds (for a single constraint): + /// either range (lb, ub), + /// or complementarity info (k, cvar), when k>0. + /// + /// For a complementarity constraint to hold, if cvar is at + /// its lower bound, then body >= 0; if cvar is at its upper + /// bound, then body <= 0; + /// and if cvar is strictly between its bounds, then body = 0. + /// The integer k in a complementarity constraint line indicates + /// which bounds on cvar are finite: 1 and 3 imply a finite + /// lower bound; 2 and 3 imply a finite upper bound; 0 (which + /// should not occur) would imply no finite bounds, i.e., + /// body = 0 must always hold. + /// + /// Example: + /// + /// .. code-block:: ampl + /// + /// ampl: var x; var y; var z; + /// ampl: s.t. Compl1: x+y >= 3 complements x-z <= 15; + /// ampl: s.t. Compl2: -2 <= 2*y+3*z <= 13 complements 6*z-2*x; + /// ampl: expand; + /// subject to Compl1: + /// 3 <= x + y + /// complements + /// x - z <= 15; + /// + /// subject to Compl2: + /// -2 <= 2*y + 3*z <= 13 + /// complements + /// -2*x + 6*z; + /// + /// ampl: solexpand; + /// Nonsquare complementarity system: + /// 4 complementarities including 2 equations + /// 5 variables + /// subject to Compl1.L: + /// x + y + Compl1$cvar = 0; + /// + /// subject to Compl1.R: + /// -15 + x - z <= 0 + /// complements + /// Compl1$cvar <= -3; + /// + /// subject to Compl2.L: + /// 2*y + 3*z - Compl2$cvar = 0; + /// + /// subject to Compl2.R: + /// -2*x + 6*z + /// complements + /// -2 <= Compl2$cvar <= 13; + /// + /// \endrst + struct AlgConRange { + double L{}, U{}; + int k{0}, cvar{0}; // k>0 means complementarity to cvar + }; + + /** Bounds/complementarity for all algebraic constraints + * (\a num_algebraic_cons). + * + * Implementation skeleton: + * for (int j=0; j negInfinity) + * bnd.k = 1; + * if (vub[j] < Infinity) + * bnd.k |= 2; + * assert(bnd.k); + * bnd.cvar = j; + * } else { + * bnd.L = clb[j]; + * bnd.U = cub[j]; + * } + * cbw.WriteAlgConRange(bnd); + * } + */ + template + void FeedConBounds(ConBoundsWriter& ) { } + + + ///////////////////// 6. CONSTRAINTS ///////////////////// + /** Description of constraint \a i + * (\a i in 0..num_algebraic_cons+num_logical_cons-1). + * With WantNLComments()==true, this is + * written to text-format NL as a comment. */ + const char* ConDescription(int ) { return ""; } + + /** Feed the linear part of algebraic constraint \a i. + * For smooth solvers, should contain entries for all + * potential nonzeros (Jacobian sparsity pattern). + * + * Implementation skeleton: + * if (con_grad[i].size()) { + * auto sv = svw.MakeVectorWriter(con_grad[i].size()); + * for (size_t j=0; j + void FeedLinearConExpr(int i, ConLinearExprWriterFactory& ) { } + + /** Feed nonlinear expression of constraint \a i. + * Algebraic constraints (num_algebraic_cons) + * come before logical (num_logical_cons). + * For linear constraints, the expression should be + * constant 0. + */ + template + void FeedConExpression(int i, ConExprWriter& ) { } + + + ///////////////////// 7. EXPRESSIONS ///////////////////// + /** Feed native expression. + * This method is recursively called from NLWriter, + * when Feeder uses ExprWriter::EPut(). + * Feeder should not call this method itself. + * + * Details of ExprWriter: see NLWriter2. + */ + template + void FeedExpr(Expr e, ExprWriter& ) { } + + + ///////////////////// 8. PL-SOS CONSTRAINTS //////////// + /** + * The below feature is for AMPL's internal + * linearization of piecewise-linear functions. + * For user-definable SOS constraints, use suffixes + * .sosno/.ref. + * + * The below is a feeder interface + * for .sos/.sosref suffixes. + * The feeder can provide 3 sparse vectors: + * - .sos for variables: + * Each nonzero value defines SOS group number. + * Negative means SOS Type 2, positive - SOS Type 1. + * - .sos for constraints: + * Each nonzero value denotes a constraint used in a + * linearization of an SOS. The constraint can be deleted + * by the solver driver if using solver's SOS. + * - .sosref for variables: + * SOS weights. Variables participating in an SOS having + * zero weights are involved in linearization and can be + * deleted if the solver accepts SOS natively. + * + * Implementation: + * auto sosv = plsos.StartSOSVars(nvsos); + * for (int i=0; insosref); + * .... + */ + template + void FeedPLSOS(PLSOSWriter& ) { } + + + ///////////////////// 9. FUNCTIONS ///////////////////// + /** Function definition. */ + struct FuncDef { + const char* Name() { return ""; } + int NumArgs() { return 0; } + /** Function type. + * 0 - numeric; + * 1 - symbolic. */ + int Type() { return 0; } + }; + + /** Provide definition + * of function \a i, i=0..num_funcs-1. */ + FuncDef Function(int i) { return {}; } + + + ///////////////////// 10. RANDOM VARIABLES ///////////////////// + /// Random variables. + /// Undocumented feature. SNL2006. + /// Example: + /// var z >= 0; + /// let z.stage := 1; + /// var x{0..1, 0..1} random := Uniform(0,2); + /// for {i in 0..1, j in 0..1} {let x[i,j].stage := 1;}; + /// display z.stage, x.stage; + /// c: z * sum{i in 0..1, j in 0..1} x[i,j] <= 3 + Sample(Uniform(0,2)); + /// + /// Feed random variables. + /// Indexes: num_vars+num_common_exprs + /// .. num_vars+num_common_exprs+num_rand_vars-1. + /// + /// Implementation skeleton: + /// for(j = num_vars+num_common_exprs; + /// j < num_vars+num_common_exprs+num_rand_vars; j++) { + /// auto ew = rvw.StartRandVar(j, rand_var_comment(j)); + /// ew.EPut(rand_var_root_expr(j)); + /// } + template + void FeedRandomVariables(RandVarWriterFactory& ) { } + + + ///////////////////// 11. COLUMN SIZES ///////////////////// + + /** Jacobian column sizes. + * Should feed LP column sizes + * for all but the last variable. + * + * Implementation skeleton: + * if (WantColumnSizes()) + * for (int i=0; i < num_vars+num_rand_vars-1; ++i) + * csw.Write(col_size[i]); + */ + template + void FeedColumnSizes(ColSizeWriter& ) { } + + + ///////////////////// 12. INITIAL GUESSES ///////////////////// + /** Initial primal guesses. + * + * Implementation: + * if (ini_guess.size()) { + * auto ig = igw.MakeVectorWriter(ini_guess.size()); + * for (size_t i=0; i + void FeedInitialGuesses(IGWriter& ) { } + + /** Initial dual guesses. */ + template + void FeedInitialDualGuesses(IDGWriter& ) { } + + + ///////////////////// 13. SUFFIXES ///////////////////// + /** Feed suffixes. + * + * For constraints, assume ordering: + * first algebraic, then logical. + * + * Implementation: + * while (....) { + * auto sw = swf.StartIntSuffix( // or ...DblSuffix + * suf_name, kind, n_nonzeros); + * for (int i=0; i + void FeedSuffixes(SuffixWriterFactory& ) { } + + + //////////////////// 14. ROW/COLUMN NAMES ETC ///////////////////// + /** FeedRowAndObjNames: + * Provide constraint, then objective names. + * Name information is optional. + * + * Implementation: + * if ((output_desired) && wrt) + * for (i: ....) + * wrt << name[i].c_str(); + */ + template + void FeedRowAndObjNames(RowObjNameWriter& wrt) { } + + /** Provide deleted row names.*/ + template + void FeedDelRowNames(DelRowNameWriter& ) { } + + /** Provide variable names. */ + template + void FeedColNames(ColNameWriter& ) { } + + /** Provide unused variable names. */ + template + void FeedUnusedVarNames(UnusedVarNameWriter& ) { } + + /** Provide {fixed variable, extra info} pairs. + * This includes defined eliminated variables. + * + * Implementation: + * if ((output_desired) && wrt) + * for (....) + * wrt << typename Writer::StrStrValue + * { name[i].c_str(), comment[i].c_str() }; + */ + template + void FeedFixedVarNames(FixedVarNameWriter& ) { } + + /** Provide {obj name, constant term} pairs. + * + * Implementation: + * if (wrt) + * for (....) + * wrt << typename Writer::StrDblValue + * { name[i].c_str(), (double)obj_offset[i] }; + */ + template + void FeedObjAdj(ObjOffsetWriter& ) { } + +}; + +} // namespace mp + +#endif // NLFEEDER2_H diff --git a/nl-writer2/include/mp/nl-header.h b/nl-writer2/include/mp/nl-header.h new file mode 100644 index 000000000..bd8ac6f01 --- /dev/null +++ b/nl-writer2/include/mp/nl-header.h @@ -0,0 +1,375 @@ +/** + Common definitions for NL reader and NL writer. + + Copyright (C) 2014 - 2023 AMPL Optimization Inc. + + Permission to use, copy, modify, and distribute this software and its + documentation for any purpose and without fee is hereby granted, + provided that the above copyright notice appear in all copies and that + both that the copyright notice and this permission notice and warranty + disclaimer appear in supporting documentation. + + The author and AMPL Optimization Inc disclaim all warranties with + regard to this software, including all implied warranties of + merchantability and fitness. In no event shall the author be liable + for any special, indirect or consequential damages or any damages + whatsoever resulting from loss of use, data or profits, whether in an + action of contract, negligence or other tortious action, arising out + of or in connection with the use or performance of this software. + + Authors: Victor Zverovich, Gleb Belov + */ + +#ifndef NLHEADER_H +#define NLHEADER_H + +#include +#include +#include +#include + +/// Namespace mp. +namespace mp { + + +/** Information about an optimization problem. + * Full documentation on the NL format: + * technical report "Writing .nl Files" + * (https://ampl.github.io/nlwrite.pdf.) + */ +struct ProblemInfo { + /** Total number of variables. */ + int num_vars {0}; + + /** + Number of algebraic constraints including ranges and equality constraints. + It doesn't include logical constraints. + */ + int num_algebraic_cons {0}; + + /** Total number of objectives. */ + int num_objs {0}; + + /** Number of ranges (constraints with -Infinity < LHS < RHS < Infinity). */ + int num_ranges {0}; + + /** + Number of equality constraints or -1 if unknown (AMPL prior to 19970627). + */ + int num_eqns {0}; + + /** Number of logical constraints. */ + int num_logical_cons {0}; + + /** Number of random variables. */ + int num_rand_vars {0}; + + /** Number of random defined variables. */ + int num_rand_common_exprs {0}; + + /** Number of random constraints. */ + int num_rand_cons {0}; + + /** Number of random objectives. */ + int num_rand_objs {0}; + + /** Number of random calls. */ + int num_rand_calls {0}; + + /** Number of stages. */ + int num_stages {0}; + + /** Returns the number of integer variables (includes binary variables). */ + int num_integer_vars() const { + return num_linear_binary_vars + num_linear_integer_vars + + num_nl_integer_vars_in_both + num_nl_integer_vars_in_cons + + num_nl_integer_vars_in_objs; + } + + /** Returns the number of continuous variables. */ + int num_continuous_vars() const { return num_vars - num_integer_vars(); } + + // Nonlinear and complementarity information + // ----------------------------------------- + + /** Total number of nonlinear constraints. */ + int num_nl_cons {0}; + + /** Total number of nonlinear objectives. */ + int num_nl_objs {0}; + + /** Total number of complementarity conditions. */ + int num_compl_conds {0}; + + /** Number of nonlinear complementarity conditions. */ + int num_nl_compl_conds {0}; + + /** Number of complementarities involving double inequalities. */ + int num_compl_dbl_ineqs {0}; + + /** Number of complemented variables with a nonzero lower bound. */ + int num_compl_vars_with_nz_lb {0}; + + // Information about network constraints + // ------------------------------------- + + /** Number of nonlinear network constraints. */ + int num_nl_net_cons {0}; + + /** Number of linear network constraints. */ + int num_linear_net_cons {0}; + + // Information about nonlinear variables + // ------------------------------------- + + /** + Number of nonlinear variables in constraints including nonlinear + variables in both constraints and objectives. + */ + int num_nl_vars_in_cons {0}; + + /** + Number of nonlinear variables in objectives including nonlinear + variables in both constraints and objectives. + */ + int num_nl_vars_in_objs {0}; + + /** Number of nonlinear variables in both constraints and objectives. */ + int num_nl_vars_in_both {0}; + + // Miscellaneous + // ------------- + + /** Number of linear network variables (arcs). */ + int num_linear_net_vars {0}; + + /** Number of functions. */ + int num_funcs {0}; + + // Information about discrete variables + // ------------------------------------ + + /** Number of linear binary variables. */ + int num_linear_binary_vars {0}; + + /** Number of linear non-binary integer variables. */ + int num_linear_integer_vars {0}; + + /** + Number of integer nonlinear variables in both constraints and objectives. + */ + int num_nl_integer_vars_in_both {0}; + + /** Number of integer nonlinear variables just in constraints. */ + int num_nl_integer_vars_in_cons {0}; + + /** Number of integer nonlinear variables just in objectives. */ + int num_nl_integer_vars_in_objs {0}; + + // Information about nonzeros + // -------------------------- + + /** Number of nonzeros in constraints' Jacobian. */ + std::size_t num_con_nonzeros {0}; + + /** Number of nonzeros in all objective gradients. */ + std::size_t num_obj_nonzeros {0}; + + // Information about names. + // Does not have to be filled for NLWriter2. + // ----------------------- + + /** Length of longest constraint or objective name + * if names are available. */ + int max_con_name_len {0}; + + /** Length of longest variable name if names are available. */ + int max_var_name_len {0}; + + // Information about common expressions + // ------------------------------------ + + /** + Number of common expressions that appear both in constraints + and objectives. + */ + int num_common_exprs_in_both {0}; + + /** + Number of common expressions that appear in multiple constraints + and don't appear in objectives. + */ + int num_common_exprs_in_cons {0}; + + /** + Number of common expressions that appear in multiple objectives + and don't appear in constraints. + */ + int num_common_exprs_in_objs {0}; + + /** + Number of common expressions that only appear in a single constraint + and don't appear in objectives. + */ + int num_common_exprs_in_single_cons {0}; + + /** + Number of common expressions that only appear in a single objective + and don't appear in constraints. + */ + int num_common_exprs_in_single_objs {0}; + + /** Returns the total number of common expressions. */ + int num_common_exprs() const { + return num_common_exprs_in_both + num_common_exprs_in_cons + + num_common_exprs_in_objs + num_common_exprs_in_single_cons + + num_common_exprs_in_single_objs; + } +}; + + +enum { + /** Maximum number of options reserved for AMPL use in NL and SOL formats. */ + MAX_AMPL_OPTIONS = 9 +}; + +enum { + VBTOL_OPTION_INDEX = 1, + USE_VBTOL_FLAG = 3 +}; + +/// Namespace arith +namespace arith { + +/** Floating-point arithmetic kind. */ +enum Kind { + + /** Unknown floating-point arithmetic. */ + UNKNOWN = 0, + + /** + \rst + Standard `IEEE-754 floating point + `_ - little endian. + \endrst + */ + IEEE_LITTLE_ENDIAN = 1, + + /** Standard IEEE-754 floating point - big endian. */ + IEEE_BIG_ENDIAN = 2, + + /** + \rst + `IBM floating point + `_. + \endrst + */ + IBM = 3, + + /** VAX floating point (legacy). */ + VAX = 4, + + /** Cray floating point. */ + CRAY = 5, + + /** Last floating point. */ + LAST = CRAY +}; + +/// Returns floating-point arithmetic kind used on the current system. +Kind GetKind(); + +/// IsIEEE +inline bool IsIEEE(arith::Kind k) { + return k == IEEE_LITTLE_ENDIAN || k == IEEE_BIG_ENDIAN; +} +} // namespace arith + + +/** + \rst + An NL `header `_ + which contains information about problem dimensions, such as the number of + variables and constraints, and the input format. + + Base class: `mp::ProblemInfo` + \endrst + */ +struct NLHeader : ProblemInfo { + /** Input/output format */ + enum Format { + /** + Text format. The text format is fully portable meaning that an .nl file + can be written on a machine of one architecture and then read on a + machine of a different architecture. + */ + TEXT = 0, + + /** + Binary format. The binary format is not generally portable and should + normally be used on a single machine. + */ + BINARY = 1 + }; + + /** Input/output format. */ + Format format {TEXT}; + + /** The number of options reserved for AMPL use. */ + int num_ampl_options {3}; + + /** + Values of options reserved for AMPL use. + Leave the default values if not using AMPL. + */ + long ampl_options[MAX_AMPL_OPTIONS] = {1, 1, 0}; + + /** + Extra info for writing a solution reserved for AMPL use. + Leave the default value if not using AMPL. + */ + double ampl_vbtol {0.0}; + + /** + * Problem name. + */ + std::string prob_name; + + /** + \rst + Floating-point arithmetic kind used with binary format to check + if an .nl file is written using a compatible representation of + floating-point numbers. It is not used with the text format and normally + set to `mp::arith::UNKNOWN` there. + \endrst + */ + arith::Kind arith_kind {arith::IEEE_LITTLE_ENDIAN}; + + /** Flags. */ + enum { + /** Flag that specifies whether to write output suffixes to a .sol file. */ + WANT_OUTPUT_SUFFIXES = 1 + }; + + /** + \rst + Flags. Can be either 0 or `mp::NLHeader::WANT_OUTPUT_SUFFIXES`. + \endrst + */ + int flags {WANT_OUTPUT_SUFFIXES}; + + NLHeader() + : ProblemInfo(), + format(TEXT), + num_ampl_options(3), ampl_vbtol(0), + arith_kind(arith::IEEE_LITTLE_ENDIAN), + flags(WANT_OUTPUT_SUFFIXES) { + std::fill(ampl_options, ampl_options + MAX_AMPL_OPTIONS, 0); + std::array opt_default {1, 1, 0}; + std::copy(opt_default.begin(), opt_default.end(), ampl_options); + } +}; + +} // namespace mp + +#endif // NLHEADER_H diff --git a/nl-writer2/include/mp/nl-utils2.h b/nl-writer2/include/mp/nl-utils2.h new file mode 100644 index 000000000..5ce625d54 --- /dev/null +++ b/nl-writer2/include/mp/nl-utils2.h @@ -0,0 +1,95 @@ +#ifndef NLUTILS2_H +#define NLUTILS2_H + +#include +#include +#include +#include + + +namespace mp { + +/// A wrapper for file I/O. +class File { +public: + /// Destructor + ~File() { Close(); } + /// Constructor. + File() { } + /// Constructor. + File(const File& f) { assert(!f_ && !f.f_); } + /// Constructor. + File(File&& f) : f_(f.f_) { f.f_ = nullptr; } + /// copy assign + void operator=(const File& f) { assert(!f_ && !f.f_); } + /// move assign + void operator=(File&& f) noexcept { std::swap(f_, f.f_); } + + /// Check being open for now. + operator bool() const { return f_; } + /// Open file. + bool Open(const char* nm, const char* mode) { + f_ = std::fopen(nm, mode); + return (*this); + } + /// Close file. + void Close() { + if (f_) { + std::fclose(f_); + } + f_ = nullptr; + } + + /// Printf + void Printf(const char* format, ...); + + /// Obtain handler + FILE* GetHandle() const { return f_; } + +private: + FILE *f_{nullptr}; +}; + + +/// NL writer and SOL reader utilities. +/// It provides default facilities for logging +/// and error handling. +/// The default error handler exit()s. +class NLUtils { +public: + /// Destructor + virtual ~NLUtils() { } + /// File opener + /// @param Close: remove file instead + virtual File openf(const std::string& fname, + int Close, const char *mode); + /// log message + virtual void log_message(const char* format, ...); + /// log warning + virtual void log_warning(const char* format, ...); + /// Whether to show file names + /// in the default openf() via log_message(). + virtual bool if_show_filenames() { return show_fln_; } + /// Set: whether to show file names + virtual void set_show_filenames(bool v) { show_fln_ = v; } + + /// Override this to your error handler. + /// Not using exceptions by default. + /// Only called with wrong output format string + /// (internal error.) + virtual void myexit(const std::string& msg) { + // No help including after on Mac + using namespace std; + fprintf(stderr, "%s\n", msg.c_str()); + exit(1); + } + + +private: + bool show_fln_{false}; +}; + + +} // namespace mp + +#endif // NLUTILS2_H diff --git a/nl-writer2/include/mp/nl-writer2-misc.h b/nl-writer2/include/mp/nl-writer2-misc.h new file mode 100644 index 000000000..fc0e8d8ad --- /dev/null +++ b/nl-writer2/include/mp/nl-writer2-misc.h @@ -0,0 +1,80 @@ +#ifndef NLWRITER2MISC_H +#define NLWRITER2MISC_H + +#include +#include +#include +#include +#include + +#include "mp/nl-header.h" +#include "mp/nl-utils2.h" + +namespace mp { + +/// Text formatter +class TextFormatter { +public: + /// Construct + TextFormatter(NLUtils& u, + bool nlc, + int outprec=0) + : utils_(u), + nl_comments(nlc), + output_prec(outprec) + { } + /// Mode query + NLHeader::Format Mode() const { return NLHeader::TEXT; } + /// Text printf + int apr(File&, const char*, ...); + /// Text nput + void nput(File&, double); + /// Retrieve utils + NLUtils& Utils() { return utils_; } + + +private: + NLUtils& utils_; + const bool nl_comments{false}; + const int output_prec{0}; // 0 means full precision +}; + + +/// Binary formatter +class BinaryFormatter { +public: + /// Construct + BinaryFormatter(NLUtils& u, bool, int) + : utils_(u) { } + /// Mode query + NLHeader::Format Mode() const { return NLHeader::BINARY; } + /// Binary printf + int apr(File&, const char*, ...); + /// Binary nput + void nput(File&, double); + /// Retrieve utils + NLUtils& Utils() { return utils_; } + + +private: + NLUtils& utils_; +}; + + +class StringFileWriter; + + +/// NLWriter2 parameters. +/// @param Formatter: low-level writer: binary or text. +/// @param Feeder2: a class implementing the NLFeeder2 +/// concept that feeds model information on request. +/// @param Utils: writer utilities +template +struct NLWriter2Params { + using FormatterType=Formatter; + using FeederType=Feeder2; +}; + +} // namespace mp + +#endif // NLWRITER2MISC_H diff --git a/nl-writer2/include/mp/nl-writer2.h b/nl-writer2/include/mp/nl-writer2.h new file mode 100644 index 000000000..8c9eceb3a --- /dev/null +++ b/nl-writer2/include/mp/nl-writer2.h @@ -0,0 +1,597 @@ +/** + NL writer2 + + NL is a format for representing optimization problems such as linear, + quadratic, nonlinear, complementarity and constraint programming problems + in discrete or continuous variables. It is described in the technical report + "Writing .nl Files" (http://www.cs.sandia.gov/~dmgay/nlwrite.pdf). + + This is a complete reusable C++ implementation of an NL writer. + + Usage: + /// Write an NL file: + auto result = mp::WriteNLFile(filenamebase, feeder, utils); + if (mp::WriteNL_OK != result.first) { + ... + } + + where feeder is an object that provides information on model + components. See NLFeeder2 for an interface of a feeder class. + + See also NLReader and NLHandler classes. + + Copyright (C) 2023 AMPL Optimization Inc. + + Permission to use, copy, modify, and distribute this software and its + documentation for any purpose and without fee is hereby granted, + provided that the above copyright notice appear in all copies and that + both that the copyright notice and this permission notice and warranty + disclaimer appear in supporting documentation. + + The author and AMPL Optimization Inc disclaim all warranties with + regard to this software, including all implied warranties of + merchantability and fitness. In no event shall the author be liable + for any special, indirect or consequential damages or any damages + whatsoever resulting from loss of use, data or profits, whether in an + action of contract, negligence or other tortious action, arising out + of or in connection with the use or performance of this software. + + Author: Gleb Belov + */ + +#ifndef MP_NLWriter22_H +#define MP_NLWriter22_H + +#include +#include +#include +#include +#include +#include + +#include "mp/nl-writer2-misc.h" +/// Just for checking compilation +#include "mp/nl-feeder2.h" + + +namespace mp { + +/// WriteNLResult enum. +enum WriteNLResult { + WriteNL_Unset = 0, + WriteNL_OK = 1, + WriteNL_CantOpen, + WriteNL_Failed +}; + +/// Typedef WriteNLReport: +/// result code and error message +using WriteNLReport = std::pair; + +/// Write NL file and any necessary auxiliary files. +/// @param namebase: name without extension, +/// to be augmented by .nl, .col, ...etc. +/// @return Write status. +/// Note that warnings are reported via \a utl. +template +inline WriteNLReport WriteNLFile( + const std::string& namebase, + NLFeeder2& nlf, NLUtils& utl); + + +/// NLWriter2. +/// +/// mp::NLWriter2 is a zero-overhead NL model writer +/// implemented with inline template code. In particular, +/// it does not store any intermediate model representation. +/// +/// Usage: recommended via mp::NLSOL class. +/// If you need the NL writing part only, proceed as follows: +/// +/// Write an NL file: +/// WriteNLFile(filenamebase, feeder, utils); +/// +/// Write an NL string: +/// WriteNLString(string, feeder, utils); +/// +/// where feeder is an object that provides information on model +/// components. See NLFeeder2 for an interface of a feeder class. +/// +/// See also NLReader and NLHandler classes. +/// +/// @param Params: a specialization of class NLWriter2Params<>. +template +class NLWriter2 : + public Params::FormatterType { +public: + /// Destructor. + virtual ~NLWriter2() { } + + using FormatterType = typename Params::FormatterType; + using FeederType = typename Params::FeederType; + + /// Constructor. + NLWriter2(FeederType& f, NLHeader h, NLUtils& utl) : + FormatterType(utl, f.WantNLComments(), f.OutputPrecision()), + feeder_(f), header_(h) { } + + /// Write to files with the given namebase. + /// @param namebase will be used for filenames + /// namebase.nl, namebase.row, etc. + /// @return Write status. + WriteNLReport WriteFiles(const std::string& namebase); + + /// Retrieve WriteNLResult and error message + WriteNLReport GetResult() const + { return result_; } + + +protected: + /// Write auxiliary files .row, .slc, .col, etc, + /// if data provided by the corresp. feeders + /// (otherwise, any existing files are deleted). + void WriteAuxFiles(const std::string& namebase); + + /// Write NL file, if any variables + void WriteNL(const std::string& namebase); + + + void WriteRowObjNames(const std::string& namebase); + void WriteRowDelNames(const std::string& namebase); + void WriteColNames(const std::string& namebase); + void WriteUnusedVarNames(const std::string& namebase); + void WriteFixedVars(const std::string& namebase); + void WriteObjAdj(const std::string& namebase); + + + void WriteNLHeader(); + void WriteFunctions(); // function defs + void WriteSuffixes(); + void WritePLSOSConstraints(); + + void WriteVarBounds(); + void WriteInitialGuesses(); + void WriteConBounds(); + void WriteDualInitialGuesses(); + + /// i=0: used in several places, + /// i>0 increasing: used in 1 con, + /// i<0 decreasing: used in 1 obj + void WriteDefinedVariables(int i); + + void WriteConObjExpressions(); + + void WriteColumnSizes(); + + void WriteLinearConExpr(); + void WriteObjGradients(); + + void WriteRandomVariables(); + + +protected: + //////////////////////////////////////////////////////// + /// Writer callbacks. + /// The callbacks are employed by user-specialized feeder + /// when asked to feed something. + + /// Sparse vector writer. + /// Is constructed knowing the number + /// of nozeros to be written. + template + class SparseVectorWriter { + public: + /// Constructor + SparseVectorWriter(NLWriter2& nlw, size_t n); + /// Destructor + ~SparseVectorWriter() { assert(0 == n_entries_); } + /// Write next entry + void Write(Index , Value ); + /// Number of outstanding elements + int NLeft() const { return n_entries_; } + + private: + NLWriter2& nlw_; + size_t n_entries_; + }; + + /// Typedef SparseDblVecWriter + using SparseDblVecWriter = SparseVectorWriter; + /// Typedef SparseIntVecWriter + using SparseIntVecWriter = SparseVectorWriter; + + + /// Single sparse vector writer factory. + /// Can be used once to produce a SparseVectorWriter. + /// At first it writes the number of nonzeros + /// using the given format string. + template + class SingleSparseVecWrtFactory { + public: + /// Construct. + /// @param fmt: format string for printf() + /// containing %d for the number of sparse elements + /// and finishing with '\n', + /// e.g., "x%d\t# initial guess\n" + SingleSparseVecWrtFactory(NLWriter2& nlw, const char* fmt) + : nlw_(nlw), fmt_(fmt) { } + /// Construct. More general. + /// @param formatter: a lambda printing something + /// given N nonzeros. + SingleSparseVecWrtFactory(NLWriter2& nlw, + std::function hdr_prn) + : nlw_(nlw), hdr_prn_(hdr_prn) { } + /// Create a vector writer + SparseVectorWriter MakeVectorWriter(size_t nnz) { + assert(0 == nInst_++); + if (fmt_) + nlw_.apr(nlw_.nm, fmt_, (int)nnz); + else if (hdr_prn_) + hdr_prn_(nnz); + else { + assert(0); + } + return {nlw_, nnz}; + } + + private: + NLWriter2& nlw_; + const char* fmt_ = nullptr; + std::function hdr_prn_; + int nInst_ = 0; + }; + + /// Typedef SingleSparseDblVecWrtFactory + using SingleSparseDblVecWrtFactory + = SingleSparseVecWrtFactory; + /// Typedef SingleSparseIntVecWrtFactory + using SingleSparseIntVecWrtFactory + = SingleSparseVecWrtFactory; + + + /** Declare actual expression writer. */ + class ExprWriter; + + /** Expression argument writer. + * An object of this class is returned by FuncPut() + * and Oput..() of the parent expression's ExprWriter. + * This class acts as a proxy for ExprWriter's + * of the arguments. + * + * ExprArgWriter(na=1) is passed to write the root of a tree, + * such as of the tree for a certain constraint or + * a defined variable, or recursively, for + * a user expression type, when user calls EPut(). + */ + class ExprArgWriter { + public: + /// Construct for \a na arguments. + ExprArgWriter(NLWriter2& nlw, int na); + /// Destruct. Check that all arguments are written. + ~ExprArgWriter(); + + /// Write the next arg as Feeder's native expression. + /// This recursively calls Feeder::FeedExpr(). + void EPut(typename FeederType::Expr e); + + /** Write the next arg as 'variable reference'. + * 0 <= index < num_vars is a solver variable; + * index >= num_vars is a defined variable. */ + void VPut(int v, const char* descr=""); + + /** Write numeric constant expression. */ + void NPut(double x); + + /** Write string constant expression. */ + void StrPut(const char* ); + + /** Write the next arg as function call expression. */ + ExprArgWriter FuncPut( + int index, int nArgs, const char* descr=""); + + /// Write the next arg as AMPL opcode for a unary op. + /// @return 1-arg writer. + ExprArgWriter OPut1(int opcode, const char* descr=""); + /// Write AMPL opcode for a binary op. + ExprArgWriter OPut2(int opcode, const char* descr=""); + /// Write AMPL opcode for a 3-arg op. + ExprArgWriter OPut3(int opcode, const char* descr=""); + + /// Write AMPL opcode for an iterated op (min, exists, sum, etc). + /// For a piecewise-linear expression, \a nArgs should be + /// 2*(N slopes) and the arguments are: + /// break points, slopes, argument variable. + ExprArgWriter OPutN( + int opcode, int nArgs, const char* descr=""); + + protected: + /// Add an argument explicitly. + /// @return the expr writer for the argument. + /// Note that the ..Put. methods transfer their call + /// to the result of an AddArg(). + ExprWriter AddArg(); + + private: + NLWriter2& nlw_; + int nargs_; + }; + + +protected: + /** Actual expression writer. */ + class ExprWriter { + public: + /// Construct + ExprWriter(NLWriter2& nlw) : nlw_(nlw) { } + + /// Write Feeder's native expression. + /// This recursively calls Feeder::FeedExpr(). + void EPut(typename FeederType::Expr e); + + /** Write expression 'variable reference'. + * 0 <= index < num_vars is a solver variable; + * index >= num_vars is a defined variable. */ + void VPut(int v, const char* descr=""); + + /** Write numeric constant expression. */ + void NPut(double x); + + /** Write string constant expression. */ + void StrPut(const char* ); + + /** Write the next arg as function call expression. */ + ExprArgWriter FuncPut( + int index, int nArgs, const char* descr=""); + + /// Write the next arg as AMPL opcode for a unary op. + /// @return 1-arg writer. + ExprArgWriter OPut1(int opcode, const char* descr=""); + /// Write AMPL opcode for a binary op. + ExprArgWriter OPut2(int opcode, const char* descr=""); + /// Write AMPL opcode for a 3-arg op. + ExprArgWriter OPut3(int opcode, const char* descr=""); + + /// Write AMPL opcode for an iterated op (min, exists, sum, etc). + /// For a piecewise-linear expression, \a nArgs should be + /// 2*(N slopes) and the arguments are: + /// break points, slopes, argument variable. + ExprArgWriter OPutN( + int opcode, int nArgs, const char* descr=""); + + private: + NLWriter2& nlw_; + }; + + + /** Writer of a defined variable. */ + class DefVarWriter { + public: + /// Construct + DefVarWriter(NLWriter2& nlw, int nnzlin); + + /// Write entries c*var[v] to the linear part + /// of the defining expression. + /// All nnz entries should be written before + /// the nonlinear expression. + /// This method can be called only once. + SparseDblVecWriter GetLinExprWriter(); + + /// Retrieve the nonlinear expression writer. + /// Should be used after the linear expression. + ExprArgWriter GetExprWriter(); + + private: + int nnzlin_; + NLWriter2& nlw_; + }; + + + /** Write a sequence of defined variables, + * for example, all such used in a certain constraint. */ + class DefVarWriterFactory { + public: + /// Construct. + /// For the meaning of \a k, see NLFeeder2. + DefVarWriterFactory(NLWriter2& nlw, int k) + : nlw_(nlw), k_(k) { } + + /// Start writing a defined variable. + /// + /// @param index: defined variable index, used to + /// reference it in subsequent expression graphs. + /// Thus, the index should be >= num_vars, + /// < num_vars+num_common_exprs. + /// Providing the index explicitly, because classical NL + /// likes special order of defined variables. + /// + /// @param nnz: number of nonzeros in the linear part. + /// + /// @param descr: meta-information - what is this variable, + /// for example, "nl(t[2])". + /// Providing it here because it's not + /// included in ColNames(). + /// + /// @return A callback object writing + /// a single defined variable. + DefVarWriter StartDefVar( + int index, int nnz, const char* descr=""); + + private: + NLWriter2& nlw_; + int k_; // for which constraint / obj + }; + + + /** Write \a num_vars variable bounds + * (all except defined variables). */ + class VarBndWriter { + public: + /// Construct + VarBndWriter(NLWriter2& nlw) : nlw_(nlw) { } + /// Write range for the next variable. + void WriteLbUb(double lb, double ub); + /// Get N written + int GetNWritten() const { return nWrt_; } + + private: + NLWriter2& nlw_; + int nWrt_ = 0; + }; + + /** Write \a num_algebraic_cons constraint bounds. */ + class ConBndWriter { + public: + /// Construct + ConBndWriter(NLWriter2& nlw) : nlw_(nlw) { } + /// Write range/complementarity for the next constraint. + template + void WriteAlgConRange(AlgConRange ); + /// Get N written + int GetNWritten() const { return nWrt_; } + + private: + NLWriter2& nlw_; + int nWrt_ = 0; + }; + + /** Write num_vars+num_rand_vars-1 + * column sizes */ + class ColSizeWriter { + public: + /// Construct + ColSizeWriter(NLWriter2& nlw, int k) + : nlw_(nlw), kind_(k) { } + /// Write next col's size + void Write(int s) { + switch(kind_) { + case 1: + sum_ += s; + nlw_.apr(nlw_.nm, "%z\n", sum_); + break; + case 2: + nlw_.apr(nlw_.nm, "%d\n", s); + break; + default: + assert(0 && "why am I here?"); + } + ++nWrt_; + } + /// Report N written + int GetNWritten() const { return nWrt_; } + + private: + NLWriter2& nlw_; + const int kind_; + std::size_t sum_ = 0; + int nWrt_ = 0; + }; + + /** Write a sequence of random variables. */ + class RandVarWriterFactory { + public: + /// Construct + RandVarWriterFactory(NLWriter2& nlw) : nlw_(nlw) { } + + /// Start writing a random variable. + /// (To be finished by feeding the defining expression.) + /// + /// @return A callback writing the defining expression. + ExprArgWriter StartRandVar( + int index, const char* descr); + + private: + NLWriter2& nlw_; + }; + + using SuffixIntWriter = SparseVectorWriter; + using SuffixDblWriter = SparseVectorWriter; + + /** Suffixes. */ + class SuffixWriterFactory { + public: + /// Construct + SuffixWriterFactory(NLWriter2& nlw) + : nlw_(nlw) { } + /// Start writing an int-valued suffix. + SuffixIntWriter StartIntSuffix( + const char* name, int kind, int nnz); + /// Start writing an dbl-valued suffix. + SuffixDblWriter StartDblSuffix( + const char* name, int kind, int nnz); + + private: + NLWriter2& nlw_; + }; + + /** PL-SOS constraints. */ + class PLSOSWriter { + public: + /// Construct. + PLSOSWriter(NLWriter2& nlw) : swf_(nlw) { } + /// .sos for variables + SuffixIntWriter StartSOSVars(int nnz); + /// .sos for constraints + SuffixIntWriter StartSOSCons(int nnz); + /// .sosref for variables + SuffixDblWriter StartSOSREFVars(int nnz); + + private: + SuffixWriterFactory swf_; + }; + +protected: + /// Write string vector calling the given feeder. + /// @return max string length. + int WriteStringVec2File( + const std::string& name, std::function ); + + /// Writes nothing if !sz + template + void WriteSuffix(File& nm, + int kind, const char* name, + SparseVecFeeder&& sv); + + void WriteSparseEntry(File& nm, int i, int v); + void WriteSparseEntry(File& nm, int i, double x); + + void WriteBndRangeOrCompl(File& nm, + double L, double U, int k=0, int cvar=0); + + template + void WriteExprArgs(File& nm, ExprArgsFeeder&& args); + + /// Reuse Writer's facilities + using FormatterType::apr; + using FormatterType::nput; + using FormatterType::Utils; + + /// Retrieve Writer + FormatterType& Formatter() { return *this; } + /// Retrieve feeder + FeederType& Feeder() const { return feeder_; } + /// Retrieve header + const NLHeader& Hdr() const { return header_; } + + /// Infinity + double Infty() const; + /// Negative infinity + double NegInfty() const; + +private: + FeederType &feeder_; + NLHeader header_; + + int num_vars_and_exprs_; // Number of variables and common expressions. + + int maxLen_ColName_ {0}; + int maxLen_UnvName_ {0}; + int maxLen_FixName_ {0}; + + File nm; + + WriteNLReport result_ {WriteNL_Unset, ""}; +}; + +} // namespace mp + +#endif // MP_NLWriter22_H diff --git a/nl-writer2/include/mp/nl-writer2.hpp b/nl-writer2/include/mp/nl-writer2.hpp new file mode 100644 index 000000000..ef2f51875 --- /dev/null +++ b/nl-writer2/include/mp/nl-writer2.hpp @@ -0,0 +1,878 @@ +#ifndef NLWriter22_HPP +#define NLWriter22_HPP + +#include +#include + +#include "mp/nl-writer2.h" + +namespace mp { + +template +inline WriteNLReport WriteNLFile( + const std::string& namebase, + NLFeeder2& nlf, NLUtils& utl) { + auto h = nlf.Header(); + if (NLHeader::TEXT == h.format) { + mp::NLWriter2< + mp::NLWriter2Params< + mp::TextFormatter, NLFeeder2 > > + writer(nlf, h, utl); + return writer.WriteFiles(namebase); + } else { + mp::NLWriter2< + mp::NLWriter2Params< + mp::BinaryFormatter, NLFeeder2 > > + writer(nlf, h, utl); + return writer.WriteFiles(namebase); + } +} + +template +WriteNLReport NLWriter2::WriteFiles( + const std::string& namebase) { + WriteAuxFiles(namebase); + WriteNL(namebase); + return GetResult(); +} + +template +void NLWriter2::WriteAuxFiles( + const std::string& namebase) { + WriteRowObjNames(namebase); + WriteRowDelNames(namebase); + WriteColNames(namebase); + WriteUnusedVarNames(namebase); + WriteFixedVars(namebase); + header_.max_var_name_len = // maxLen_FixName_ can be longer + maxLen_ColName_ + maxLen_UnvName_ + maxLen_FixName_; + WriteObjAdj(namebase); +} + +template +void NLWriter2::WriteNL( + const std::string& namebase) { + auto fln = namebase + ".nl"; + nm = Utils().openf(fln, + Hdr().num_vars<=0, // Delete file if true + "wb"); + if (nm) { + WriteNLHeader(); + WriteFunctions(); + WriteSuffixes(); + WritePLSOSConstraints(); + if (Feeder().WantBoundsFirst()) { + WriteVarBounds(); + WriteInitialGuesses(); + WriteConBounds(); + WriteDualInitialGuesses(); + } + WriteDefinedVariables(0); + WriteConObjExpressions(); + if (!Feeder().WantBoundsFirst()) { + WriteDualInitialGuesses(); + WriteInitialGuesses(); + WriteConBounds(); + WriteVarBounds(); + } + if (Feeder().WantColumnSizes()) + WriteColumnSizes(); + WriteLinearConExpr(); + WriteObjGradients(); + WriteRandomVariables(); + + if (std::ferror(nm.GetHandle())) { + result_.first = WriteNL_Failed; + result_.second = fln + ": " + std::strerror(errno); + } else + result_.first = WriteNL_OK; + } else { + result_.first = WriteNL_CantOpen; + result_.second = fln + ": " + std::strerror(errno); + } +} + + +/// Write strings to a file. +/// The file is only opened when the user calls operator bool() +/// or upon 1st write. +/// If nothing written, the file is removed (if exists). +class StringFileWriter { +public: + /// bool: true if should remove the file + using TOpener = std::function; + /// Construct + StringFileWriter(int& lmax, TOpener opener) + : len_max_(lmax), opener_(opener) { } + /// No copy + StringFileWriter(const StringFileWriter& ) = delete; + /// Destroy + ~StringFileWriter() { + if (!cnt_ && !fTriedOpen_) // remove file if not opened + opener_(true); // or nothing written + } + /// No assign + void operator=(const StringFileWriter& ) = delete; + /// oper bool() + operator bool() { + return !fTriedOpen_ ? Open() : nm; + } + /// Write const char*, StrStrValue, StrDblValue + template + void Write(Value s) { + if (!fTriedOpen_) + Open(); + ++cnt_; + auto l = Print_ReturnLen1(s); + if (l>len_max_) + len_max_ = l; + } + /// Typedef StrStrValue + using StrStrValue = std::pair; + /// Typedef StrDblValue + using StrDblValue = std::pair; + +protected: + bool Open() { + fTriedOpen_ = true; + nm = opener_(false); + return nm; + } + int Print_ReturnLen1(const char* s) { + nm.Printf("%s\n", s); + return std::strlen(s); + } + int Print_ReturnLen1(StrStrValue p) { + nm.Printf("%s\t%s\n", p.first, p.second); + return std::strlen(p.first); + } + int Print_ReturnLen1(StrDblValue p) { + nm.Printf("%s\t%.17g\n", p.first, p.second); + return std::strlen(p.first); + } + +private: + int& len_max_; + TOpener opener_; + File nm; + bool fTriedOpen_{false}; + int cnt_{0}; +}; + +/// operator<< (StringFileWriter&, Value) +template +StringFileWriter& +operator << (StringFileWriter& wrt, Value s) { + wrt.Write(s); + return wrt; +} + +template +void NLWriter2::WriteRowObjNames( + const std::string& namebase) { + header_.max_con_name_len = + WriteStringVec2File(namebase + ".row", + [this](StringFileWriter& w){ + feeder_.FeedRowAndObjNames(w); }); +} + +template +void NLWriter2::WriteRowDelNames( + const std::string& namebase) { + WriteStringVec2File(namebase + ".slc", + [this](StringFileWriter& w){ + feeder_.FeedDelRowNames(w); }); +} + +template +void NLWriter2::WriteColNames( + const std::string& namebase) { + maxLen_ColName_ = + WriteStringVec2File(namebase + ".col", + [this](StringFileWriter& w){ + feeder_.FeedColNames(w); }); +} + +template +void NLWriter2::WriteUnusedVarNames( + const std::string& namebase) { + maxLen_UnvName_ = + WriteStringVec2File(namebase + ".unv", + [this](StringFileWriter& w){ + feeder_.FeedUnusedVarNames(w); +}); +} + +template +void NLWriter2::WriteFixedVars( + const std::string& namebase) { + maxLen_FixName_ = + WriteStringVec2File(namebase + ".fix", + [this](StringFileWriter& w){ + feeder_.FeedFixedVarNames(w); +}); +} + +template +void NLWriter2::WriteObjAdj( + const std::string& namebase) { + WriteStringVec2File(namebase + ".adj", + [this](StringFileWriter& w){ + feeder_.FeedObjAdj(w); + }); +} + +template +int NLWriter2::WriteStringVec2File( + const std::string& name, std::function swf) { + int lMax=0; + StringFileWriter sw(lMax, + [name, this](bool fClose){ + return Utils().openf(name.c_str(), + fClose, "w"); + }); + swf(sw); + return lMax; +} + + +/* song and dance because of the interaction of cfront's treatment + of adjacent string literals, lcc's treatment of multi-line strings, + and MSDOS's stupid \r\n conventions.... + */ +#ifdef X_AMPL +#if defined(_WIN32) || defined(_WIN64) +#define EOL "\r\n" +#else +#define EOL "\n" +#endif /* _WIN32 && _WIN64 */ +#else /* X_AMPL */ +#ifdef MSDOS +#define EOL "\r\n" +#else +#define EOL "\n" +#endif +#endif /* X_AMPL */ + +#ifdef NL_LIB2_ORIG_HDR +static constexpr char +gl_1__short[] = "%c%d", +gl_2a[] = " %d %d %d %d %d", +gl_1a[] = "\t# problem %s" EOL, +gl_2[] = "\t# vars, constraints, objectives, ranges, eqns%s" EOL, +gl_3[] = " %d %d\t# nonlinear constraints, objectives" EOL, +gl_3c[] = "\t# nonlinear constrs, objs; ccons: lin, nonlin, nd, nzlb%s" EOL, +gl_4[] = "\t# network constraints: nonlinear, linear" EOL, +gl_4r[] = " %d\t# network constraints: nonlinear, linear; stages" EOL, +gl_5[] = " %d %d %d\t# nonlinear vars in constraints, objectives, both" EOL, +gl_6[] = " %d %d\t# linear network variables; functions" EOL, +gl_6x[] = " %d %d %d %d\t# linear network variables; functions; arith, flags" EOL, +gl_6y[] = " %d %d %d %d %d\t# linear network variables; functions; arith, flags; randcalls" EOL, +gl_7[] = " %d %d %d %d %d\t# discrete variables: binary, integer, nonlinear (b,c,o)" EOL, +gl_8[] = " %zd %zd\t# nonzeros in Jacobian, gradients" EOL, +gl_9[] = " %d %d\t# max name lengths: constraints, variables" EOL, +gl_10[] = "\t# common exprs: b,c,o,c1,o1%s" EOL; +#else +static constexpr char +gl_1__short[] = "%c%d", +gl_2a[] = " %d %d %d %d %d", +gl_1a[] = "\t# problem %s" EOL, +gl_2[] = "\t# vars, algcons, objs, ranges, eqns%s" EOL, +gl_3[] = " %d %d\t# nonlinear cons, objs" EOL, +gl_3c[] = "\t# nonlinear cons, objs; compl: lin, nonlin, range, nzlb%s" EOL, +gl_4[] = "\t# network cons: nonlinear, linear" EOL, +gl_4r[] = " %d\t# network cons: nonlinear, linear; stages" EOL, +gl_5[] = " %d %d %d\t# nonlinear vars in cons, objs, both" EOL, +gl_6[] = " %d %d\t# linear network vars; funcs" EOL, +gl_6x[] = " %d %d %d %d\t# linear network vars; funcs; arith, flags" EOL, +gl_6y[] = " %d %d %d %d %d\t# linear network vars; funcs; arith, flags; randcalls" EOL, +gl_7[] = " %d %d %d %d %d\t# discrete vars: binary, integer, nonlinear (b,c,o)" EOL, +gl_8[] = " %zd %zd\t# nonzeros in Jacobian, gradients" EOL, +gl_9[] = " %d %d\t# max name lengths: cons/objs, vars" EOL, +gl_10[] = "\t# common exprs: b,c,o,c1,o1%s" EOL; +#endif +#undef EOL + + +template +void NLWriter2::WriteNLHeader() { + assert(Formatter().Mode() == Hdr().format); + + /// But the header is always text. + nm.Printf(gl_1__short, + NLHeader::TEXT==Hdr().format ? 'g' : 'b', + Hdr().num_ampl_options); + for (int i = 0; i < Hdr().num_ampl_options; ++i) + nm.Printf(" %ld", Hdr().ampl_options[i]); + if (Hdr().ampl_options[VBTOL_OPTION_INDEX] == USE_VBTOL_FLAG) + nm.Printf(" %.g", Hdr().ampl_vbtol); + nm.Printf(gl_1a, Hdr().prob_name.c_str()); + + /// Num variables, constraints, obj, ... + nm.Printf(gl_2a, + Hdr().num_vars, + Hdr().num_algebraic_cons, + Hdr().num_objs, + Hdr().num_ranges, + Hdr().num_eqns); + const char* s = ""; + if (Hdr().num_rand_vars) { /*SNL2006*/ + s = ", lcons, randvars"; + nm.Printf(" %d %d", + Hdr().num_logical_cons, + Hdr().num_rand_vars); + } + else if (Hdr().num_logical_cons) { + s = ", lcons"; + nm.Printf(" %d", Hdr().num_logical_cons); + } + nm.Printf(gl_2, s); + + /// Nonlinear / random cons/obj + if (Hdr().num_compl_conds | + Hdr().num_rand_cons | Hdr().num_rand_objs) { + nm.Printf(" %d %d %d %d %d %d", + Hdr().num_nl_cons, Hdr().num_nl_objs, + Hdr().num_compl_conds - Hdr().num_nl_compl_conds, + Hdr().num_nl_compl_conds, Hdr().num_compl_dbl_ineqs, + Hdr().num_compl_vars_with_nz_lb); + s = ""; + if (Hdr().num_rand_cons | Hdr().num_rand_objs) { + nm.Printf(" %d %d", + Hdr().num_rand_cons, Hdr().num_rand_objs); + s = "; rand constrs, objs"; + } + nm.Printf(gl_3c, s); + } + else + nm.Printf(gl_3, Hdr().num_nl_cons, Hdr().num_nl_objs); + + /// Network cons, stages + nm.Printf(" %d %d", + Hdr().num_nl_net_cons, Hdr().num_linear_net_cons); + nm.Printf(Hdr().num_stages > 1 ? gl_4r : gl_4, + Hdr().num_stages); + + /// NL vars in cons, objs, both + nm.Printf(gl_5, Hdr().num_nl_vars_in_cons, + Hdr().num_nl_vars_in_objs, Hdr().num_nl_vars_in_both); + + /// Linear network vars, functions, ak, flags, rnd calls + const char* fmt = + Hdr().num_rand_vars ? gl_6y : + Hdr().flags | Hdr().arith_kind ? + gl_6x : gl_6; + nm.Printf(fmt, Hdr().num_linear_net_vars, Hdr().num_funcs, + NLHeader::TEXT==Hdr().format ? 0 : Hdr().arith_kind, + Hdr().flags, Hdr().num_rand_calls); + + /// Integer vars + nm.Printf(gl_7, Hdr().num_linear_binary_vars, + Hdr().num_linear_integer_vars, + Hdr().num_nl_integer_vars_in_both, + Hdr().num_nl_integer_vars_in_cons, + Hdr().num_nl_integer_vars_in_objs); + + /// Nonzeros + nm.Printf(gl_8, Hdr().num_con_nonzeros, + Hdr().num_obj_nonzeros); + + /// Max name lengths + nm.Printf(gl_9, Hdr().max_con_name_len, + Hdr().max_var_name_len); + + /// Common exprs + nm.Printf(" %d %d %d %d %d", + Hdr().num_common_exprs_in_both, + Hdr().num_common_exprs_in_cons, + Hdr().num_common_exprs_in_objs, + Hdr().num_common_exprs_in_single_cons, + Hdr().num_common_exprs_in_single_objs); + s = ""; + if (Hdr().num_rand_common_exprs) { /*SNL2006*/ + nm.Printf(" %d", Hdr().num_rand_common_exprs); + s = ",rand"; + } + nm.Printf(gl_10, s); +} + +template +void NLWriter2::WriteFunctions() { + for (int i=0; i +void NLWriter2:: +WriteSparseEntry(File& nm, int i, int v) { + apr(nm, "%d %d\n", i, v); +} + +template +void NLWriter2:: +WriteSparseEntry(File& nm, int i, double x) { + apr(nm, "%d %g\n", i, x); +} + +template +void NLWriter2::WriteSuffixes() { + SuffixWriterFactory swf(*this); + Feeder().FeedSuffixes(swf); +} + +template +void NLWriter2::WritePLSOSConstraints() { + PLSOSWriter plsos(*this); + Feeder().FeedPLSOS(plsos); +} + +template +void NLWriter2::WriteVarBounds() { + apr(nm, "b\t# %d bounds (on variables)\n", + (int)Hdr().num_vars); + VarBndWriter vbw(*this); + Feeder().FeedVarBounds(vbw); + assert((int)Hdr().num_vars == vbw.GetNWritten()); +} + +template +void NLWriter2::WriteInitialGuesses() { + SingleSparseDblVecWrtFactory + vwf(*this, "x%d\t# initial guess\n"); + Feeder().FeedInitialGuesses(vwf); +} + +template +void NLWriter2::WriteConBounds() { + if (Hdr().num_algebraic_cons) { + apr(nm, "r\t# %d ranges (rhs's) and complementarity\n", + (int)Hdr().num_algebraic_cons); + ConBndWriter cbw(*this); + Feeder().FeedConBounds(cbw); + assert((int)Hdr().num_algebraic_cons == cbw.GetNWritten()); + } +} + +template +void NLWriter2::WriteDualInitialGuesses() { + SingleSparseDblVecWrtFactory + vwf(*this, "d%d\t# initial dual guess\n"); + Feeder().FeedInitialDualGuesses(vwf); +} + +template +void NLWriter2::WriteBndRangeOrCompl( + File& nm, + double L, double U, int k, int cvar) { + if (k<=0) { // normal algebraic constraint + if (L <= NegInfty()) + apr(nm, U >= Infty() + ? "3\n" : "1 %.16g\n", U); + else + apr(nm, U >= Infty() ? "2 %.16g\n" + : L == U ? "4 %.16g\n" + : "0 %.16g %.16g\n", + L, U); + } else { // complementarity + apr(nm, "5 %d %d\n", k, cvar + 1); // add +1 here + } +} + + +template +template +void NLWriter2::WriteExprArgs(File& nm, + ExprArgsFeeder&& args) { + while (args) { + auto e = args.Next(); + WriteExpression(nm, e); + } +} + +template +void NLWriter2::WriteDefinedVariables(int i) { + DefVarWriterFactory dvwf(*this, i); + Feeder().FeedDefinedVariables(i, dvwf); +} + +template +void NLWriter2::WriteConObjExpressions() { + int i=0; + for (; i +void NLWriter2::WriteColumnSizes() { + switch(Feeder().WantColumnSizes()) { + case 1: + apr(nm, + "k%d\t#intermediate Jacobian column lengths (cumulative)\n", + Hdr().num_vars + Hdr().num_rand_vars - 1); + { + ColSizeWriter csw(*this, 1); + Feeder().FeedColumnSizes(csw); + assert(Hdr().num_vars + Hdr().num_rand_vars - 1 + == csw.GetNWritten()); + } + break; + case 2: + apr(nm, "K%d\t#intermediate Jacobian column lengths\n", + Hdr().num_vars + Hdr().num_rand_vars - 1); + { + ColSizeWriter csw(*this, 2); + Feeder().FeedColumnSizes(csw); + assert(Hdr().num_vars + Hdr().num_rand_vars - 1 + == csw.GetNWritten()); + } + break; + case 0: + break; // SKIP + default: + assert(0 && "why am I here?"); + } +} + +template +void NLWriter2::WriteLinearConExpr() { + for (int i=0; iapr(this->nm, "J%d %d\n", i, nnz); + }); + Feeder().FeedLinearConExpr(i, vwf); + } +} + +template +void NLWriter2::WriteObjGradients() { + for (int i=0; iapr(this->nm, "G%d %d\n", i, nnz); + }); + Feeder().FeedObjGradient(i, vwf); + } +} + +template +void NLWriter2::WriteRandomVariables() { + RandVarWriterFactory rvwf(*this); + Feeder().FeedRandomVariables(rvwf); +} + + +///////////////////// SparseVectorWriter ////////////////// +template +template +NLWriter2::SparseVectorWriter:: +SparseVectorWriter(NLWriter2& nlw, size_t n) + : nlw_(nlw), n_entries_(n) { } + +template +template +void NLWriter2::SparseVectorWriter:: +Write(Index i, Value v) { + --n_entries_; + nlw_.WriteSparseEntry(nlw_.nm, i, v); +} + + +///////////////////// ExprArgWriter /////////////////////// +template +NLWriter2::ExprArgWriter:: +ExprArgWriter(NLWriter2& nlw, int na) + : nlw_(nlw), nargs_(na) { assert(nargs_>0); } + +template +NLWriter2::ExprArgWriter:: +~ExprArgWriter() +{ assert(0==nargs_); } // Check all written. + +template +typename NLWriter2::ExprWriter +NLWriter2::ExprArgWriter::AddArg() { + --nargs_; + return {nlw_}; +} + +template +void NLWriter2::ExprArgWriter::EPut( + typename FeederType::Expr e) +{ AddArg().EPut(e); } + +template +void NLWriter2::ExprArgWriter::VPut( + int v, const char* descr) +{ AddArg().VPut(v, descr); } + +template +void NLWriter2::ExprArgWriter::NPut( + double x) { AddArg().NPut(x); } + +template +void NLWriter2::ExprArgWriter::StrPut( + const char* s) { AddArg().StrPut(s); } + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprArgWriter::FuncPut( + int index, int nArgs, const char* descr) +{ return AddArg().FuncPut(index, nArgs, descr); } + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprArgWriter::OPut1( + int opcode, const char* descr) +{ return AddArg().OPut1(opcode, descr); } + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprArgWriter::OPut2( + int opcode, const char* descr) +{ return AddArg().OPut2(opcode, descr); } + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprArgWriter::OPut3( + int opcode, const char* descr) +{ return AddArg().OPut3(opcode, descr); } + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprArgWriter::OPutN( + int opcode, int nArgs, const char* descr) +{ return AddArg().OPutN(opcode, nArgs, descr); } + + +///////////////////// ExprWriter /////////////////////// +template +void NLWriter2::ExprWriter::EPut( + typename FeederType::Expr e) { + ExprArgWriter ew(nlw_, 1); + nlw_.Feeder().FeedExpr(e, ew); +} + +template +void NLWriter2::ExprWriter::VPut( + int v, const char* descr) { + nlw_.apr(nlw_.nm, + "v%d\t#%s\n", v, descr); +} + +template +void NLWriter2::ExprWriter::NPut( + double x) { nlw_.nput(nlw_.nm, x); } + +template +void NLWriter2::ExprWriter::StrPut( + const char* s) { + nlw_.apr(nlw_.nm, + "h%d:%s\n", (int)std::strlen(s), s); +} + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprWriter::FuncPut( + int index, int nArgs, const char* descr) { + nlw_.apr(nlw_.nm, + "f%d %d\t#%s\n", index, nArgs, descr); + return ExprArgWriter(nlw_, nArgs); +} + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprWriter::OPut1( + int opcode, const char* descr) { + nlw_.apr(nlw_.nm, + "o%d\t#%s\n", opcode, descr); + return ExprArgWriter(nlw_, 1); +} + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprWriter::OPut2( + int opcode, const char* descr) { + nlw_.apr(nlw_.nm, + "o%d\t#%s\n", opcode, descr); + return ExprArgWriter(nlw_, 2); +} + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprWriter::OPut3( + int opcode, const char* descr) { + nlw_.apr(nlw_.nm, + "o%d\t#%s\n", opcode, descr); + return ExprArgWriter(nlw_, 3); +} + +template +typename NLWriter2::ExprArgWriter +NLWriter2::ExprWriter::OPutN( + int opcode, int nArgs, const char* descr) { + nlw_.apr(nlw_.nm, + "o%d\t#%s\n", opcode, descr); + int n2write = nArgs; + if (64 == opcode) { // piecewise-linear + assert(0 == n2write % 2); + n2write /= 2; + } + nlw_.apr(nlw_.nm, "%d\n", n2write); + return ExprArgWriter(nlw_, nArgs); +} + + +///////////////////// DEFINED VARIABLES /////////////////// +template +NLWriter2::DefVarWriter:: +DefVarWriter(NLWriter2& nlw, int nnzlin) + : nnzlin_(nnzlin), nlw_(nlw) { } + +template +typename NLWriter2::SparseDblVecWriter +NLWriter2::DefVarWriter::GetLinExprWriter() +{ return SparseDblVecWriter(nlw_, nnzlin_); } + +template +typename NLWriter2::ExprArgWriter +NLWriter2::DefVarWriter::GetExprWriter() { + // Check that linw_ used? Still might not be written! + return ExprArgWriter(nlw_, 1); +} + +template +typename NLWriter2::DefVarWriter +NLWriter2::DefVarWriterFactory:: +StartDefVar(int index, int nnz, const char* descr) { + nlw_.apr(nlw_.nm, + "V%d %d %d\t#%s\n", + index, nnz, + k_>=0 ? k_ + : nlw_.Hdr().num_algebraic_cons + + nlw_.Hdr().num_logical_cons - k_, + descr); + return DefVarWriter(nlw_, nnz); +} + + +/////////////////////// BOUNDS //////////////////////////// +template +void +NLWriter2::VarBndWriter::WriteLbUb( + double lb, double ub) { + nlw_.WriteBndRangeOrCompl(nlw_.nm, lb, ub); + ++nWrt_; +} + +template +template +void +NLWriter2::ConBndWriter::WriteAlgConRange( + AlgConRange acr) { + nlw_.WriteBndRangeOrCompl(nlw_.nm, + acr.L, acr.U, acr.k, acr.cvar); + ++nWrt_; +} + + +////////////////////// SUFFIXES /////////////////////////// +template +typename NLWriter2::SuffixIntWriter +NLWriter2::SuffixWriterFactory::StartIntSuffix( + const char* name, int kind, int nnz) { + assert(0 == (kind & 4)); + if (nnz) + nlw_.apr(nlw_.nm, "S%d %d %s\n", kind, nnz, name); + return SuffixIntWriter(nlw_, nnz); +} + +template +typename NLWriter2::SuffixDblWriter +NLWriter2::SuffixWriterFactory::StartDblSuffix( + const char* name, int kind, int nnz) { + assert(0 != (kind & 4)); + if (nnz) + nlw_.apr(nlw_.nm, "S%d %d %s\n", kind, nnz, name); + return SuffixDblWriter(nlw_, nnz); +} + + +////////////////////// PL-SOS SUFFIXES ///////////// +template +typename NLWriter2::SuffixIntWriter +NLWriter2::PLSOSWriter::StartSOSVars(int nnz) +{ + return swf_.StartIntSuffix("sos", 0, nnz); +} + +template +typename NLWriter2::SuffixIntWriter +NLWriter2::PLSOSWriter::StartSOSCons(int nnz) +{ + return swf_.StartIntSuffix("sos", 1, nnz); +} + +template +typename NLWriter2::SuffixDblWriter +NLWriter2::PLSOSWriter::StartSOSREFVars(int nnz) +{ + return swf_.StartDblSuffix("sosref", 4, nnz); +} + + +////////////////////// RANDOM VARS ///////////////// +template +typename NLWriter2::ExprArgWriter +NLWriter2::RandVarWriterFactory:: +StartRandVar(int index, const char* descr) { + nlw_.apr(nlw_.nm, + "R%d\t# %s\n", index, descr); + return ExprArgWriter(nlw_, 1); +} + + +////////////////////// MATH //////////////////////// +template +double NLWriter2::Infty() const { + return std::numeric_limits::max(); +} + +template +double NLWriter2::NegInfty() const { + return -Infty(); +} + +} // namespace mp + +#endif // NLWriter22_HPP diff --git a/nl-writer2/include/mp/nlsol.h b/nl-writer2/include/mp/nlsol.h new file mode 100644 index 000000000..b4f7df468 --- /dev/null +++ b/nl-writer2/include/mp/nlsol.h @@ -0,0 +1,189 @@ +/** + mp::NLSOL. + Manager for solving optimization models via NL files. + It performs zero-overhead model/solution transmission. + In particular, it does not store any intermediate + model representation. + + NL is a format for representing optimization problems such as linear, + quadratic, nonlinear, complementarity and constraint programming problems + in discrete or continuous variables. It is described in the technical report + "Writing .nl Files" (http://www.cs.sandia.gov/~dmgay/nlwrite.pdf). + + + Copyright (C) 2023 AMPL Optimization, Inc. + + Permission to use, copy, modify, and distribute this software and its + documentation for any purpose and without fee is hereby granted, + provided that the above copyright notice appear in all copies and that + both that the copyright notice and this permission notice and warranty + disclaimer appear in supporting documentation. + + The author and AMPL Optimization Inc. disclaim all warranties with + regard to this software, including all implied warranties of + merchantability and fitness. In no event shall the author be liable + for any special, indirect or consequential damages or any damages + whatsoever resulting from loss of use, data or profits, whether in an + action of contract, negligence or other tortious action, arising out + of or in connection with the use or performance of this software. + + Author: Gleb Belov + */ +#ifndef NLSOL_H +#define NLSOL_H + +#include +#include +#include + +#include "mp/nl-writer2.h" +#include "mp/sol-reader2.h" + +namespace mp { + +/// Class NLSOL. +/// +/// Manager for solving optimization models via NL files. +/// It performs zero-overhead model/solution transmission. +/// In particular, it does not store any intermediate +/// model representation. +/// +/// Usage: +/// +/// MyNLFeeder feeder; +/// MySOLHandler handler; +/// mp::NLUtils nlutils; +/// mp::NLSOL +/// nlsol(feeder, handler, nlutils); +/// nlsol.SetSolver("minos"); +/// nlsol.SetSolverOptions("outlev=1"); +/// if (!nlsol.Solve("filestub")) { +/// printf("%s\n", nlsol.GetErrorMessage()); +/// } +/// +/// If separate steps are needed, use a subset of +/// +/// if (!nlsol.WriteNLFile(filestub) +/// || !nlsol.InvokeSolver(filestub) +/// || !nlsol.ReadSolution(filestub + ".sol")) ... +/// +/// See mp::NLFeeder2 and mp::SOLHandler2 interfaces +/// for model/solution transmission. +/// +/// @param MyNLFeeder: a class implementing +/// the mp::NLFeeder2 interface. +/// @param MySOLHandler: a class implementing +/// the mp::SOLHandler2 interface. +template +class NLSOL { +public: + /// Construct + NLSOL(MyNLFeeder& fd, MySOLHandler& sh, mp::NLUtils& ut) + : feeder_(fd), handler_(sh), utils_(ut) { } + + /// NLFeederType + using NLFeederType = MyNLFeeder; + /// SOLHandlerType + using SOLHandlerType = MySOLHandler; + + /// Set solver, such as "gurobi", "highs", "ipopt" + void SetSolver(std::string solver); + + /// Set solver options, such as "outlev=1 lim:time=500" + void SetSolverOptions(std::string sopts); + + /// Solve. + /// @param filestub: filename stub to be used + /// for input files (.nl, .col., .row, etc.), + /// and output files (.sol). + /// @return true if all ok. + bool Solve(const std::string& filestub); + + /// Get error message. + const char* GetErrorMessage() const { return err_msg_.c_str(); } + + /// Substep: write NL and any accompanying files. + bool WriteNLFile(const std::string& filestub); + + /// Substep: invoke chosen solver for \a filestub. + bool InvokeSolver(const std::string& filestub); + + /// Substep: read solution. + /// @param filename: complete file name, + /// normally (stub).sol. + bool ReadSolution(const std::string& filename); + + +private: + MyNLFeeder& feeder_; + MySOLHandler& handler_; + mp::NLUtils& utils_; + + std::string solver_; + std::string solver_options_; + + std::string err_msg_; +}; + + +template +void NLSOL:: +SetSolver(std::string solver) { solver_ = std::move(solver); } + +template +void NLSOL:: +SetSolverOptions(std::string sopts) +{ solver_options_ = std::move(sopts); } + +template +bool NLSOL:: +Solve(const std::string& filestub) { + return (WriteNLFile(filestub) + && InvokeSolver(filestub) + && ReadSolution(filestub + ".sol")); +} + +template +bool NLSOL:: +WriteNLFile(const std::string& filestub) { + if (filestub.empty()) + return (err_msg_="WriteNL error: provide filestub.", false); + auto result = mp::WriteNLFile(filestub, feeder_, utils_); + if (mp::WriteNL_OK != result.first) + return (err_msg_ = "WriteNL error: " + result.second, false); + return true; +} + +template +bool NLSOL:: +InvokeSolver(const std::string& filestub) { + if (filestub.empty()) + return (err_msg_="NLSOL: provide filestub.", false); + if (solver_.empty()) + return (err_msg_="NLSOL: provide solver.", false); + auto call = solver_ + + ' ' + filestub + + " -AMPL " + + solver_options_; + if (auto status = std::system(call.c_str())) + return (err_msg_="NLSOL: call \"" + + call + "\" failed (code " + + std::to_string(status) + ").", false); + return true; +} + +template +bool NLSOL:: +ReadSolution(const std::string& filename) { + if (filename.empty()) + return (err_msg_="SOLReader: provide filename.", false); + auto status = mp::ReadSOLFile( + filename, handler_, utils_); + if (mp::SOL_Read_OK != status.first) + return (err_msg_="SOLReader error: "+status.second, false); + return true; +} + +} // namespace mp + +#endif // NLSOL_H diff --git a/nl-writer2/include/mp/sol-handler2.h b/nl-writer2/include/mp/sol-handler2.h new file mode 100644 index 000000000..56302f632 --- /dev/null +++ b/nl-writer2/include/mp/sol-handler2.h @@ -0,0 +1,176 @@ +/** + mp::SOLHandler2. + Interface for a solution handler to be used with mp::SOLReader2. + mp::SOLReader2 is a zero-overhead reader + implemented with inline template code. + + AMPL SOL is a format for representing solutions of optimization problems + such as linear, quadratic, nonlinear, complementarity + and constraint programming problems in discrete or continuous variables, + used by AMPL. It is not documented yet. + + Usage: recommended via mp::NLSOL class. + + If you need to read solutions separately, proceed as follows: + + MySOLHandler handler; + mp::NLUtils utils; + auto status = mp::ReadSOLFile( + "filestub", handler, utils); + if (mp::SOL_Read_OK != status.first) { + printf(status.second.c_str()); + } + + where handler is an object that receives information on solution + components. Below is an interface for such a handler. + + Copyright (C) 2023 AMPL Optimization, Inc. + + Permission to use, copy, modify, and distribute this software and its + documentation for any purpose and without fee is hereby granted, + provided that the above copyright notice appear in all copies and that + both that the copyright notice and this permission notice and warranty + disclaimer appear in supporting documentation. + + The author and AMPL Optimization Inc. disclaim all warranties with + regard to this software, including all implied warranties of + merchantability and fitness. In no event shall the author be liable + for any special, indirect or consequential damages or any damages + whatsoever resulting from loss of use, data or profits, whether in an + action of contract, negligence or other tortious action, arising out + of or in connection with the use or performance of this software. + + Author: Gleb Belov + */ + +#ifndef SOLHANDLER2_H +#define SOLHANDLER2_H + +#include +#include + +#include "mp/nl-header.h" + +namespace mp { + +/** + \rst + SOLHandler2: reads solution details on request + via provided callback objects. + + `~mp::SOLHandler2` can be used as a base class for other handlers, + or just be an interface prototype. + + \endrst + */ +class SOLHandler2 { +public: + /** Destroys the object. */ + virtual ~SOLHandler2() { } + + /** The NLHeader used to write the NL file. */ + NLHeader Header() const { return {}; } + + /** Receive solve message. + * The message always ends with '\n'. + * + * @param nbs: number of backspaces + * in the original solve message. + * So many characters should be skipped + * from the message if printed straightaway. + * AMPL solver drivers can supply the message + * with initial backspaces to indicate + * that so many characters should be skipped + * when printing. For example, if the driver prints + * MINOS 5.51: + * and exits, and the message starts with that again, + * this part should be skipped. + */ + void OnSolveMessage(const char* s, int nbs) { } + + /** + * AMPL internal options. + */ + struct AMPLOptions { + std::vector options_; + bool has_vbtol_; + double vbtol_; + }; + + /** + * Can be ignored by external systems. + * @return non-zero to stop solution input. + */ + int OnAMPLOptions(const AMPLOptions& ) { return 0; } + + /** + * Dual values for algebraic constraints, + * if provided in the solution. + * Number of values <= NumAlgCons(). + * Implementation: + * + * duals.reserve(rd.Size()); + * while (rd.Size()) + * duals.push_back(rd.ReadNext()); + */ + template + void OnDualSolution(VecReader& ) { } + + /** + * Variable values, if provided. + * Number of values <= NumVars(). + */ + template + void OnPrimalSolution(VecReader& ) { } + + /** + * Receive notification of the objective index + * used by the driver (solver option 'objno'). + */ + void OnObjno(int ) { } + + /** + * Receive notification of the solve code. + */ + void OnSolveCode(int ) { } + + /** + * OnIntSuffix(). + * + * For constraints, can include values for + * logical constraints (after algebraic.) + * Sparse representation - can be empty + * (i.e., all values zero.) + * + * const auto& si = sr.SufInfo(); + * int kind = si.Kind(); + * int nmax = nitems_max[kind & 3]; + * const std::string& name = si.Name(); + * const std::string& table = si.Table(); + * while (sr.Size()) { + * std::pair val = sr.ReadNext(); + * if (val.first<0 || val.first>=nmax) { + * sr.SetError(mp::SOL_Read_Bad_Suffix, + * "bad suffix element index"); + * return; + * } + * suf[val.first] = val.second; + * } + * if (mp::SOL_Read_OK == sr.ReadResult()) // Can check + * RegisterSuffix(kind, name, table, suf); + */ + template + void OnIntSuffix(SuffixReader& ) { } + + /** + * Same as OnIntSuffix(), but + * sr.ReadNext() returns pair + */ + template + void OnDblSuffix(SuffixReader& ) { } + +}; + +} // namespace mp + +#endif // SOLHANDLER2_H diff --git a/nl-writer2/include/mp/sol-reader2.h b/nl-writer2/include/mp/sol-reader2.h new file mode 100644 index 000000000..26cf5b918 --- /dev/null +++ b/nl-writer2/include/mp/sol-reader2.h @@ -0,0 +1,324 @@ +/** + SOL reader2. + + SOL is a format for representing solutions of optimization problems + such as linear, quadratic, nonlinear, complementarity + and constraint programming problems in discrete or continuous variables, + used by AMPL. It is not documented yet. + + Usage: + MySOLHandler handler; + mp::NLUtils utils; + + auto status = ReadSOLFile(name, handler, utils); + if (status.first) { + printf("Error: %s\n", status.second.c_str()); + exit(EXIT_FAILURE); + } + + where handler is an object that receives information on solution + components. + + Copyright (C) 2023 AMPL Optimization Inc. + + Permission to use, copy, modify, and distribute this software and its + documentation for any purpose and without fee is hereby granted, + provided that the above copyright notice appear in all copies and that + both that the copyright notice and this permission notice and warranty + disclaimer appear in supporting documentation. + + The author and AMPL Optimization Inc disclaim all warranties with + regard to this software, including all implied warranties of + merchantability and fitness. In no event shall the author be liable + for any special, indirect or consequential damages or any damages + whatsoever resulting from loss of use, data or profits, whether in an + action of contract, negligence or other tortious action, arising out + of or in connection with the use or performance of this software. + + Author: Gleb Belov + */ + +#ifndef SOLREADER2_H +#define SOLREADER2_H + +#include +#include +#include +#include + +#include "mp/sol-handler2.h" +#include "mp/nl-utils2.h" + +namespace mp { + +/// SOL read result +enum SOLReadResult { + SOL_Read_Result_Not_Set = -1, + SOL_Read_OK = 0, + SOL_Read_Fail_Open, + SOL_Read_Early_EOF, + SOL_Read_Bad_Format, + SOL_Read_Bad_Line, + SOL_Read_Bad_Options, + SOL_Read_Vector_Not_Finished, + SOL_Read_Bad_Suffix +}; + + +/// Vector reader +template +class VecReader { +public: + /// Construct + VecReader(FILE* f, int b, int n) + : f_(f), binary_(b), n_(n) { } + + /// Destruct + ~VecReader() { assert(!n_); } + + /// Typedef value_type + using value_type = Value; + + /// Size remaining to read + int Size() const { return n_; } + + /// Read next value + Value ReadNext(); + + /// Set error status + void SetError(SOLReadResult res, std::string msg) + { rr_ = res; err_msg_ = std::move(msg); n_ = 0; } + + /// status + SOLReadResult ReadResult() const { return rr_; } + + /// error message + const std::string& ErrorMessage() const + { return err_msg_; } + + +private: + FILE* f_; + int binary_; + int n_; + + SOLReadResult rr_{SOL_Read_OK}; + std::string err_msg_; +}; + + +/// Suffix info +class SuffixInfo { +public: + /// Construct + SuffixInfo(int k, std::string n, + std::string t) + : kind_(k), name_(std::move(n)), + table_(std::move(t)) { } + /// Suffix kind + int Kind() const { return kind_; } + /// Suffix name + const std::string& Name() const { return name_; } + /// Suffix table + const std::string& Table() const { return table_; } + +private: + int kind_; + std::string name_, table_; +}; + + +/// Typedef SparseVecReader +template +using SparseVecReader + = VecReader< std::pair >; + + +/// Class SuffixReader<> +template +class SuffixReader + : public SparseVecReader { +public: + /// Construct + SuffixReader(SuffixInfo si, FILE* f, int b, int n) + : SparseVecReader(f, b, n), + si_(std::move(si)) { } + /// Suffix info + const SuffixInfo& SufInfo() const { return si_; } + +private: + SuffixInfo si_; +}; + +/// Declare struct SufRead +struct SufRead; + + +/// SOL reader. +/// +/// AMPL SOL is a format for representing solutions of optimization problems +/// such as linear, quadratic, nonlinear, complementarity +/// and constraint programming problems in discrete or continuous variables, +/// used by AMPL. It is not documented yet. +/// +/// Usage: recommended via mp::NLSOL class. +/// If you need to read solutions separately, proceed as follows: +/// +/// MySOLHandler handler; +/// mp::NLUtils utils; +/// +/// auto status = ReadSOLFile(name, handler, utils); +/// if (status.first) { +/// printf("Error: %s\n", status.second.c_str()); +/// exit(EXIT_FAILURE); +/// } +/// +/// @param SOLHandler2: class that receives information on solution +/// components. +template +class SOLReader2 { +public: + /// Destructor. + virtual ~SOLReader2() { } + + /// Constructor. + SOLReader2(SOLHandler2& sh, NLUtils& utl) : + solh_(sh), utils_(utl), hdr_(sh.Header()) { } + + /// Read .sol file. + SOLReadResult ReadSOLFile(const std::string& name); + + /// Error message, if any. + /// Warnings are printed independently + /// via NLUtils::log_warning(). + const std::string& ErrorMessage(SOLReadResult ) const + { return err_msg_; } + + /// Internal AMPL result code + int internal_rv() const { return internal_rv_; } + + +protected: + /// Get const handler& + const SOLHandler2& Handler() const { return solh_; } + /// Get handler& + SOLHandler2& Handler() { return solh_; } + + /// Get const utils& + const NLUtils& Utils() const { return utils_; } + /// Get utils& + NLUtils& Utils() { return utils_; } + + /// Get const NLHeader& + const NLHeader& Header() const { return hdr_; } + + /// NumVars + int NumVars() const { return Header().num_vars; } + + /// NumAlgCons + int NumAlgCons() const { return Header().num_algebraic_cons; } + + /// NumLogCons + int NumLogCons() const { return Header().num_logical_cons; } + + /// Check result of a [vector] reader + template + bool CheckReader(const Reader& rd, SOLReadResult& rr) { + if (SOL_Read_Early_EOF == rd.ReadResult()) + return (rr=ReportEarlyEof(), false); + if (SOL_Read_Bad_Line == rd.ReadResult()) + return (rr=ReportBadLine(rd.ErrorMessage()), false); + if (rd.Size()) { // unfinished + serror("vector not read completely"); + return (rr=ReportBadFormat(), false); + } + if (SOL_Read_OK != rd.ReadResult()) { + serror( rd.ErrorMessage().c_str() ); + ReportBadFormat(); + return (rr=rd.ReadResult(), false); + } + return true; + } + + /// Read suffixes, binary + SOLReadResult bsufread(FILE* f); + + /// Read suffixes, text + SOLReadResult gsufread(FILE* f); + + /// Suffix head check + int sufheadcheck(SufRead *sr); + + /// Report Early Eof + SOLReadResult ReportEarlyEof(); + + /// Report bad format + SOLReadResult ReportBadFormat(); + + /// Report bad line + SOLReadResult ReportBadLine(const std::string& line); + + /// Save error message + void serror(const char* format, ...); + + +#ifndef Long + using Long = int; + using uLong = unsigned int; + /** cfront screws up when we use typedef Long uiolen; */ + /** It renders Long as long rather than Long. */ + using uiolen = uLong; +#else + static_assert (sizeof(Long)==sizeof(int), "Unusable Long"); +#endif + + using real = double; + +#ifdef CRAY + typedef int integer; +#else + typedef uLong integer; +#endif + +private: + SOLHandler2& solh_; + NLUtils& utils_; + NLHeader hdr_; + + std::string solve_msg_; + std::string err_msg_; + int internal_rv_ {0}; + SOLReadResult readresult_ {SOL_Read_Result_Not_Set}; + + const char* stub_ {nullptr}; + const char *bkind[2] = { "ASCII", "binary" }; + char *b1, buf[512], *from_solve, *s, *se; + Long Objno[2] = {-2, -2}; /* Objno[1] = solve_result_num */ + Long nOpts, Lt, Options[14], *z; + uiolen L, L1, L2; + int binary, bs, c, have_options, hs, i, j, j1, je; + int need_vbtol, newsufs, nsv, objno=-2, + presolved, rv, slin, solmsg; + double vbtol, x; + size_t len, n, n1, nbs, ui; + int ndebug, sstatus_seen; +}; + + +/// Read SOL file. +template +inline std::pair +ReadSOLFile( + const std::string& name, + SOLHandler2& solh, NLUtils& utl, int* p_internal_rv=nullptr) { + SOLReader2 solr(solh, utl); + auto res = solr.ReadSOLFile(name); + if (p_internal_rv) + *p_internal_rv = solr.internal_rv(); + return { res, solr.ErrorMessage(res) }; +} + + +} // namespace mp + +#endif // SOLREADER2_H diff --git a/nl-writer2/include/mp/sol-reader2.hpp b/nl-writer2/include/mp/sol-reader2.hpp new file mode 100644 index 000000000..a53138b8a --- /dev/null +++ b/nl-writer2/include/mp/sol-reader2.hpp @@ -0,0 +1,626 @@ +#ifndef SOLREADER2_HPP +#define SOLREADER2_HPP + +#include + +#include "mp/sol-reader2.h" + +namespace mp { + +/// Parse double +inline int decstring(const char *buf, double *val) { + char *be; + int c; + *val = std::strtod(buf, &be); + return be <= buf + || (((c = be[-1]) < '0' || c > '9') && c != '.'); +} + +/// Read a double from text or binary file +inline SOLReadResult Read( + FILE* f, int binary, double& v, std::string& err) { + err.resize(512); + if (binary ? !std::fread((char *)&v, sizeof(double), 1, f) + : !fgets((char*)err.data(), err.size()-1, f)) + return SOL_Read_Early_EOF; + if (!binary && decstring(err.data(), &v)) + return SOL_Read_Bad_Line; + return SOL_Read_OK; +} + + +/// Read a pair from text or binary file +template +inline SOLReadResult Read( + FILE* f, int binary, + std::pair& v, std::string& err) { + err.resize(512); + if (binary) { + if (fread(&v.first, sizeof(int), 1, f) != 1 + || fread(&v.second, sizeof(El), 1, f) != 1) + return SOL_Read_Early_EOF; + } else { + if (!fgets((char*)err.data(), err.size()-1, f)) + return SOL_Read_Early_EOF; + const char *s; + char *se; + v.first = (int)strtol(s = err.c_str(), &se, 10); + if (se <= s) + return SOL_Read_Bad_Line; + auto el = strtod(s = se, &se); + if (se <= s) + return SOL_Read_Bad_Line; + v.second = (El)el; + } + return SOL_Read_OK; +} + + + +template +Value VecReader::ReadNext() { + Value v; + --n_; // decrement counter + if (SOL_Read_OK != + (rr_ = Read(f_, binary_, v, err_msg_))) { + n_ = 0; // return error status + } + return v; +} + + +/// Some refurbished old code. +template +SOLReadResult +SOLReader2::ReadSOLFile( + const std::string& name) { + File file; + stub_ = name.c_str(); + + file.Open(stub_, "rb"); + internal_rv_ = 997; + if (!file) { + serror("can't open '%s'", stub_); + internal_rv_ = 998; + return SOL_Read_Fail_Open; + } + FILE* f = file.GetHandle(); + if (fread((char *)&L, sizeof(uiolen), 1, f) + && L == 6) { + /* binary files may be written by Fortran unformatted writes */ + binary = 1; + if (!fread(buf, 6, 1, f) + || strncmp(buf,"binary",6) + || !fread((char *)&L, sizeof(uiolen), 1, f) + || L != 6) { + return ReportBadFormat(); + } + } + else { + binary = 0; + rewind(f); + } + have_options = need_vbtol = 0; + bs = 1; /* omit initial backspaces */ + nbs = 0; /* number of backspaces omitted */ + if (binary) { /////////////// BINARY FORMAT ////////// + for(;;) { /////////////// SOLVE MESSAGE ////////// + if (!fread((char *)&L,sizeof(uiolen),1,f)) + return ReportEarlyEof(); + if ((L1 = L)) { + do { + n = L < sizeof(buf) ? (int)L : (int)sizeof(buf); + L -= n; + if (!fread(buf, n, 1, f)) + return ReportEarlyEof(); + if (!L) { + while(n > 0) { + if (buf[--n] != ' ') { + ++n; + break; + } + } + } + b1 = buf; + n1 = n; + if (buf[0] == '\b' && bs) { + while(--n1 > 0 + && *++b1 == '\b'); + nbs += n - n1; + if (n1 > 0) + bs = 0; + else + continue; + } + solve_msg_.append(b1, n1); + } + while(L); + } + if (!fread((char *)&L, sizeof(uiolen), 1, f)) + return ReportEarlyEof(); + if (L != L1) + return ReportBadFormat(); + if (!L) + break; + } + L1 = NumAlgCons() * sizeof(real); + if (!fread((char *)&L, sizeof(uiolen), 1, f)) + return ReportEarlyEof(); + L2 = L - (8*sizeof(Long) + 7); + if (L2 <= 4*sizeof(Long) + sizeof(real) + && !(L2 & (sizeof(Long)-1))) { + /////////////// check for Options /////////////// + if (!fread(buf, 7, 1, f)) + return ReportEarlyEof(); + if (strncmp(buf, "Options", 7)) + return ReportBadFormat(); + if (!fread((char *)Options, sizeof(Long), 4, f)) + return ReportEarlyEof(); + nOpts = Options[0]; + if (nOpts < 3 || nOpts > 9) { +bad_nOpts: + serror( + "expected nOpts between 3 and 9; got %d: ", + nOpts); + return ReportBadFormat(); + } + if (Options[2] == 3) { + nOpts -= 2; + need_vbtol = 1; + } + if (!fread((char *)(Options+4), sizeof(Long), + (size_t)(nOpts+1), f)) + return ReportEarlyEof(); + if (need_vbtol + && !fread((char *)&vbtol, sizeof(real), 1, f)) + return ReportEarlyEof(); + if (!fread((char *)&L2, sizeof(uiolen), 1, f) + || L != L2) + return ReportBadFormat(); + have_options = 1; + } + else if (L != L1) + return ReportBadFormat(); + } + else { ///////////////// TEXT FORMAT /////////////// + for(;;) { ///////////////// SOLVE MESSAGE ///////////// + if (!fgets(buf, sizeof(buf), f)) { + return ReportEarlyEof(); + } + for(se = buf; *se; se++) + if (*se == '\r' && se[1] == '\n') { + *se = '\n'; + *++se = 0; + break; + } + if (*buf == '\n') + break; + n1 = se - buf; + b1 = buf; + if (buf[0] == '\b' && bs) { + n = n1; + while(--n1 > 0 && *++b1 == '\b'); + nbs += n - n1; + if (n1 > 0) + bs = 0; + else + continue; + } + solve_msg_.append(b1, n1); + } + while((j = getc(f)) == '\n' || j == '\r'); + if (j != 'O') ////////// Check for Options ////////// + ungetc(j,f); + else { + if (!fgets(buf, sizeof(buf), f)) + return ReportEarlyEof(); + if (!strncmp(buf, "ptions", 6)) { + have_options = 1; + for(j = 0; j < 4; j++) { + if (!fgets(buf, sizeof(buf), f)) + return ReportEarlyEof(); + Options[j] = (int)strtol(buf,&se,10); + if (se == buf) + return ReportBadLine(buf); + } + nOpts = Options[0]; + if (nOpts < 3 || nOpts > 9) + goto bad_nOpts; + if (Options[2] == 3) { + nOpts -= 2; + need_vbtol = 1; + } + je = (int)(nOpts+5); + for(j = 4; j < je; j++) { + if (!fgets(buf, sizeof(buf), f)) + return ReportEarlyEof(); + Options[j] = (int)strtol(buf,&se,10); + if (se == buf) + return ReportBadLine(buf); + } + if (need_vbtol) { + if (!fgets(buf, sizeof(buf), f)) + return ReportEarlyEof(); + vbtol = strtod(buf,&se); + if (se == buf) + return ReportBadLine(buf); + } + } + } + } + + /* send termination msg */ + if (nbs) { + // Should not have \b here but... + auto b=solve_msg_.begin(); + while (solve_msg_.end()!=b && '\b'==*b) + ++b; + solve_msg_.erase(solve_msg_.begin(), b); + } + if (solve_msg_.size()) { + if (binary) + solve_msg_ += "\n"; + Handler().OnSolveMessage( + solve_msg_.c_str(), nbs); + } + + if (have_options) { + z = Options + nOpts + 1; + internal_rv_ = 996; + // Send options to Handler: + typename SOLHandler2::AMPLOptions ao; + ao.options_.assign(Options, Options+nOpts+5); + ao.has_vbtol_ = need_vbtol; + ao.vbtol_ = vbtol; + // Handler says we should stop: + if (auto rv = Handler().OnAMPLOptions(ao)) { + internal_rv_ = rv; + return SOL_Read_Bad_Options; + } + + // Some checks. + j = (int)z[3]; + internal_rv_ = 997; + if (j > NumVars() || j < 0) { + serror("Wrong NumVars %d, expected 0..%d: ", + j, NumVars()); + return ReportBadFormat(); + } + j = (int)z[1]; + if (j > NumAlgCons() || j < 0) { + serror("Wrong NumAlgCons %d, expected 0..%d: ", + j, NumAlgCons()); + return ReportBadFormat(); + } + + if (binary) { // read on for binary + L1 = j * sizeof(real); + if (!fread((char *)&L, sizeof(uiolen), 1, f)) + return ReportEarlyEof(); + if (L != L1) + return ReportBadFormat(); + } + } + else j = NumAlgCons(); + + // Read duals + if (j) { + VecReader vr(f, binary, j); + Handler().OnDualSolution(vr); + if (!CheckReader( vr, readresult_ )) + return readresult_; + } + + i = nsv = have_options ? (int)z[3] : NumVars(); +// objno = ac_->no >= 0 ? 0 : -1; + sstatus_seen = 0; + if (binary) { + if (!fread((char *)&L, sizeof(uiolen), 1, f) || L != L1) + return ReportBadFormat(); + L1 = i * sizeof(real); + if (!fread((char *)&L, sizeof(uiolen), 1, f) || L != L1) + return ReportBadFormat(); + + // Read primal solution + if (i) { + VecReader vr(f, binary, i); + Handler().OnPrimalSolution(vr); + if (!CheckReader( vr, readresult_ )) + return readresult_; + } + + if (!fread((char *)&L, sizeof(uiolen), 1, f) || L != L1) + return ReportBadFormat(); + + /* do we have an objno ? */ + if (fread((char *)&L, sizeof(uiolen), 1, f)) { + if (L != 2*sizeof(Long) && L != sizeof(integer)) + return ReportBadFormat(); + ui = (size_t) L; + if (fread(Objno, ui, 1, f) != 1 + || fread(&L1, sizeof(uiolen), 1, f) != 1 + || L != L1) + return ReportBadFormat(); + objno = (int)Objno[0]; + + /* Submit objno and solve_code to Handler. */ + Handler().OnObjno(objno); + Handler().OnSolveCode(Objno[1]); + + if (L == 2*sizeof(integer)) { + switch(bsufread(f)) { + case SOL_Read_Bad_Suffix: +badsuftable: + serror("Bad suffix in '%s'\n", stub_); + return SOL_Read_Bad_Suffix; + case SOL_Read_OK: + break; + default: + return readresult_; + } + } + while(fread(&L, sizeof(uiolen), 1, f)) {} + } + } + else { + // Read primal solution + if (i) { + VecReader vr(f, binary, i); + Handler().OnPrimalSolution(vr); + if (!CheckReader( vr, readresult_ )) + return readresult_; + } + + if (!fgets(buf,sizeof(buf), f)) + goto f_done; + if (strncmp(buf,"objno ",6)) { +bad_objno: + return ReportBadLine(buf); + } + x = strtod(s = buf+6, &se); + if (se <= s) + goto bad_objno; + objno = (int)x; + x = strtod(s = se, &se); + if (se <= s) + goto f_done; + Objno[1] = (Long)x; + + /* Submit objno and solve_code to Handler. */ + Handler().OnObjno(objno); + Handler().OnSolveCode(Objno[1]); + + switch(gsufread(f)) { + case SOL_Read_Bad_Suffix: + goto badsuftable; + case SOL_Read_OK: + break; + default: + return readresult_; + } + } +f_done: + internal_rv_ = 0; + return SOL_Read_OK; +} + + +/// Suffix head +struct SufHead { + char sufid[8]; + int kind; + int n; + int namelen; + int tablen; +}; + +/// Suffix read data +struct SufRead { + SufHead h; + char *name; + char *tabname; + char *table; +// int *ip; +// double *rp; + std::vector xp; + int tablines; +// int k; +// int nmax; +}; + +template +SOLReadResult SOLReader2::bsufread(FILE* f) { + uiolen L, L1; + + while(fread(&L, sizeof(uiolen), 1, f)) { + SufRead SR; + if (L < sizeof(SufHead)) + return SOL_Read_Bad_Suffix; + if (fread(&SR.h, sizeof(SufHead), 1, f) != 1) + return ReportEarlyEof(); + SR.tablines = SR.h.tablen - 1; + if (strncmp(SR.h.sufid, "\nSuffix\n", 8) + || sufheadcheck(&SR)) + return SOL_Read_Bad_Suffix; + if (fread(SR.name, (size_t)SR.h.namelen, 1, f) != 1) + return ReportEarlyEof(); + if (SR.h.tablen && fread(SR.table, (size_t)SR.h.tablen, + 1, f) != 1) + return SOL_Read_Bad_Suffix; + SuffixInfo si(SR.h.kind, SR.name, SR.table); + if (SR.h.kind & 4) { // real-valued + SuffixReader sr(std::move(si), f, 1, SR.h.n); + Handler().OnDblSuffix(sr); + if (!CheckReader( sr, readresult_ )) + return readresult_; + } else { // int-valued + SuffixReader sr(std::move(si), f, 1, SR.h.n); + Handler().OnIntSuffix(sr); + if (!CheckReader( sr, readresult_ )) + return readresult_; + } + if (!fread(&L1, sizeof(uiolen), 1, f) || L != L1) + return ReportEarlyEof(); + // sufput(&SR, ac, newsufs); + } + return SOL_Read_OK; +} + +/// Parse int +inline int +Lget(char **sp, int *Lp) +{ + int c; + int L = 0; + char *s = *sp; + while(*s == ' ') + s++; + if ((c = *s++) < '0' || c > '9') + return 1; + L = c - '0'; + while((c = *s) >= '0' && c <= '9') { + L = 10*L + c - '0'; + s++; + } + *Lp = L; + *sp = s; + switch(*s) { + case '\r': + if (*++s != '\n') + break; + *sp = s; + /* no break */ + case 0: + case ' ': + case '\n': + return 0; + } + return 1; +} + + +template +SOLReadResult SOLReader2::gsufread(FILE* f) { + char *s, *se; + size_t L; + char buf[512]; + + while(fgets(buf, sizeof(buf)-1, f)) { + SufRead SR; + if (strncmp(buf, "suffix ", 7)) + return ReportBadLine(buf); + s = buf + 7; + if (Lget(&s, &SR.h.kind) + || Lget(&s, &SR.h.n) + || Lget(&s, &SR.h.namelen) + || Lget(&s, &SR.h.tablen) + || Lget(&s, &SR.tablines)) + return ReportBadLine(buf); + if (sufheadcheck(&SR)) + return ReportBadLine(buf); + if (!fgets(buf, sizeof(buf)-1, f) + || buf[SR.h.namelen-1] != '\n' + && (buf[SR.h.namelen-1] != '\r' + || buf[SR.h.namelen] != '\n')) + return ReportBadLine(buf); + buf[SR.h.namelen-1] = 0; + strcpy(SR.name, buf); + if (SR.h.tablen) { + s = SR.table; + se = s + SR.h.tablen; + for(int i = 1; i < SR.tablines; i++) { + if (!fgets(s, se-s, f)) + return ReportEarlyEof(); + s += strlen(s); + } + if (!fgets(buf, sizeof(buf)-1, f)) + return ReportEarlyEof(); + if (!(L = strlen(buf)) || buf[--L] != '\n' + || L >= se - s) + return ReportBadLine(buf); + if (L) { + if (buf[L-1] != '\r' || --L) { + buf[L] = 0; + memcpy(s, buf, L); + } + } + } + SuffixInfo si(SR.h.kind, SR.name, SR.table); + if (SR.h.kind & 4) { // real-valued + SuffixReader sr(std::move(si), f, binary, SR.h.n); + Handler().OnDblSuffix(sr); + if (!CheckReader( sr, readresult_ )) + return readresult_; + } else { // int-valued + SuffixReader sr(std::move(si), f, binary, SR.h.n); + Handler().OnIntSuffix(sr); + if (!CheckReader( sr, readresult_ )) + return readresult_; + } + // sufput(&SR, ac, newsufs); + } + return SOL_Read_OK; +} + +template +int SOLReader2::sufheadcheck(SufRead* sr) { + int n; + + n = (int)sr->h.n; + if (sr->h.kind < 0 || sr->h.kind > 15 || n < 0 || sr->h.namelen < 2 + || sr->h.tablen < 0) + return 1; + i = (int)sr->h.kind & 3; + if (sr->h.tablen + && (sr->tablines > sr->h.tablen + 1 || sr->tablines < 1)) + return 1; + sr->xp.resize((sr->h.tablen + 2*sr->h.namelen + 6)); + sr->name = (char*)sr->xp.data(); + sr->table = sr->name + sr->h.namelen; + sr->tabname = sr->table + sr->h.tablen; + return 0; +} + +template +SOLReadResult SOLReader2::ReportEarlyEof() { + serror("error reading '%s' (errno=%d)", stub_, errno); + return readresult_ = SOL_Read_Early_EOF; +} + +template +SOLReadResult SOLReader2::ReportBadFormat() { + serror("Bad %s solution file '%s' (errno=%d)", + bkind[binary], stub_, errno); + return readresult_ = SOL_Read_Bad_Format; +} + +template +SOLReadResult SOLReader2:: +ReportBadLine(const std::string& line) { + serror("Bad line in '%s': %s", stub_, line.c_str()); + return readresult_ = SOL_Read_Bad_Line; +} + + +template +void SOLReader2::serror( + const char* format, ...) { + va_list args; + va_start (args, format); + auto n = std::vsnprintf (0, 0, format, args); + va_end (args); + va_start (args, format); + std::vector buf(n+1); + auto n1 = std::vsnprintf (buf.data(), n+1, format, args); + assert(n==n1); + va_end (args); + if (err_msg_.size()) + err_msg_ += '\n'; + err_msg_ += buf.data(); +} + + + +} // namespace mp + +#endif // SOLREADER2_HPP diff --git a/nl-writer2/src/dtoa.c b/nl-writer2/src/dtoa.c new file mode 100644 index 000000000..11469089c --- /dev/null +++ b/nl-writer2/src/dtoa.c @@ -0,0 +1,6210 @@ +/**************************************************************** + * + * The author of this software is David M. Gay. + * + * Copyright (c) 1991, 2000, 2001 by Lucent Technologies. + * + * Permission to use, copy, modify, and distribute this software for any + * purpose without fee is hereby granted, provided that this entire notice + * is included in all copies of any software which is or includes a copy + * or modification of this software and in all copies of the supporting + * documentation for such software. + * + * THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR IMPLIED + * WARRANTY. IN PARTICULAR, NEITHER THE AUTHOR NOR LUCENT MAKES ANY + * REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE MERCHANTABILITY + * OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR PURPOSE. + * + ***************************************************************/ + +/* Please send bug reports to David M. Gay (dmg at acm dot org, + * with " at " changed at "@" and " dot " changed to "."). */ + +/* On a machine with IEEE extended-precision registers, it is + * necessary to specify double-precision (53-bit) rounding precision + * before invoking strtod or dtoa. If the machine uses (the equivalent + * of) Intel 80x87 arithmetic, the call + * _control87(PC_53, MCW_PC); + * does this with many compilers. Whether this or another call is + * appropriate depends on the compiler; for this to work, it may be + * necessary to #include "float.h" or another system-dependent header + * file. + */ + +/* strtod for IEEE-, VAX-, and IBM-arithmetic machines. + * (Note that IEEE arithmetic is disabled by gcc's -ffast-math flag.) + * + * This strtod returns a nearest machine number to the input decimal + * string (or sets errno to ERANGE). With IEEE arithmetic, ties are + * broken by the IEEE round-even rule. Otherwise ties are broken by + * biased rounding (add half and chop). + * + * Inspired loosely by William D. Clinger's paper "How to Read Floating + * Point Numbers Accurately" [Proc. ACM SIGPLAN '90, pp. 92-101]. + * + * Modifications: + * + * 1. We only require IEEE, IBM, or VAX double-precision + * arithmetic (not IEEE double-extended). + * 2. We get by with floating-point arithmetic in a case that + * Clinger missed -- when we're computing d * 10^n + * for a small integer d and the integer n is not too + * much larger than 22 (the maximum integer k for which + * we can represent 10^k exactly), we may be able to + * compute (d*10^k) * 10^(e-k) with just one roundoff. + * 3. Rather than a bit-at-a-time adjustment of the binary + * result in the hard case, we use floating-point + * arithmetic to determine the adjustment to within + * one bit; only in really hard cases do we need to + * compute a second residual. + * 4. Because of 3., we don't need a large table of powers of 10 + * for ten-to-e (just some small tables, e.g. of 10^k + * for 0 <= k <= 22). + */ + +/* + * #define IEEE_8087 for IEEE-arithmetic machines where the least + * significant byte has the lowest address. + * #define IEEE_MC68k for IEEE-arithmetic machines where the most + * significant byte has the lowest address. + * #define Long int on machines with 32-bit ints and 64-bit longs. + * #define IBM for IBM mainframe-style floating-point arithmetic. + * #define VAX for VAX-style floating-point arithmetic (D_floating). + * #define No_leftright to omit left-right logic in fast floating-point + * computation of dtoa. This will cause dtoa modes 4 and 5 to be + * treated the same as modes 2 and 3 for some inputs. + * #define Honor_FLT_ROUNDS if FLT_ROUNDS can assume the values 2 or 3 + * and strtod and dtoa should round accordingly. Unless Trust_FLT_ROUNDS + * is also #defined, fegetround() will be queried for the rounding mode. + * Note that both FLT_ROUNDS and fegetround() are specified by the C99 + * standard (and are specified to be consistent, with fesetround() + * affecting the value of FLT_ROUNDS), but that some (Linux) systems + * do not work correctly in this regard, so using fegetround() is more + * portable than using FLT_ROUNDS directly. + * #define Check_FLT_ROUNDS if FLT_ROUNDS can assume the values 2 or 3 + * and Honor_FLT_ROUNDS is not #defined. + * #define RND_PRODQUOT to use rnd_prod and rnd_quot (assembly routines + * that use extended-precision instructions to compute rounded + * products and quotients) with IBM. + * #define ROUND_BIASED for IEEE-format with biased rounding and arithmetic + * that rounds toward +Infinity. + * #define ROUND_BIASED_without_Round_Up for IEEE-format with biased + * rounding when the underlying floating-point arithmetic uses + * unbiased rounding. This prevent using ordinary floating-point + * arithmetic when the result could be computed with one rounding error. + * #define Inaccurate_Divide for IEEE-format with correctly rounded + * products but inaccurate quotients, e.g., for Intel i860. + * #define NO_LONG_LONG on machines that do not have a "long long" + * integer type (of >= 64 bits). On such machines, you can + * #define Just_16 to store 16 bits per 32-bit Long when doing + * high-precision integer arithmetic. Whether this speeds things + * up or slows things down depends on the machine and the number + * being converted. If long long is available and the name is + * something other than "long long", #define Llong to be the name, + * and if "unsigned Llong" does not work as an unsigned version of + * Llong, #define #ULLong to be the corresponding unsigned type. + * #define Bad_float_h if your system lacks a float.h or if it does not + * define some or all of DBL_DIG, DBL_MAX_10_EXP, DBL_MAX_EXP, + * FLT_RADIX, FLT_ROUNDS, and DBL_MAX. + * #define MALLOC your_malloc, where your_malloc(n) acts like malloc(n) + * if memory is available and otherwise does something you deem + * appropriate. If MALLOC is undefined, malloc will be invoked + * directly -- and assumed always to succeed. Similarly, if you + * want something other than the system's free() to be called to + * recycle memory acquired from MALLOC, #define FREE to be the + * name of the alternate routine. (FREE or free is only called in + * pathological cases, e.g., in a dtoa call after a dtoa return in + * mode 3 with thousands of digits requested.) + * #define Omit_Private_Memory to omit logic (added Jan. 1998) for making + * memory allocations from a private pool of memory when possible. + * When used, the private pool is PRIVATE_MEM bytes long: 2304 bytes, + * unless #defined to be a different length. This default length + * suffices to get rid of MALLOC calls except for unusual cases, + * such as decimal-to-binary conversion of a very long string of + * digits. The longest string dtoa can return is about 751 bytes + * long. For conversions by strtod of strings of 800 digits and + * all dtoa conversions in single-threaded executions with 8-byte + * pointers, PRIVATE_MEM >= 7400 appears to suffice; with 4-byte + * pointers, PRIVATE_MEM >= 7112 appears adequate. + * #define NO_INFNAN_CHECK if you do not wish to have INFNAN_CHECK + * #defined automatically on IEEE systems. On such systems, + * when INFNAN_CHECK is #defined, strtod checks + * for Infinity and NaN (case insensitively). On some systems + * (e.g., some HP systems), it may be necessary to #define NAN_WORD0 + * appropriately -- to the most significant word of a quiet NaN. + * (On HP Series 700/800 machines, -DNAN_WORD0=0x7ff40000 works.) + * When INFNAN_CHECK is #defined and No_Hex_NaN is not #defined, + * strtod also accepts (case insensitively) strings of the form + * NaN(x), where x is a string of hexadecimal digits and spaces; + * if there is only one string of hexadecimal digits, it is taken + * for the 52 fraction bits of the resulting NaN; if there are two + * or more strings of hex digits, the first is for the high 20 bits, + * the second and subsequent for the low 32 bits, with intervening + * white space ignored; but if this results in none of the 52 + * fraction bits being on (an IEEE Infinity symbol), then NAN_WORD0 + * and NAN_WORD1 are used instead. + * #define MULTIPLE_THREADS if the system offers preemptively scheduled + * multiple threads. In this case, you must provide (or suitably + * #define) two locks, acquired by ACQUIRE_DTOA_LOCK(n) and freed + * by FREE_DTOA_LOCK(n) for n = 0 or 1. (The second lock, accessed + * in pow5mult, ensures lazy evaluation of only one copy of high + * powers of 5; omitting this lock would introduce a small + * probability of wasting memory, but would otherwise be harmless.) + * You must also invoke freedtoa(s) to free the value s returned by + * dtoa. You may do so whether or not MULTIPLE_THREADS is #defined. + + * When MULTIPLE_THREADS is #defined, this source file provides + * void set_max_dtoa_threads(unsigned int n); + * and expects + * unsigned int dtoa_get_threadno(void); + * to be available (possibly provided by + * #define dtoa_get_threadno omp_get_thread_num + * if OpenMP is in use or by + * #define dtoa_get_threadno pthread_self + * if Pthreads is in use), to return the current thread number. + * If set_max_dtoa_threads(n) was called and the current thread + * number is k with k < n, then calls on ACQUIRE_DTOA_LOCK(...) and + * FREE_DTOA_LOCK(...) are avoided; instead each thread with thread + * number < n has a separate copy of relevant data structures. + * After set_max_dtoa_threads(n), a call set_max_dtoa_threads(m) + * with m <= n has has no effect, but a call with m > n is honored. + * Such a call invokes REALLOC (assumed to be "realloc" if REALLOC + * is not #defined) to extend the size of the relevant array. + + * #define NO_IEEE_Scale to disable new (Feb. 1997) logic in strtod that + * avoids underflows on inputs whose result does not underflow. + * If you #define NO_IEEE_Scale on a machine that uses IEEE-format + * floating-point numbers and flushes underflows to zero rather + * than implementing gradual underflow, then you must also #define + * Sudden_Underflow. + * #define USE_LOCALE to use the current locale's decimal_point value. + * #define SET_INEXACT if IEEE arithmetic is being used and extra + * computation should be done to set the inexact flag when the + * result is inexact and avoid setting inexact when the result + * is exact. In this case, dtoa.c must be compiled in + * an environment, perhaps provided by #include "dtoa.c" in a + * suitable wrapper, that defines two functions, + * int get_inexact(void); + * void clear_inexact(void); + * such that get_inexact() returns a nonzero value if the + * inexact bit is already set, and clear_inexact() sets the + * inexact bit to 0. When SET_INEXACT is #defined, strtod + * also does extra computations to set the underflow and overflow + * flags when appropriate (i.e., when the result is tiny and + * inexact or when it is a numeric value rounded to +-infinity). + * #define NO_ERRNO if strtod should not assign errno = ERANGE when + * the result overflows to +-Infinity or underflows to 0. + * When errno should be assigned, under seemingly rare conditions + * it may be necessary to define Set_errno(x) suitably, e.g., in + * a local errno.h, such as + * #include + * #define Set_errno(x) _set_errno(x) + * #define NO_HEX_FP to omit recognition of hexadecimal floating-point + * values by strtod. + * #define NO_STRTOD_BIGCOMP (on IEEE-arithmetic systems only for now) + * to disable logic for "fast" testing of very long input strings + * to strtod. This testing proceeds by initially truncating the + * input string, then if necessary comparing the whole string with + * a decimal expansion to decide close cases. This logic is only + * used for input more than STRTOD_DIGLIM digits long (default 40). + */ + +#ifndef Long +#define Long int +#endif +#ifndef ULong +typedef unsigned Long ULong; +#endif + +#ifdef DEBUG +#include +#include "stdio.h" +#define Bug(x) {fprintf(stderr, "%s\n", x); exit(1);} +#define Debug(x) x +int dtoa_stats[7]; /* strtod_{64,96,bigcomp},dtoa_{exact,64,96,bigcomp} */ +#else +#define assert(x) /*nothing*/ +#define Debug(x) /*nothing*/ +#endif + +#include "stdlib.h" +#include "string.h" + +#ifdef USE_LOCALE +#include "locale.h" +#endif + +#ifdef Honor_FLT_ROUNDS +#ifndef Trust_FLT_ROUNDS +#include +#endif +#endif + +#ifdef __cplusplus +extern "C" { +#endif +#ifdef MALLOC +extern void *MALLOC(size_t); +#else +#define MALLOC malloc +#endif + +#ifdef REALLOC +extern void *REALLOC(void*,size_t); +#else +#define REALLOC realloc +#endif + +#ifndef FREE +#define FREE free +#endif + +#ifdef __cplusplus + } +#endif + + +#ifndef Omit_Private_Memory +#ifndef PRIVATE_MEM +#define PRIVATE_MEM 2304 +#endif +#define PRIVATE_mem ((PRIVATE_MEM+sizeof(double)-1)/sizeof(double)) +static double private_mem[PRIVATE_mem], *pmem_next = private_mem; +#endif + +#undef IEEE_Arith +#undef Avoid_Underflow +#ifdef IEEE_MC68k +#define IEEE_Arith +#endif +#ifdef IEEE_8087 +#define IEEE_Arith +#endif + +#ifdef IEEE_Arith +#ifndef NO_INFNAN_CHECK +#undef INFNAN_CHECK +#define INFNAN_CHECK +#endif +#else +#undef INFNAN_CHECK +#define NO_STRTOD_BIGCOMP +#endif + +#include "errno.h" + +#ifdef NO_ERRNO /*{*/ +#undef Set_errno +#define Set_errno(x) +#else +#ifndef Set_errno +#define Set_errno(x) errno = x +#endif +#endif /*}*/ + +#ifdef Bad_float_h + +#ifdef IEEE_Arith +#define DBL_DIG 15 +#define DBL_MAX_10_EXP 308 +#define DBL_MAX_EXP 1024 +#define FLT_RADIX 2 +#endif /*IEEE_Arith*/ + +#ifdef IBM +#define DBL_DIG 16 +#define DBL_MAX_10_EXP 75 +#define DBL_MAX_EXP 63 +#define FLT_RADIX 16 +#define DBL_MAX 7.2370055773322621e+75 +#endif + +#ifdef VAX +#define DBL_DIG 16 +#define DBL_MAX_10_EXP 38 +#define DBL_MAX_EXP 127 +#define FLT_RADIX 2 +#define DBL_MAX 1.7014118346046923e+38 +#endif + +#ifndef LONG_MAX +#define LONG_MAX 2147483647 +#endif + +#else /* ifndef Bad_float_h */ +#include "float.h" +#endif /* Bad_float_h */ + +#ifndef __MATH_H__ +#include "math.h" +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +#define IEEE_8087 +#define IEEE_Arith +#if defined(IEEE_8087) + defined(IEEE_MC68k) + defined(VAX) + defined(IBM) != 1 +Exactly one of IEEE_8087, IEEE_MC68k, VAX, or IBM should be defined. +#endif + +#undef USE_BF96 + +#ifdef NO_LONG_LONG /*{{*/ +#undef ULLong +#ifdef Just_16 +#undef Pack_32 +/* When Pack_32 is not defined, we store 16 bits per 32-bit Long. + * This makes some inner loops simpler and sometimes saves work + * during multiplications, but it often seems to make things slightly + * slower. Hence the default is now to store 32 bits per Long. + */ +#endif +#else /*}{ long long available */ +#ifndef Llong +#define Llong long long +#endif +#ifndef ULLong +#define ULLong unsigned Llong +#endif +#ifndef NO_BF96 /*{*/ +#define USE_BF96 + +#ifdef SET_INEXACT +#define dtoa_divmax 27 +#else +static int dtoa_divmax = 2; /* Permit experimenting: on some systems, 64-bit integer */ + /* division is slow enough that we may sometimes want to */ + /* avoid using it. We assume (but do not check) that */ + /* dtoa_divmax <= 27.*/ +#endif + +typedef struct BF96 { /* Normalized 96-bit software floating point numbers */ + unsigned int b0,b1,b2; /* b0 = most significant, binary point just to its left */ + int e; /* number represented = b * 2^e, with .5 <= b < 1 */ + } BF96; + + static BF96 pten[667] = { + { 0xeef453d6, 0x923bd65a, 0x113faa29, -1136 }, + { 0x9558b466, 0x1b6565f8, 0x4ac7ca59, -1132 }, + { 0xbaaee17f, 0xa23ebf76, 0x5d79bcf0, -1129 }, + { 0xe95a99df, 0x8ace6f53, 0xf4d82c2c, -1126 }, + { 0x91d8a02b, 0xb6c10594, 0x79071b9b, -1122 }, + { 0xb64ec836, 0xa47146f9, 0x9748e282, -1119 }, + { 0xe3e27a44, 0x4d8d98b7, 0xfd1b1b23, -1116 }, + { 0x8e6d8c6a, 0xb0787f72, 0xfe30f0f5, -1112 }, + { 0xb208ef85, 0x5c969f4f, 0xbdbd2d33, -1109 }, + { 0xde8b2b66, 0xb3bc4723, 0xad2c7880, -1106 }, + { 0x8b16fb20, 0x3055ac76, 0x4c3bcb50, -1102 }, + { 0xaddcb9e8, 0x3c6b1793, 0xdf4abe24, -1099 }, + { 0xd953e862, 0x4b85dd78, 0xd71d6dad, -1096 }, + { 0x87d4713d, 0x6f33aa6b, 0x8672648c, -1092 }, + { 0xa9c98d8c, 0xcb009506, 0x680efdaf, -1089 }, + { 0xd43bf0ef, 0xfdc0ba48, 0x0212bd1b, -1086 }, + { 0x84a57695, 0xfe98746d, 0x014bb630, -1082 }, + { 0xa5ced43b, 0x7e3e9188, 0x419ea3bd, -1079 }, + { 0xcf42894a, 0x5dce35ea, 0x52064cac, -1076 }, + { 0x818995ce, 0x7aa0e1b2, 0x7343efeb, -1072 }, + { 0xa1ebfb42, 0x19491a1f, 0x1014ebe6, -1069 }, + { 0xca66fa12, 0x9f9b60a6, 0xd41a26e0, -1066 }, + { 0xfd00b897, 0x478238d0, 0x8920b098, -1063 }, + { 0x9e20735e, 0x8cb16382, 0x55b46e5f, -1059 }, + { 0xc5a89036, 0x2fddbc62, 0xeb2189f7, -1056 }, + { 0xf712b443, 0xbbd52b7b, 0xa5e9ec75, -1053 }, + { 0x9a6bb0aa, 0x55653b2d, 0x47b233c9, -1049 }, + { 0xc1069cd4, 0xeabe89f8, 0x999ec0bb, -1046 }, + { 0xf148440a, 0x256e2c76, 0xc00670ea, -1043 }, + { 0x96cd2a86, 0x5764dbca, 0x38040692, -1039 }, + { 0xbc807527, 0xed3e12bc, 0xc6050837, -1036 }, + { 0xeba09271, 0xe88d976b, 0xf7864a44, -1033 }, + { 0x93445b87, 0x31587ea3, 0x7ab3ee6a, -1029 }, + { 0xb8157268, 0xfdae9e4c, 0x5960ea05, -1026 }, + { 0xe61acf03, 0x3d1a45df, 0x6fb92487, -1023 }, + { 0x8fd0c162, 0x06306bab, 0xa5d3b6d4, -1019 }, + { 0xb3c4f1ba, 0x87bc8696, 0x8f48a489, -1016 }, + { 0xe0b62e29, 0x29aba83c, 0x331acdab, -1013 }, + { 0x8c71dcd9, 0xba0b4925, 0x9ff0c08b, -1009 }, + { 0xaf8e5410, 0x288e1b6f, 0x07ecf0ae, -1006 }, + { 0xdb71e914, 0x32b1a24a, 0xc9e82cd9, -1003 }, + { 0x892731ac, 0x9faf056e, 0xbe311c08, -999 }, + { 0xab70fe17, 0xc79ac6ca, 0x6dbd630a, -996 }, + { 0xd64d3d9d, 0xb981787d, 0x092cbbcc, -993 }, + { 0x85f04682, 0x93f0eb4e, 0x25bbf560, -989 }, + { 0xa76c5823, 0x38ed2621, 0xaf2af2b8, -986 }, + { 0xd1476e2c, 0x07286faa, 0x1af5af66, -983 }, + { 0x82cca4db, 0x847945ca, 0x50d98d9f, -979 }, + { 0xa37fce12, 0x6597973c, 0xe50ff107, -976 }, + { 0xcc5fc196, 0xfefd7d0c, 0x1e53ed49, -973 }, + { 0xff77b1fc, 0xbebcdc4f, 0x25e8e89c, -970 }, + { 0x9faacf3d, 0xf73609b1, 0x77b19161, -966 }, + { 0xc795830d, 0x75038c1d, 0xd59df5b9, -963 }, + { 0xf97ae3d0, 0xd2446f25, 0x4b057328, -960 }, + { 0x9becce62, 0x836ac577, 0x4ee367f9, -956 }, + { 0xc2e801fb, 0x244576d5, 0x229c41f7, -953 }, + { 0xf3a20279, 0xed56d48a, 0x6b435275, -950 }, + { 0x9845418c, 0x345644d6, 0x830a1389, -946 }, + { 0xbe5691ef, 0x416bd60c, 0x23cc986b, -943 }, + { 0xedec366b, 0x11c6cb8f, 0x2cbfbe86, -940 }, + { 0x94b3a202, 0xeb1c3f39, 0x7bf7d714, -936 }, + { 0xb9e08a83, 0xa5e34f07, 0xdaf5ccd9, -933 }, + { 0xe858ad24, 0x8f5c22c9, 0xd1b3400f, -930 }, + { 0x91376c36, 0xd99995be, 0x23100809, -926 }, + { 0xb5854744, 0x8ffffb2d, 0xabd40a0c, -923 }, + { 0xe2e69915, 0xb3fff9f9, 0x16c90c8f, -920 }, + { 0x8dd01fad, 0x907ffc3b, 0xae3da7d9, -916 }, + { 0xb1442798, 0xf49ffb4a, 0x99cd11cf, -913 }, + { 0xdd95317f, 0x31c7fa1d, 0x40405643, -910 }, + { 0x8a7d3eef, 0x7f1cfc52, 0x482835ea, -906 }, + { 0xad1c8eab, 0x5ee43b66, 0xda324365, -903 }, + { 0xd863b256, 0x369d4a40, 0x90bed43e, -900 }, + { 0x873e4f75, 0xe2224e68, 0x5a7744a6, -896 }, + { 0xa90de353, 0x5aaae202, 0x711515d0, -893 }, + { 0xd3515c28, 0x31559a83, 0x0d5a5b44, -890 }, + { 0x8412d999, 0x1ed58091, 0xe858790a, -886 }, + { 0xa5178fff, 0x668ae0b6, 0x626e974d, -883 }, + { 0xce5d73ff, 0x402d98e3, 0xfb0a3d21, -880 }, + { 0x80fa687f, 0x881c7f8e, 0x7ce66634, -876 }, + { 0xa139029f, 0x6a239f72, 0x1c1fffc1, -873 }, + { 0xc9874347, 0x44ac874e, 0xa327ffb2, -870 }, + { 0xfbe91419, 0x15d7a922, 0x4bf1ff9f, -867 }, + { 0x9d71ac8f, 0xada6c9b5, 0x6f773fc3, -863 }, + { 0xc4ce17b3, 0x99107c22, 0xcb550fb4, -860 }, + { 0xf6019da0, 0x7f549b2b, 0x7e2a53a1, -857 }, + { 0x99c10284, 0x4f94e0fb, 0x2eda7444, -853 }, + { 0xc0314325, 0x637a1939, 0xfa911155, -850 }, + { 0xf03d93ee, 0xbc589f88, 0x793555ab, -847 }, + { 0x96267c75, 0x35b763b5, 0x4bc1558b, -843 }, + { 0xbbb01b92, 0x83253ca2, 0x9eb1aaed, -840 }, + { 0xea9c2277, 0x23ee8bcb, 0x465e15a9, -837 }, + { 0x92a1958a, 0x7675175f, 0x0bfacd89, -833 }, + { 0xb749faed, 0x14125d36, 0xcef980ec, -830 }, + { 0xe51c79a8, 0x5916f484, 0x82b7e127, -827 }, + { 0x8f31cc09, 0x37ae58d2, 0xd1b2ecb8, -823 }, + { 0xb2fe3f0b, 0x8599ef07, 0x861fa7e6, -820 }, + { 0xdfbdcece, 0x67006ac9, 0x67a791e0, -817 }, + { 0x8bd6a141, 0x006042bd, 0xe0c8bb2c, -813 }, + { 0xaecc4991, 0x4078536d, 0x58fae9f7, -810 }, + { 0xda7f5bf5, 0x90966848, 0xaf39a475, -807 }, + { 0x888f9979, 0x7a5e012d, 0x6d8406c9, -803 }, + { 0xaab37fd7, 0xd8f58178, 0xc8e5087b, -800 }, + { 0xd5605fcd, 0xcf32e1d6, 0xfb1e4a9a, -797 }, + { 0x855c3be0, 0xa17fcd26, 0x5cf2eea0, -793 }, + { 0xa6b34ad8, 0xc9dfc06f, 0xf42faa48, -790 }, + { 0xd0601d8e, 0xfc57b08b, 0xf13b94da, -787 }, + { 0x823c1279, 0x5db6ce57, 0x76c53d08, -783 }, + { 0xa2cb1717, 0xb52481ed, 0x54768c4b, -780 }, + { 0xcb7ddcdd, 0xa26da268, 0xa9942f5d, -777 }, + { 0xfe5d5415, 0x0b090b02, 0xd3f93b35, -774 }, + { 0x9efa548d, 0x26e5a6e1, 0xc47bc501, -770 }, + { 0xc6b8e9b0, 0x709f109a, 0x359ab641, -767 }, + { 0xf867241c, 0x8cc6d4c0, 0xc30163d2, -764 }, + { 0x9b407691, 0xd7fc44f8, 0x79e0de63, -760 }, + { 0xc2109436, 0x4dfb5636, 0x985915fc, -757 }, + { 0xf294b943, 0xe17a2bc4, 0x3e6f5b7b, -754 }, + { 0x979cf3ca, 0x6cec5b5a, 0xa705992c, -750 }, + { 0xbd8430bd, 0x08277231, 0x50c6ff78, -747 }, + { 0xece53cec, 0x4a314ebd, 0xa4f8bf56, -744 }, + { 0x940f4613, 0xae5ed136, 0x871b7795, -740 }, + { 0xb9131798, 0x99f68584, 0x28e2557b, -737 }, + { 0xe757dd7e, 0xc07426e5, 0x331aeada, -734 }, + { 0x9096ea6f, 0x3848984f, 0x3ff0d2c8, -730 }, + { 0xb4bca50b, 0x065abe63, 0x0fed077a, -727 }, + { 0xe1ebce4d, 0xc7f16dfb, 0xd3e84959, -724 }, + { 0x8d3360f0, 0x9cf6e4bd, 0x64712dd7, -720 }, + { 0xb080392c, 0xc4349dec, 0xbd8d794d, -717 }, + { 0xdca04777, 0xf541c567, 0xecf0d7a0, -714 }, + { 0x89e42caa, 0xf9491b60, 0xf41686c4, -710 }, + { 0xac5d37d5, 0xb79b6239, 0x311c2875, -707 }, + { 0xd77485cb, 0x25823ac7, 0x7d633293, -704 }, + { 0x86a8d39e, 0xf77164bc, 0xae5dff9c, -700 }, + { 0xa8530886, 0xb54dbdeb, 0xd9f57f83, -697 }, + { 0xd267caa8, 0x62a12d66, 0xd072df63, -694 }, + { 0x8380dea9, 0x3da4bc60, 0x4247cb9e, -690 }, + { 0xa4611653, 0x8d0deb78, 0x52d9be85, -687 }, + { 0xcd795be8, 0x70516656, 0x67902e27, -684 }, + { 0x806bd971, 0x4632dff6, 0x00ba1cd8, -680 }, + { 0xa086cfcd, 0x97bf97f3, 0x80e8a40e, -677 }, + { 0xc8a883c0, 0xfdaf7df0, 0x6122cd12, -674 }, + { 0xfad2a4b1, 0x3d1b5d6c, 0x796b8057, -671 }, + { 0x9cc3a6ee, 0xc6311a63, 0xcbe33036, -667 }, + { 0xc3f490aa, 0x77bd60fc, 0xbedbfc44, -664 }, + { 0xf4f1b4d5, 0x15acb93b, 0xee92fb55, -661 }, + { 0x99171105, 0x2d8bf3c5, 0x751bdd15, -657 }, + { 0xbf5cd546, 0x78eef0b6, 0xd262d45a, -654 }, + { 0xef340a98, 0x172aace4, 0x86fb8971, -651 }, + { 0x9580869f, 0x0e7aac0e, 0xd45d35e6, -647 }, + { 0xbae0a846, 0xd2195712, 0x89748360, -644 }, + { 0xe998d258, 0x869facd7, 0x2bd1a438, -641 }, + { 0x91ff8377, 0x5423cc06, 0x7b6306a3, -637 }, + { 0xb67f6455, 0x292cbf08, 0x1a3bc84c, -634 }, + { 0xe41f3d6a, 0x7377eeca, 0x20caba5f, -631 }, + { 0x8e938662, 0x882af53e, 0x547eb47b, -627 }, + { 0xb23867fb, 0x2a35b28d, 0xe99e619a, -624 }, + { 0xdec681f9, 0xf4c31f31, 0x6405fa00, -621 }, + { 0x8b3c113c, 0x38f9f37e, 0xde83bc40, -617 }, + { 0xae0b158b, 0x4738705e, 0x9624ab50, -614 }, + { 0xd98ddaee, 0x19068c76, 0x3badd624, -611 }, + { 0x87f8a8d4, 0xcfa417c9, 0xe54ca5d7, -607 }, + { 0xa9f6d30a, 0x038d1dbc, 0x5e9fcf4c, -604 }, + { 0xd47487cc, 0x8470652b, 0x7647c320, -601 }, + { 0x84c8d4df, 0xd2c63f3b, 0x29ecd9f4, -597 }, + { 0xa5fb0a17, 0xc777cf09, 0xf4681071, -594 }, + { 0xcf79cc9d, 0xb955c2cc, 0x7182148d, -591 }, + { 0x81ac1fe2, 0x93d599bf, 0xc6f14cd8, -587 }, + { 0xa21727db, 0x38cb002f, 0xb8ada00e, -584 }, + { 0xca9cf1d2, 0x06fdc03b, 0xa6d90811, -581 }, + { 0xfd442e46, 0x88bd304a, 0x908f4a16, -578 }, + { 0x9e4a9cec, 0x15763e2e, 0x9a598e4e, -574 }, + { 0xc5dd4427, 0x1ad3cdba, 0x40eff1e1, -571 }, + { 0xf7549530, 0xe188c128, 0xd12bee59, -568 }, + { 0x9a94dd3e, 0x8cf578b9, 0x82bb74f8, -564 }, + { 0xc13a148e, 0x3032d6e7, 0xe36a5236, -561 }, + { 0xf18899b1, 0xbc3f8ca1, 0xdc44e6c3, -558 }, + { 0x96f5600f, 0x15a7b7e5, 0x29ab103a, -554 }, + { 0xbcb2b812, 0xdb11a5de, 0x7415d448, -551 }, + { 0xebdf6617, 0x91d60f56, 0x111b495b, -548 }, + { 0x936b9fce, 0xbb25c995, 0xcab10dd9, -544 }, + { 0xb84687c2, 0x69ef3bfb, 0x3d5d514f, -541 }, + { 0xe65829b3, 0x046b0afa, 0x0cb4a5a3, -538 }, + { 0x8ff71a0f, 0xe2c2e6dc, 0x47f0e785, -534 }, + { 0xb3f4e093, 0xdb73a093, 0x59ed2167, -531 }, + { 0xe0f218b8, 0xd25088b8, 0x306869c1, -528 }, + { 0x8c974f73, 0x83725573, 0x1e414218, -524 }, + { 0xafbd2350, 0x644eeacf, 0xe5d1929e, -521 }, + { 0xdbac6c24, 0x7d62a583, 0xdf45f746, -518 }, + { 0x894bc396, 0xce5da772, 0x6b8bba8c, -514 }, + { 0xab9eb47c, 0x81f5114f, 0x066ea92f, -511 }, + { 0xd686619b, 0xa27255a2, 0xc80a537b, -508 }, + { 0x8613fd01, 0x45877585, 0xbd06742c, -504 }, + { 0xa798fc41, 0x96e952e7, 0x2c481138, -501 }, + { 0xd17f3b51, 0xfca3a7a0, 0xf75a1586, -498 }, + { 0x82ef8513, 0x3de648c4, 0x9a984d73, -494 }, + { 0xa3ab6658, 0x0d5fdaf5, 0xc13e60d0, -491 }, + { 0xcc963fee, 0x10b7d1b3, 0x318df905, -488 }, + { 0xffbbcfe9, 0x94e5c61f, 0xfdf17746, -485 }, + { 0x9fd561f1, 0xfd0f9bd3, 0xfeb6ea8b, -481 }, + { 0xc7caba6e, 0x7c5382c8, 0xfe64a52e, -478 }, + { 0xf9bd690a, 0x1b68637b, 0x3dfdce7a, -475 }, + { 0x9c1661a6, 0x51213e2d, 0x06bea10c, -471 }, + { 0xc31bfa0f, 0xe5698db8, 0x486e494f, -468 }, + { 0xf3e2f893, 0xdec3f126, 0x5a89dba3, -465 }, + { 0x986ddb5c, 0x6b3a76b7, 0xf8962946, -461 }, + { 0xbe895233, 0x86091465, 0xf6bbb397, -458 }, + { 0xee2ba6c0, 0x678b597f, 0x746aa07d, -455 }, + { 0x94db4838, 0x40b717ef, 0xa8c2a44e, -451 }, + { 0xba121a46, 0x50e4ddeb, 0x92f34d62, -448 }, + { 0xe896a0d7, 0xe51e1566, 0x77b020ba, -445 }, + { 0x915e2486, 0xef32cd60, 0x0ace1474, -441 }, + { 0xb5b5ada8, 0xaaff80b8, 0x0d819992, -438 }, + { 0xe3231912, 0xd5bf60e6, 0x10e1fff6, -435 }, + { 0x8df5efab, 0xc5979c8f, 0xca8d3ffa, -431 }, + { 0xb1736b96, 0xb6fd83b3, 0xbd308ff8, -428 }, + { 0xddd0467c, 0x64bce4a0, 0xac7cb3f6, -425 }, + { 0x8aa22c0d, 0xbef60ee4, 0x6bcdf07a, -421 }, + { 0xad4ab711, 0x2eb3929d, 0x86c16c98, -418 }, + { 0xd89d64d5, 0x7a607744, 0xe871c7bf, -415 }, + { 0x87625f05, 0x6c7c4a8b, 0x11471cd7, -411 }, + { 0xa93af6c6, 0xc79b5d2d, 0xd598e40d, -408 }, + { 0xd389b478, 0x79823479, 0x4aff1d10, -405 }, + { 0x843610cb, 0x4bf160cb, 0xcedf722a, -401 }, + { 0xa54394fe, 0x1eedb8fe, 0xc2974eb4, -398 }, + { 0xce947a3d, 0xa6a9273e, 0x733d2262, -395 }, + { 0x811ccc66, 0x8829b887, 0x0806357d, -391 }, + { 0xa163ff80, 0x2a3426a8, 0xca07c2dc, -388 }, + { 0xc9bcff60, 0x34c13052, 0xfc89b393, -385 }, + { 0xfc2c3f38, 0x41f17c67, 0xbbac2078, -382 }, + { 0x9d9ba783, 0x2936edc0, 0xd54b944b, -378 }, + { 0xc5029163, 0xf384a931, 0x0a9e795e, -375 }, + { 0xf64335bc, 0xf065d37d, 0x4d4617b5, -372 }, + { 0x99ea0196, 0x163fa42e, 0x504bced1, -368 }, + { 0xc06481fb, 0x9bcf8d39, 0xe45ec286, -365 }, + { 0xf07da27a, 0x82c37088, 0x5d767327, -362 }, + { 0x964e858c, 0x91ba2655, 0x3a6a07f8, -358 }, + { 0xbbe226ef, 0xb628afea, 0x890489f7, -355 }, + { 0xeadab0ab, 0xa3b2dbe5, 0x2b45ac74, -352 }, + { 0x92c8ae6b, 0x464fc96f, 0x3b0b8bc9, -348 }, + { 0xb77ada06, 0x17e3bbcb, 0x09ce6ebb, -345 }, + { 0xe5599087, 0x9ddcaabd, 0xcc420a6a, -342 }, + { 0x8f57fa54, 0xc2a9eab6, 0x9fa94682, -338 }, + { 0xb32df8e9, 0xf3546564, 0x47939822, -335 }, + { 0xdff97724, 0x70297ebd, 0x59787e2b, -332 }, + { 0x8bfbea76, 0xc619ef36, 0x57eb4edb, -328 }, + { 0xaefae514, 0x77a06b03, 0xede62292, -325 }, + { 0xdab99e59, 0x958885c4, 0xe95fab36, -322 }, + { 0x88b402f7, 0xfd75539b, 0x11dbcb02, -318 }, + { 0xaae103b5, 0xfcd2a881, 0xd652bdc2, -315 }, + { 0xd59944a3, 0x7c0752a2, 0x4be76d33, -312 }, + { 0x857fcae6, 0x2d8493a5, 0x6f70a440, -308 }, + { 0xa6dfbd9f, 0xb8e5b88e, 0xcb4ccd50, -305 }, + { 0xd097ad07, 0xa71f26b2, 0x7e2000a4, -302 }, + { 0x825ecc24, 0xc873782f, 0x8ed40066, -298 }, + { 0xa2f67f2d, 0xfa90563b, 0x72890080, -295 }, + { 0xcbb41ef9, 0x79346bca, 0x4f2b40a0, -292 }, + { 0xfea126b7, 0xd78186bc, 0xe2f610c8, -289 }, + { 0x9f24b832, 0xe6b0f436, 0x0dd9ca7d, -285 }, + { 0xc6ede63f, 0xa05d3143, 0x91503d1c, -282 }, + { 0xf8a95fcf, 0x88747d94, 0x75a44c63, -279 }, + { 0x9b69dbe1, 0xb548ce7c, 0xc986afbe, -275 }, + { 0xc24452da, 0x229b021b, 0xfbe85bad, -272 }, + { 0xf2d56790, 0xab41c2a2, 0xfae27299, -269 }, + { 0x97c560ba, 0x6b0919a5, 0xdccd879f, -265 }, + { 0xbdb6b8e9, 0x05cb600f, 0x5400e987, -262 }, + { 0xed246723, 0x473e3813, 0x290123e9, -259 }, + { 0x9436c076, 0x0c86e30b, 0xf9a0b672, -255 }, + { 0xb9447093, 0x8fa89bce, 0xf808e40e, -252 }, + { 0xe7958cb8, 0x7392c2c2, 0xb60b1d12, -249 }, + { 0x90bd77f3, 0x483bb9b9, 0xb1c6f22b, -245 }, + { 0xb4ecd5f0, 0x1a4aa828, 0x1e38aeb6, -242 }, + { 0xe2280b6c, 0x20dd5232, 0x25c6da63, -239 }, + { 0x8d590723, 0x948a535f, 0x579c487e, -235 }, + { 0xb0af48ec, 0x79ace837, 0x2d835a9d, -232 }, + { 0xdcdb1b27, 0x98182244, 0xf8e43145, -229 }, + { 0x8a08f0f8, 0xbf0f156b, 0x1b8e9ecb, -225 }, + { 0xac8b2d36, 0xeed2dac5, 0xe272467e, -222 }, + { 0xd7adf884, 0xaa879177, 0x5b0ed81d, -219 }, + { 0x86ccbb52, 0xea94baea, 0x98e94712, -215 }, + { 0xa87fea27, 0xa539e9a5, 0x3f2398d7, -212 }, + { 0xd29fe4b1, 0x8e88640e, 0x8eec7f0d, -209 }, + { 0x83a3eeee, 0xf9153e89, 0x1953cf68, -205 }, + { 0xa48ceaaa, 0xb75a8e2b, 0x5fa8c342, -202 }, + { 0xcdb02555, 0x653131b6, 0x3792f412, -199 }, + { 0x808e1755, 0x5f3ebf11, 0xe2bbd88b, -195 }, + { 0xa0b19d2a, 0xb70e6ed6, 0x5b6aceae, -192 }, + { 0xc8de0475, 0x64d20a8b, 0xf245825a, -189 }, + { 0xfb158592, 0xbe068d2e, 0xeed6e2f0, -186 }, + { 0x9ced737b, 0xb6c4183d, 0x55464dd6, -182 }, + { 0xc428d05a, 0xa4751e4c, 0xaa97e14c, -179 }, + { 0xf5330471, 0x4d9265df, 0xd53dd99f, -176 }, + { 0x993fe2c6, 0xd07b7fab, 0xe546a803, -172 }, + { 0xbf8fdb78, 0x849a5f96, 0xde985204, -169 }, + { 0xef73d256, 0xa5c0f77c, 0x963e6685, -166 }, + { 0x95a86376, 0x27989aad, 0xdde70013, -162 }, + { 0xbb127c53, 0xb17ec159, 0x5560c018, -159 }, + { 0xe9d71b68, 0x9dde71af, 0xaab8f01e, -156 }, + { 0x92267121, 0x62ab070d, 0xcab39613, -152 }, + { 0xb6b00d69, 0xbb55c8d1, 0x3d607b97, -149 }, + { 0xe45c10c4, 0x2a2b3b05, 0x8cb89a7d, -146 }, + { 0x8eb98a7a, 0x9a5b04e3, 0x77f3608e, -142 }, + { 0xb267ed19, 0x40f1c61c, 0x55f038b2, -139 }, + { 0xdf01e85f, 0x912e37a3, 0x6b6c46de, -136 }, + { 0x8b61313b, 0xbabce2c6, 0x2323ac4b, -132 }, + { 0xae397d8a, 0xa96c1b77, 0xabec975e, -129 }, + { 0xd9c7dced, 0x53c72255, 0x96e7bd35, -126 }, + { 0x881cea14, 0x545c7575, 0x7e50d641, -122 }, + { 0xaa242499, 0x697392d2, 0xdde50bd1, -119 }, + { 0xd4ad2dbf, 0xc3d07787, 0x955e4ec6, -116 }, + { 0x84ec3c97, 0xda624ab4, 0xbd5af13b, -112 }, + { 0xa6274bbd, 0xd0fadd61, 0xecb1ad8a, -109 }, + { 0xcfb11ead, 0x453994ba, 0x67de18ed, -106 }, + { 0x81ceb32c, 0x4b43fcf4, 0x80eacf94, -102 }, + { 0xa2425ff7, 0x5e14fc31, 0xa1258379, -99 }, + { 0xcad2f7f5, 0x359a3b3e, 0x096ee458, -96 }, + { 0xfd87b5f2, 0x8300ca0d, 0x8bca9d6e, -93 }, + { 0x9e74d1b7, 0x91e07e48, 0x775ea264, -89 }, + { 0xc6120625, 0x76589dda, 0x95364afe, -86 }, + { 0xf79687ae, 0xd3eec551, 0x3a83ddbd, -83 }, + { 0x9abe14cd, 0x44753b52, 0xc4926a96, -79 }, + { 0xc16d9a00, 0x95928a27, 0x75b7053c, -76 }, + { 0xf1c90080, 0xbaf72cb1, 0x5324c68b, -73 }, + { 0x971da050, 0x74da7bee, 0xd3f6fc16, -69 }, + { 0xbce50864, 0x92111aea, 0x88f4bb1c, -66 }, + { 0xec1e4a7d, 0xb69561a5, 0x2b31e9e3, -63 }, + { 0x9392ee8e, 0x921d5d07, 0x3aff322e, -59 }, + { 0xb877aa32, 0x36a4b449, 0x09befeb9, -56 }, + { 0xe69594be, 0xc44de15b, 0x4c2ebe68, -53 }, + { 0x901d7cf7, 0x3ab0acd9, 0x0f9d3701, -49 }, + { 0xb424dc35, 0x095cd80f, 0x538484c1, -46 }, + { 0xe12e1342, 0x4bb40e13, 0x2865a5f2, -43 }, + { 0x8cbccc09, 0x6f5088cb, 0xf93f87b7, -39 }, + { 0xafebff0b, 0xcb24aafe, 0xf78f69a5, -36 }, + { 0xdbe6fece, 0xbdedd5be, 0xb573440e, -33 }, + { 0x89705f41, 0x36b4a597, 0x31680a88, -29 }, + { 0xabcc7711, 0x8461cefc, 0xfdc20d2b, -26 }, + { 0xd6bf94d5, 0xe57a42bc, 0x3d329076, -23 }, + { 0x8637bd05, 0xaf6c69b5, 0xa63f9a49, -19 }, + { 0xa7c5ac47, 0x1b478423, 0x0fcf80dc, -16 }, + { 0xd1b71758, 0xe219652b, 0xd3c36113, -13 }, + { 0x83126e97, 0x8d4fdf3b, 0x645a1cac, -9 }, + { 0xa3d70a3d, 0x70a3d70a, 0x3d70a3d7, -6 }, + { 0xcccccccc, 0xcccccccc, 0xcccccccc, -3 }, + { 0x80000000, 0x00000000, 0x00000000, 1 }, + { 0xa0000000, 0x00000000, 0x00000000, 4 }, + { 0xc8000000, 0x00000000, 0x00000000, 7 }, + { 0xfa000000, 0x00000000, 0x00000000, 10 }, + { 0x9c400000, 0x00000000, 0x00000000, 14 }, + { 0xc3500000, 0x00000000, 0x00000000, 17 }, + { 0xf4240000, 0x00000000, 0x00000000, 20 }, + { 0x98968000, 0x00000000, 0x00000000, 24 }, + { 0xbebc2000, 0x00000000, 0x00000000, 27 }, + { 0xee6b2800, 0x00000000, 0x00000000, 30 }, + { 0x9502f900, 0x00000000, 0x00000000, 34 }, + { 0xba43b740, 0x00000000, 0x00000000, 37 }, + { 0xe8d4a510, 0x00000000, 0x00000000, 40 }, + { 0x9184e72a, 0x00000000, 0x00000000, 44 }, + { 0xb5e620f4, 0x80000000, 0x00000000, 47 }, + { 0xe35fa931, 0xa0000000, 0x00000000, 50 }, + { 0x8e1bc9bf, 0x04000000, 0x00000000, 54 }, + { 0xb1a2bc2e, 0xc5000000, 0x00000000, 57 }, + { 0xde0b6b3a, 0x76400000, 0x00000000, 60 }, + { 0x8ac72304, 0x89e80000, 0x00000000, 64 }, + { 0xad78ebc5, 0xac620000, 0x00000000, 67 }, + { 0xd8d726b7, 0x177a8000, 0x00000000, 70 }, + { 0x87867832, 0x6eac9000, 0x00000000, 74 }, + { 0xa968163f, 0x0a57b400, 0x00000000, 77 }, + { 0xd3c21bce, 0xcceda100, 0x00000000, 80 }, + { 0x84595161, 0x401484a0, 0x00000000, 84 }, + { 0xa56fa5b9, 0x9019a5c8, 0x00000000, 87 }, + { 0xcecb8f27, 0xf4200f3a, 0x00000000, 90 }, + { 0x813f3978, 0xf8940984, 0x40000000, 94 }, + { 0xa18f07d7, 0x36b90be5, 0x50000000, 97 }, + { 0xc9f2c9cd, 0x04674ede, 0xa4000000, 100 }, + { 0xfc6f7c40, 0x45812296, 0x4d000000, 103 }, + { 0x9dc5ada8, 0x2b70b59d, 0xf0200000, 107 }, + { 0xc5371912, 0x364ce305, 0x6c280000, 110 }, + { 0xf684df56, 0xc3e01bc6, 0xc7320000, 113 }, + { 0x9a130b96, 0x3a6c115c, 0x3c7f4000, 117 }, + { 0xc097ce7b, 0xc90715b3, 0x4b9f1000, 120 }, + { 0xf0bdc21a, 0xbb48db20, 0x1e86d400, 123 }, + { 0x96769950, 0xb50d88f4, 0x13144480, 127 }, + { 0xbc143fa4, 0xe250eb31, 0x17d955a0, 130 }, + { 0xeb194f8e, 0x1ae525fd, 0x5dcfab08, 133 }, + { 0x92efd1b8, 0xd0cf37be, 0x5aa1cae5, 137 }, + { 0xb7abc627, 0x050305ad, 0xf14a3d9e, 140 }, + { 0xe596b7b0, 0xc643c719, 0x6d9ccd05, 143 }, + { 0x8f7e32ce, 0x7bea5c6f, 0xe4820023, 147 }, + { 0xb35dbf82, 0x1ae4f38b, 0xdda2802c, 150 }, + { 0xe0352f62, 0xa19e306e, 0xd50b2037, 153 }, + { 0x8c213d9d, 0xa502de45, 0x4526f422, 157 }, + { 0xaf298d05, 0x0e4395d6, 0x9670b12b, 160 }, + { 0xdaf3f046, 0x51d47b4c, 0x3c0cdd76, 163 }, + { 0x88d8762b, 0xf324cd0f, 0xa5880a69, 167 }, + { 0xab0e93b6, 0xefee0053, 0x8eea0d04, 170 }, + { 0xd5d238a4, 0xabe98068, 0x72a49045, 173 }, + { 0x85a36366, 0xeb71f041, 0x47a6da2b, 177 }, + { 0xa70c3c40, 0xa64e6c51, 0x999090b6, 180 }, + { 0xd0cf4b50, 0xcfe20765, 0xfff4b4e3, 183 }, + { 0x82818f12, 0x81ed449f, 0xbff8f10e, 187 }, + { 0xa321f2d7, 0x226895c7, 0xaff72d52, 190 }, + { 0xcbea6f8c, 0xeb02bb39, 0x9bf4f8a6, 193 }, + { 0xfee50b70, 0x25c36a08, 0x02f236d0, 196 }, + { 0x9f4f2726, 0x179a2245, 0x01d76242, 200 }, + { 0xc722f0ef, 0x9d80aad6, 0x424d3ad2, 203 }, + { 0xf8ebad2b, 0x84e0d58b, 0xd2e08987, 206 }, + { 0x9b934c3b, 0x330c8577, 0x63cc55f4, 210 }, + { 0xc2781f49, 0xffcfa6d5, 0x3cbf6b71, 213 }, + { 0xf316271c, 0x7fc3908a, 0x8bef464e, 216 }, + { 0x97edd871, 0xcfda3a56, 0x97758bf0, 220 }, + { 0xbde94e8e, 0x43d0c8ec, 0x3d52eeed, 223 }, + { 0xed63a231, 0xd4c4fb27, 0x4ca7aaa8, 226 }, + { 0x945e455f, 0x24fb1cf8, 0x8fe8caa9, 230 }, + { 0xb975d6b6, 0xee39e436, 0xb3e2fd53, 233 }, + { 0xe7d34c64, 0xa9c85d44, 0x60dbbca8, 236 }, + { 0x90e40fbe, 0xea1d3a4a, 0xbc8955e9, 240 }, + { 0xb51d13ae, 0xa4a488dd, 0x6babab63, 243 }, + { 0xe264589a, 0x4dcdab14, 0xc696963c, 246 }, + { 0x8d7eb760, 0x70a08aec, 0xfc1e1de5, 250 }, + { 0xb0de6538, 0x8cc8ada8, 0x3b25a55f, 253 }, + { 0xdd15fe86, 0xaffad912, 0x49ef0eb7, 256 }, + { 0x8a2dbf14, 0x2dfcc7ab, 0x6e356932, 260 }, + { 0xacb92ed9, 0x397bf996, 0x49c2c37f, 263 }, + { 0xd7e77a8f, 0x87daf7fb, 0xdc33745e, 266 }, + { 0x86f0ac99, 0xb4e8dafd, 0x69a028bb, 270 }, + { 0xa8acd7c0, 0x222311bc, 0xc40832ea, 273 }, + { 0xd2d80db0, 0x2aabd62b, 0xf50a3fa4, 276 }, + { 0x83c7088e, 0x1aab65db, 0x792667c6, 280 }, + { 0xa4b8cab1, 0xa1563f52, 0x577001b8, 283 }, + { 0xcde6fd5e, 0x09abcf26, 0xed4c0226, 286 }, + { 0x80b05e5a, 0xc60b6178, 0x544f8158, 290 }, + { 0xa0dc75f1, 0x778e39d6, 0x696361ae, 293 }, + { 0xc913936d, 0xd571c84c, 0x03bc3a19, 296 }, + { 0xfb587849, 0x4ace3a5f, 0x04ab48a0, 299 }, + { 0x9d174b2d, 0xcec0e47b, 0x62eb0d64, 303 }, + { 0xc45d1df9, 0x42711d9a, 0x3ba5d0bd, 306 }, + { 0xf5746577, 0x930d6500, 0xca8f44ec, 309 }, + { 0x9968bf6a, 0xbbe85f20, 0x7e998b13, 313 }, + { 0xbfc2ef45, 0x6ae276e8, 0x9e3fedd8, 316 }, + { 0xefb3ab16, 0xc59b14a2, 0xc5cfe94e, 319 }, + { 0x95d04aee, 0x3b80ece5, 0xbba1f1d1, 323 }, + { 0xbb445da9, 0xca61281f, 0x2a8a6e45, 326 }, + { 0xea157514, 0x3cf97226, 0xf52d09d7, 329 }, + { 0x924d692c, 0xa61be758, 0x593c2626, 333 }, + { 0xb6e0c377, 0xcfa2e12e, 0x6f8b2fb0, 336 }, + { 0xe498f455, 0xc38b997a, 0x0b6dfb9c, 339 }, + { 0x8edf98b5, 0x9a373fec, 0x4724bd41, 343 }, + { 0xb2977ee3, 0x00c50fe7, 0x58edec91, 346 }, + { 0xdf3d5e9b, 0xc0f653e1, 0x2f2967b6, 349 }, + { 0x8b865b21, 0x5899f46c, 0xbd79e0d2, 353 }, + { 0xae67f1e9, 0xaec07187, 0xecd85906, 356 }, + { 0xda01ee64, 0x1a708de9, 0xe80e6f48, 359 }, + { 0x884134fe, 0x908658b2, 0x3109058d, 363 }, + { 0xaa51823e, 0x34a7eede, 0xbd4b46f0, 366 }, + { 0xd4e5e2cd, 0xc1d1ea96, 0x6c9e18ac, 369 }, + { 0x850fadc0, 0x9923329e, 0x03e2cf6b, 373 }, + { 0xa6539930, 0xbf6bff45, 0x84db8346, 376 }, + { 0xcfe87f7c, 0xef46ff16, 0xe6126418, 379 }, + { 0x81f14fae, 0x158c5f6e, 0x4fcb7e8f, 383 }, + { 0xa26da399, 0x9aef7749, 0xe3be5e33, 386 }, + { 0xcb090c80, 0x01ab551c, 0x5cadf5bf, 389 }, + { 0xfdcb4fa0, 0x02162a63, 0x73d9732f, 392 }, + { 0x9e9f11c4, 0x014dda7e, 0x2867e7fd, 396 }, + { 0xc646d635, 0x01a1511d, 0xb281e1fd, 399 }, + { 0xf7d88bc2, 0x4209a565, 0x1f225a7c, 402 }, + { 0x9ae75759, 0x6946075f, 0x3375788d, 406 }, + { 0xc1a12d2f, 0xc3978937, 0x0052d6b1, 409 }, + { 0xf209787b, 0xb47d6b84, 0xc0678c5d, 412 }, + { 0x9745eb4d, 0x50ce6332, 0xf840b7ba, 416 }, + { 0xbd176620, 0xa501fbff, 0xb650e5a9, 419 }, + { 0xec5d3fa8, 0xce427aff, 0xa3e51f13, 422 }, + { 0x93ba47c9, 0x80e98cdf, 0xc66f336c, 426 }, + { 0xb8a8d9bb, 0xe123f017, 0xb80b0047, 429 }, + { 0xe6d3102a, 0xd96cec1d, 0xa60dc059, 432 }, + { 0x9043ea1a, 0xc7e41392, 0x87c89837, 436 }, + { 0xb454e4a1, 0x79dd1877, 0x29babe45, 439 }, + { 0xe16a1dc9, 0xd8545e94, 0xf4296dd6, 442 }, + { 0x8ce2529e, 0x2734bb1d, 0x1899e4a6, 446 }, + { 0xb01ae745, 0xb101e9e4, 0x5ec05dcf, 449 }, + { 0xdc21a117, 0x1d42645d, 0x76707543, 452 }, + { 0x899504ae, 0x72497eba, 0x6a06494a, 456 }, + { 0xabfa45da, 0x0edbde69, 0x0487db9d, 459 }, + { 0xd6f8d750, 0x9292d603, 0x45a9d284, 462 }, + { 0x865b8692, 0x5b9bc5c2, 0x0b8a2392, 466 }, + { 0xa7f26836, 0xf282b732, 0x8e6cac77, 469 }, + { 0xd1ef0244, 0xaf2364ff, 0x3207d795, 472 }, + { 0x8335616a, 0xed761f1f, 0x7f44e6bd, 476 }, + { 0xa402b9c5, 0xa8d3a6e7, 0x5f16206c, 479 }, + { 0xcd036837, 0x130890a1, 0x36dba887, 482 }, + { 0x80222122, 0x6be55a64, 0xc2494954, 486 }, + { 0xa02aa96b, 0x06deb0fd, 0xf2db9baa, 489 }, + { 0xc83553c5, 0xc8965d3d, 0x6f928294, 492 }, + { 0xfa42a8b7, 0x3abbf48c, 0xcb772339, 495 }, + { 0x9c69a972, 0x84b578d7, 0xff2a7604, 499 }, + { 0xc38413cf, 0x25e2d70d, 0xfef51385, 502 }, + { 0xf46518c2, 0xef5b8cd1, 0x7eb25866, 505 }, + { 0x98bf2f79, 0xd5993802, 0xef2f773f, 509 }, + { 0xbeeefb58, 0x4aff8603, 0xaafb550f, 512 }, + { 0xeeaaba2e, 0x5dbf6784, 0x95ba2a53, 515 }, + { 0x952ab45c, 0xfa97a0b2, 0xdd945a74, 519 }, + { 0xba756174, 0x393d88df, 0x94f97111, 522 }, + { 0xe912b9d1, 0x478ceb17, 0x7a37cd56, 525 }, + { 0x91abb422, 0xccb812ee, 0xac62e055, 529 }, + { 0xb616a12b, 0x7fe617aa, 0x577b986b, 532 }, + { 0xe39c4976, 0x5fdf9d94, 0xed5a7e85, 535 }, + { 0x8e41ade9, 0xfbebc27d, 0x14588f13, 539 }, + { 0xb1d21964, 0x7ae6b31c, 0x596eb2d8, 542 }, + { 0xde469fbd, 0x99a05fe3, 0x6fca5f8e, 545 }, + { 0x8aec23d6, 0x80043bee, 0x25de7bb9, 549 }, + { 0xada72ccc, 0x20054ae9, 0xaf561aa7, 552 }, + { 0xd910f7ff, 0x28069da4, 0x1b2ba151, 555 }, + { 0x87aa9aff, 0x79042286, 0x90fb44d2, 559 }, + { 0xa99541bf, 0x57452b28, 0x353a1607, 562 }, + { 0xd3fa922f, 0x2d1675f2, 0x42889b89, 565 }, + { 0x847c9b5d, 0x7c2e09b7, 0x69956135, 569 }, + { 0xa59bc234, 0xdb398c25, 0x43fab983, 572 }, + { 0xcf02b2c2, 0x1207ef2e, 0x94f967e4, 575 }, + { 0x8161afb9, 0x4b44f57d, 0x1d1be0ee, 579 }, + { 0xa1ba1ba7, 0x9e1632dc, 0x6462d92a, 582 }, + { 0xca28a291, 0x859bbf93, 0x7d7b8f75, 585 }, + { 0xfcb2cb35, 0xe702af78, 0x5cda7352, 588 }, + { 0x9defbf01, 0xb061adab, 0x3a088813, 592 }, + { 0xc56baec2, 0x1c7a1916, 0x088aaa18, 595 }, + { 0xf6c69a72, 0xa3989f5b, 0x8aad549e, 598 }, + { 0x9a3c2087, 0xa63f6399, 0x36ac54e2, 602 }, + { 0xc0cb28a9, 0x8fcf3c7f, 0x84576a1b, 605 }, + { 0xf0fdf2d3, 0xf3c30b9f, 0x656d44a2, 608 }, + { 0x969eb7c4, 0x7859e743, 0x9f644ae5, 612 }, + { 0xbc4665b5, 0x96706114, 0x873d5d9f, 615 }, + { 0xeb57ff22, 0xfc0c7959, 0xa90cb506, 618 }, + { 0x9316ff75, 0xdd87cbd8, 0x09a7f124, 622 }, + { 0xb7dcbf53, 0x54e9bece, 0x0c11ed6d, 625 }, + { 0xe5d3ef28, 0x2a242e81, 0x8f1668c8, 628 }, + { 0x8fa47579, 0x1a569d10, 0xf96e017d, 632 }, + { 0xb38d92d7, 0x60ec4455, 0x37c981dc, 635 }, + { 0xe070f78d, 0x3927556a, 0x85bbe253, 638 }, + { 0x8c469ab8, 0x43b89562, 0x93956d74, 642 }, + { 0xaf584166, 0x54a6babb, 0x387ac8d1, 645 }, + { 0xdb2e51bf, 0xe9d0696a, 0x06997b05, 648 }, + { 0x88fcf317, 0xf22241e2, 0x441fece3, 652 }, + { 0xab3c2fdd, 0xeeaad25a, 0xd527e81c, 655 }, + { 0xd60b3bd5, 0x6a5586f1, 0x8a71e223, 658 }, + { 0x85c70565, 0x62757456, 0xf6872d56, 662 }, + { 0xa738c6be, 0xbb12d16c, 0xb428f8ac, 665 }, + { 0xd106f86e, 0x69d785c7, 0xe13336d7, 668 }, + { 0x82a45b45, 0x0226b39c, 0xecc00246, 672 }, + { 0xa34d7216, 0x42b06084, 0x27f002d7, 675 }, + { 0xcc20ce9b, 0xd35c78a5, 0x31ec038d, 678 }, + { 0xff290242, 0xc83396ce, 0x7e670471, 681 }, + { 0x9f79a169, 0xbd203e41, 0x0f0062c6, 685 }, + { 0xc75809c4, 0x2c684dd1, 0x52c07b78, 688 }, + { 0xf92e0c35, 0x37826145, 0xa7709a56, 691 }, + { 0x9bbcc7a1, 0x42b17ccb, 0x88a66076, 695 }, + { 0xc2abf989, 0x935ddbfe, 0x6acff893, 698 }, + { 0xf356f7eb, 0xf83552fe, 0x0583f6b8, 701 }, + { 0x98165af3, 0x7b2153de, 0xc3727a33, 705 }, + { 0xbe1bf1b0, 0x59e9a8d6, 0x744f18c0, 708 }, + { 0xeda2ee1c, 0x7064130c, 0x1162def0, 711 }, + { 0x9485d4d1, 0xc63e8be7, 0x8addcb56, 715 }, + { 0xb9a74a06, 0x37ce2ee1, 0x6d953e2b, 718 }, + { 0xe8111c87, 0xc5c1ba99, 0xc8fa8db6, 721 }, + { 0x910ab1d4, 0xdb9914a0, 0x1d9c9892, 725 }, + { 0xb54d5e4a, 0x127f59c8, 0x2503beb6, 728 }, + { 0xe2a0b5dc, 0x971f303a, 0x2e44ae64, 731 }, + { 0x8da471a9, 0xde737e24, 0x5ceaecfe, 735 }, + { 0xb10d8e14, 0x56105dad, 0x7425a83e, 738 }, + { 0xdd50f199, 0x6b947518, 0xd12f124e, 741 }, + { 0x8a5296ff, 0xe33cc92f, 0x82bd6b70, 745 }, + { 0xace73cbf, 0xdc0bfb7b, 0x636cc64d, 748 }, + { 0xd8210bef, 0xd30efa5a, 0x3c47f7e0, 751 }, + { 0x8714a775, 0xe3e95c78, 0x65acfaec, 755 }, + { 0xa8d9d153, 0x5ce3b396, 0x7f1839a7, 758 }, + { 0xd31045a8, 0x341ca07c, 0x1ede4811, 761 }, + { 0x83ea2b89, 0x2091e44d, 0x934aed0a, 765 }, + { 0xa4e4b66b, 0x68b65d60, 0xf81da84d, 768 }, + { 0xce1de406, 0x42e3f4b9, 0x36251260, 771 }, + { 0x80d2ae83, 0xe9ce78f3, 0xc1d72b7c, 775 }, + { 0xa1075a24, 0xe4421730, 0xb24cf65b, 778 }, + { 0xc94930ae, 0x1d529cfc, 0xdee033f2, 781 }, + { 0xfb9b7cd9, 0xa4a7443c, 0x169840ef, 784 }, + { 0x9d412e08, 0x06e88aa5, 0x8e1f2895, 788 }, + { 0xc491798a, 0x08a2ad4e, 0xf1a6f2ba, 791 }, + { 0xf5b5d7ec, 0x8acb58a2, 0xae10af69, 794 }, + { 0x9991a6f3, 0xd6bf1765, 0xacca6da1, 798 }, + { 0xbff610b0, 0xcc6edd3f, 0x17fd090a, 801 }, + { 0xeff394dc, 0xff8a948e, 0xddfc4b4c, 804 }, + { 0x95f83d0a, 0x1fb69cd9, 0x4abdaf10, 808 }, + { 0xbb764c4c, 0xa7a4440f, 0x9d6d1ad4, 811 }, + { 0xea53df5f, 0xd18d5513, 0x84c86189, 814 }, + { 0x92746b9b, 0xe2f8552c, 0x32fd3cf5, 818 }, + { 0xb7118682, 0xdbb66a77, 0x3fbc8c33, 821 }, + { 0xe4d5e823, 0x92a40515, 0x0fabaf3f, 824 }, + { 0x8f05b116, 0x3ba6832d, 0x29cb4d87, 828 }, + { 0xb2c71d5b, 0xca9023f8, 0x743e20e9, 831 }, + { 0xdf78e4b2, 0xbd342cf6, 0x914da924, 834 }, + { 0x8bab8eef, 0xb6409c1a, 0x1ad089b6, 838 }, + { 0xae9672ab, 0xa3d0c320, 0xa184ac24, 841 }, + { 0xda3c0f56, 0x8cc4f3e8, 0xc9e5d72d, 844 }, + { 0x88658996, 0x17fb1871, 0x7e2fa67c, 848 }, + { 0xaa7eebfb, 0x9df9de8d, 0xddbb901b, 851 }, + { 0xd51ea6fa, 0x85785631, 0x552a7422, 854 }, + { 0x8533285c, 0x936b35de, 0xd53a8895, 858 }, + { 0xa67ff273, 0xb8460356, 0x8a892aba, 861 }, + { 0xd01fef10, 0xa657842c, 0x2d2b7569, 864 }, + { 0x8213f56a, 0x67f6b29b, 0x9c3b2962, 868 }, + { 0xa298f2c5, 0x01f45f42, 0x8349f3ba, 871 }, + { 0xcb3f2f76, 0x42717713, 0x241c70a9, 874 }, + { 0xfe0efb53, 0xd30dd4d7, 0xed238cd3, 877 }, + { 0x9ec95d14, 0x63e8a506, 0xf4363804, 881 }, + { 0xc67bb459, 0x7ce2ce48, 0xb143c605, 884 }, + { 0xf81aa16f, 0xdc1b81da, 0xdd94b786, 887 }, + { 0x9b10a4e5, 0xe9913128, 0xca7cf2b4, 891 }, + { 0xc1d4ce1f, 0x63f57d72, 0xfd1c2f61, 894 }, + { 0xf24a01a7, 0x3cf2dccf, 0xbc633b39, 897 }, + { 0x976e4108, 0x8617ca01, 0xd5be0503, 901 }, + { 0xbd49d14a, 0xa79dbc82, 0x4b2d8644, 904 }, + { 0xec9c459d, 0x51852ba2, 0xddf8e7d6, 907 }, + { 0x93e1ab82, 0x52f33b45, 0xcabb90e5, 911 }, + { 0xb8da1662, 0xe7b00a17, 0x3d6a751f, 914 }, + { 0xe7109bfb, 0xa19c0c9d, 0x0cc51267, 917 }, + { 0x906a617d, 0x450187e2, 0x27fb2b80, 921 }, + { 0xb484f9dc, 0x9641e9da, 0xb1f9f660, 924 }, + { 0xe1a63853, 0xbbd26451, 0x5e7873f8, 927 }, + { 0x8d07e334, 0x55637eb2, 0xdb0b487b, 931 }, + { 0xb049dc01, 0x6abc5e5f, 0x91ce1a9a, 934 }, + { 0xdc5c5301, 0xc56b75f7, 0x7641a140, 937 }, + { 0x89b9b3e1, 0x1b6329ba, 0xa9e904c8, 941 }, + { 0xac2820d9, 0x623bf429, 0x546345fa, 944 }, + { 0xd732290f, 0xbacaf133, 0xa97c1779, 947 }, + { 0x867f59a9, 0xd4bed6c0, 0x49ed8eab, 951 }, + { 0xa81f3014, 0x49ee8c70, 0x5c68f256, 954 }, + { 0xd226fc19, 0x5c6a2f8c, 0x73832eec, 957 }, + { 0x83585d8f, 0xd9c25db7, 0xc831fd53, 961 }, + { 0xa42e74f3, 0xd032f525, 0xba3e7ca8, 964 }, + { 0xcd3a1230, 0xc43fb26f, 0x28ce1bd2, 967 }, + { 0x80444b5e, 0x7aa7cf85, 0x7980d163, 971 }, + { 0xa0555e36, 0x1951c366, 0xd7e105bc, 974 }, + { 0xc86ab5c3, 0x9fa63440, 0x8dd9472b, 977 }, + { 0xfa856334, 0x878fc150, 0xb14f98f6, 980 }, + { 0x9c935e00, 0xd4b9d8d2, 0x6ed1bf9a, 984 }, + { 0xc3b83581, 0x09e84f07, 0x0a862f80, 987 }, + { 0xf4a642e1, 0x4c6262c8, 0xcd27bb61, 990 }, + { 0x98e7e9cc, 0xcfbd7dbd, 0x8038d51c, 994 }, + { 0xbf21e440, 0x03acdd2c, 0xe0470a63, 997 }, + { 0xeeea5d50, 0x04981478, 0x1858ccfc, 1000 }, + { 0x95527a52, 0x02df0ccb, 0x0f37801e, 1004 }, + { 0xbaa718e6, 0x8396cffd, 0xd3056025, 1007 }, + { 0xe950df20, 0x247c83fd, 0x47c6b82e, 1010 }, + { 0x91d28b74, 0x16cdd27e, 0x4cdc331d, 1014 }, + { 0xb6472e51, 0x1c81471d, 0xe0133fe4, 1017 }, + { 0xe3d8f9e5, 0x63a198e5, 0x58180fdd, 1020 }, + { 0x8e679c2f, 0x5e44ff8f, 0x570f09ea, 1024 }, + { 0xb201833b, 0x35d63f73, 0x2cd2cc65, 1027 }, + { 0xde81e40a, 0x034bcf4f, 0xf8077f7e, 1030 }, + { 0x8b112e86, 0x420f6191, 0xfb04afaf, 1034 }, + { 0xadd57a27, 0xd29339f6, 0x79c5db9a, 1037 }, + { 0xd94ad8b1, 0xc7380874, 0x18375281, 1040 }, + { 0x87cec76f, 0x1c830548, 0x8f229391, 1044 }, + { 0xa9c2794a, 0xe3a3c69a, 0xb2eb3875, 1047 }, + { 0xd433179d, 0x9c8cb841, 0x5fa60692, 1050 }, + { 0x849feec2, 0x81d7f328, 0xdbc7c41b, 1054 }, + { 0xa5c7ea73, 0x224deff3, 0x12b9b522, 1057 }, + { 0xcf39e50f, 0xeae16bef, 0xd768226b, 1060 }, + { 0x81842f29, 0xf2cce375, 0xe6a11583, 1064 }, + { 0xa1e53af4, 0x6f801c53, 0x60495ae3, 1067 }, + { 0xca5e89b1, 0x8b602368, 0x385bb19c, 1070 }, + { 0xfcf62c1d, 0xee382c42, 0x46729e03, 1073 }, + { 0x9e19db92, 0xb4e31ba9, 0x6c07a2c2, 1077 } + }; + static short int Lhint[2098] = { + /*18,*/19, 19, 19, 19, 20, 20, 20, 21, 21, + 21, 22, 22, 22, 23, 23, 23, 23, 24, 24, + 24, 25, 25, 25, 26, 26, 26, 26, 27, 27, + 27, 28, 28, 28, 29, 29, 29, 29, 30, 30, + 30, 31, 31, 31, 32, 32, 32, 32, 33, 33, + 33, 34, 34, 34, 35, 35, 35, 35, 36, 36, + 36, 37, 37, 37, 38, 38, 38, 38, 39, 39, + 39, 40, 40, 40, 41, 41, 41, 41, 42, 42, + 42, 43, 43, 43, 44, 44, 44, 44, 45, 45, + 45, 46, 46, 46, 47, 47, 47, 47, 48, 48, + 48, 49, 49, 49, 50, 50, 50, 51, 51, 51, + 51, 52, 52, 52, 53, 53, 53, 54, 54, 54, + 54, 55, 55, 55, 56, 56, 56, 57, 57, 57, + 57, 58, 58, 58, 59, 59, 59, 60, 60, 60, + 60, 61, 61, 61, 62, 62, 62, 63, 63, 63, + 63, 64, 64, 64, 65, 65, 65, 66, 66, 66, + 66, 67, 67, 67, 68, 68, 68, 69, 69, 69, + 69, 70, 70, 70, 71, 71, 71, 72, 72, 72, + 72, 73, 73, 73, 74, 74, 74, 75, 75, 75, + 75, 76, 76, 76, 77, 77, 77, 78, 78, 78, + 78, 79, 79, 79, 80, 80, 80, 81, 81, 81, + 82, 82, 82, 82, 83, 83, 83, 84, 84, 84, + 85, 85, 85, 85, 86, 86, 86, 87, 87, 87, + 88, 88, 88, 88, 89, 89, 89, 90, 90, 90, + 91, 91, 91, 91, 92, 92, 92, 93, 93, 93, + 94, 94, 94, 94, 95, 95, 95, 96, 96, 96, + 97, 97, 97, 97, 98, 98, 98, 99, 99, 99, + 100, 100, 100, 100, 101, 101, 101, 102, 102, 102, + 103, 103, 103, 103, 104, 104, 104, 105, 105, 105, + 106, 106, 106, 106, 107, 107, 107, 108, 108, 108, + 109, 109, 109, 110, 110, 110, 110, 111, 111, 111, + 112, 112, 112, 113, 113, 113, 113, 114, 114, 114, + 115, 115, 115, 116, 116, 116, 116, 117, 117, 117, + 118, 118, 118, 119, 119, 119, 119, 120, 120, 120, + 121, 121, 121, 122, 122, 122, 122, 123, 123, 123, + 124, 124, 124, 125, 125, 125, 125, 126, 126, 126, + 127, 127, 127, 128, 128, 128, 128, 129, 129, 129, + 130, 130, 130, 131, 131, 131, 131, 132, 132, 132, + 133, 133, 133, 134, 134, 134, 134, 135, 135, 135, + 136, 136, 136, 137, 137, 137, 137, 138, 138, 138, + 139, 139, 139, 140, 140, 140, 141, 141, 141, 141, + 142, 142, 142, 143, 143, 143, 144, 144, 144, 144, + 145, 145, 145, 146, 146, 146, 147, 147, 147, 147, + 148, 148, 148, 149, 149, 149, 150, 150, 150, 150, + 151, 151, 151, 152, 152, 152, 153, 153, 153, 153, + 154, 154, 154, 155, 155, 155, 156, 156, 156, 156, + 157, 157, 157, 158, 158, 158, 159, 159, 159, 159, + 160, 160, 160, 161, 161, 161, 162, 162, 162, 162, + 163, 163, 163, 164, 164, 164, 165, 165, 165, 165, + 166, 166, 166, 167, 167, 167, 168, 168, 168, 169, + 169, 169, 169, 170, 170, 170, 171, 171, 171, 172, + 172, 172, 172, 173, 173, 173, 174, 174, 174, 175, + 175, 175, 175, 176, 176, 176, 177, 177, 177, 178, + 178, 178, 178, 179, 179, 179, 180, 180, 180, 181, + 181, 181, 181, 182, 182, 182, 183, 183, 183, 184, + 184, 184, 184, 185, 185, 185, 186, 186, 186, 187, + 187, 187, 187, 188, 188, 188, 189, 189, 189, 190, + 190, 190, 190, 191, 191, 191, 192, 192, 192, 193, + 193, 193, 193, 194, 194, 194, 195, 195, 195, 196, + 196, 196, 197, 197, 197, 197, 198, 198, 198, 199, + 199, 199, 200, 200, 200, 200, 201, 201, 201, 202, + 202, 202, 203, 203, 203, 203, 204, 204, 204, 205, + 205, 205, 206, 206, 206, 206, 207, 207, 207, 208, + 208, 208, 209, 209, 209, 209, 210, 210, 210, 211, + 211, 211, 212, 212, 212, 212, 213, 213, 213, 214, + 214, 214, 215, 215, 215, 215, 216, 216, 216, 217, + 217, 217, 218, 218, 218, 218, 219, 219, 219, 220, + 220, 220, 221, 221, 221, 221, 222, 222, 222, 223, + 223, 223, 224, 224, 224, 224, 225, 225, 225, 226, + 226, 226, 227, 227, 227, 228, 228, 228, 228, 229, + 229, 229, 230, 230, 230, 231, 231, 231, 231, 232, + 232, 232, 233, 233, 233, 234, 234, 234, 234, 235, + 235, 235, 236, 236, 236, 237, 237, 237, 237, 238, + 238, 238, 239, 239, 239, 240, 240, 240, 240, 241, + 241, 241, 242, 242, 242, 243, 243, 243, 243, 244, + 244, 244, 245, 245, 245, 246, 246, 246, 246, 247, + 247, 247, 248, 248, 248, 249, 249, 249, 249, 250, + 250, 250, 251, 251, 251, 252, 252, 252, 252, 253, + 253, 253, 254, 254, 254, 255, 255, 255, 256, 256, + 256, 256, 257, 257, 257, 258, 258, 258, 259, 259, + 259, 259, 260, 260, 260, 261, 261, 261, 262, 262, + 262, 262, 263, 263, 263, 264, 264, 264, 265, 265, + 265, 265, 266, 266, 266, 267, 267, 267, 268, 268, + 268, 268, 269, 269, 269, 270, 270, 270, 271, 271, + 271, 271, 272, 272, 272, 273, 273, 273, 274, 274, + 274, 274, 275, 275, 275, 276, 276, 276, 277, 277, + 277, 277, 278, 278, 278, 279, 279, 279, 280, 280, + 280, 280, 281, 281, 281, 282, 282, 282, 283, 283, + 283, 283, 284, 284, 284, 285, 285, 285, 286, 286, + 286, 287, 287, 287, 287, 288, 288, 288, 289, 289, + 289, 290, 290, 290, 290, 291, 291, 291, 292, 292, + 292, 293, 293, 293, 293, 294, 294, 294, 295, 295, + 295, 296, 296, 296, 296, 297, 297, 297, 298, 298, + 298, 299, 299, 299, 299, 300, 300, 300, 301, 301, + 301, 302, 302, 302, 302, 303, 303, 303, 304, 304, + 304, 305, 305, 305, 305, 306, 306, 306, 307, 307, + 307, 308, 308, 308, 308, 309, 309, 309, 310, 310, + 310, 311, 311, 311, 311, 312, 312, 312, 313, 313, + 313, 314, 314, 314, 315, 315, 315, 315, 316, 316, + 316, 317, 317, 317, 318, 318, 318, 318, 319, 319, + 319, 320, 320, 320, 321, 321, 321, 321, 322, 322, + 322, 323, 323, 323, 324, 324, 324, 324, 325, 325, + 325, 326, 326, 326, 327, 327, 327, 327, 328, 328, + 328, 329, 329, 329, 330, 330, 330, 330, 331, 331, + 331, 332, 332, 332, 333, 333, 333, 333, 334, 334, + 334, 335, 335, 335, 336, 336, 336, 336, 337, 337, + 337, 338, 338, 338, 339, 339, 339, 339, 340, 340, + 340, 341, 341, 341, 342, 342, 342, 342, 343, 343, + 343, 344, 344, 344, 345, 345, 345, 346, 346, 346, + 346, 347, 347, 347, 348, 348, 348, 349, 349, 349, + 349, 350, 350, 350, 351, 351, 351, 352, 352, 352, + 352, 353, 353, 353, 354, 354, 354, 355, 355, 355, + 355, 356, 356, 356, 357, 357, 357, 358, 358, 358, + 358, 359, 359, 359, 360, 360, 360, 361, 361, 361, + 361, 362, 362, 362, 363, 363, 363, 364, 364, 364, + 364, 365, 365, 365, 366, 366, 366, 367, 367, 367, + 367, 368, 368, 368, 369, 369, 369, 370, 370, 370, + 370, 371, 371, 371, 372, 372, 372, 373, 373, 373, + 374, 374, 374, 374, 375, 375, 375, 376, 376, 376, + 377, 377, 377, 377, 378, 378, 378, 379, 379, 379, + 380, 380, 380, 380, 381, 381, 381, 382, 382, 382, + 383, 383, 383, 383, 384, 384, 384, 385, 385, 385, + 386, 386, 386, 386, 387, 387, 387, 388, 388, 388, + 389, 389, 389, 389, 390, 390, 390, 391, 391, 391, + 392, 392, 392, 392, 393, 393, 393, 394, 394, 394, + 395, 395, 395, 395, 396, 396, 396, 397, 397, 397, + 398, 398, 398, 398, 399, 399, 399, 400, 400, 400, + 401, 401, 401, 402, 402, 402, 402, 403, 403, 403, + 404, 404, 404, 405, 405, 405, 405, 406, 406, 406, + 407, 407, 407, 408, 408, 408, 408, 409, 409, 409, + 410, 410, 410, 411, 411, 411, 411, 412, 412, 412, + 413, 413, 413, 414, 414, 414, 414, 415, 415, 415, + 416, 416, 416, 417, 417, 417, 417, 418, 418, 418, + 419, 419, 419, 420, 420, 420, 420, 421, 421, 421, + 422, 422, 422, 423, 423, 423, 423, 424, 424, 424, + 425, 425, 425, 426, 426, 426, 426, 427, 427, 427, + 428, 428, 428, 429, 429, 429, 429, 430, 430, 430, + 431, 431, 431, 432, 432, 432, 433, 433, 433, 433, + 434, 434, 434, 435, 435, 435, 436, 436, 436, 436, + 437, 437, 437, 438, 438, 438, 439, 439, 439, 439, + 440, 440, 440, 441, 441, 441, 442, 442, 442, 442, + 443, 443, 443, 444, 444, 444, 445, 445, 445, 445, + 446, 446, 446, 447, 447, 447, 448, 448, 448, 448, + 449, 449, 449, 450, 450, 450, 451, 451, 451, 451, + 452, 452, 452, 453, 453, 453, 454, 454, 454, 454, + 455, 455, 455, 456, 456, 456, 457, 457, 457, 457, + 458, 458, 458, 459, 459, 459, 460, 460, 460, 461, + 461, 461, 461, 462, 462, 462, 463, 463, 463, 464, + 464, 464, 464, 465, 465, 465, 466, 466, 466, 467, + 467, 467, 467, 468, 468, 468, 469, 469, 469, 470, + 470, 470, 470, 471, 471, 471, 472, 472, 472, 473, + 473, 473, 473, 474, 474, 474, 475, 475, 475, 476, + 476, 476, 476, 477, 477, 477, 478, 478, 478, 479, + 479, 479, 479, 480, 480, 480, 481, 481, 481, 482, + 482, 482, 482, 483, 483, 483, 484, 484, 484, 485, + 485, 485, 485, 486, 486, 486, 487, 487, 487, 488, + 488, 488, 488, 489, 489, 489, 490, 490, 490, 491, + 491, 491, 492, 492, 492, 492, 493, 493, 493, 494, + 494, 494, 495, 495, 495, 495, 496, 496, 496, 497, + 497, 497, 498, 498, 498, 498, 499, 499, 499, 500, + 500, 500, 501, 501, 501, 501, 502, 502, 502, 503, + 503, 503, 504, 504, 504, 504, 505, 505, 505, 506, + 506, 506, 507, 507, 507, 507, 508, 508, 508, 509, + 509, 509, 510, 510, 510, 510, 511, 511, 511, 512, + 512, 512, 513, 513, 513, 513, 514, 514, 514, 515, + 515, 515, 516, 516, 516, 516, 517, 517, 517, 518, + 518, 518, 519, 519, 519, 520, 520, 520, 520, 521, + 521, 521, 522, 522, 522, 523, 523, 523, 523, 524, + 524, 524, 525, 525, 525, 526, 526, 526, 526, 527, + 527, 527, 528, 528, 528, 529, 529, 529, 529, 530, + 530, 530, 531, 531, 531, 532, 532, 532, 532, 533, + 533, 533, 534, 534, 534, 535, 535, 535, 535, 536, + 536, 536, 537, 537, 537, 538, 538, 538, 538, 539, + 539, 539, 540, 540, 540, 541, 541, 541, 541, 542, + 542, 542, 543, 543, 543, 544, 544, 544, 544, 545, + 545, 545, 546, 546, 546, 547, 547, 547, 548, 548, + 548, 548, 549, 549, 549, 550, 550, 550, 551, 551, + 551, 551, 552, 552, 552, 553, 553, 553, 554, 554, + 554, 554, 555, 555, 555, 556, 556, 556, 557, 557, + 557, 557, 558, 558, 558, 559, 559, 559, 560, 560, + 560, 560, 561, 561, 561, 562, 562, 562, 563, 563, + 563, 563, 564, 564, 564, 565, 565, 565, 566, 566, + 566, 566, 567, 567, 567, 568, 568, 568, 569, 569, + 569, 569, 570, 570, 570, 571, 571, 571, 572, 572, + 572, 572, 573, 573, 573, 574, 574, 574, 575, 575, + 575, 575, 576, 576, 576, 577, 577, 577, 578, 578, + 578, 579, 579, 579, 579, 580, 580, 580, 581, 581, + 581, 582, 582, 582, 582, 583, 583, 583, 584, 584, + 584, 585, 585, 585, 585, 586, 586, 586, 587, 587, + 587, 588, 588, 588, 588, 589, 589, 589, 590, 590, + 590, 591, 591, 591, 591, 592, 592, 592, 593, 593, + 593, 594, 594, 594, 594, 595, 595, 595, 596, 596, + 596, 597, 597, 597, 597, 598, 598, 598, 599, 599, + 599, 600, 600, 600, 600, 601, 601, 601, 602, 602, + 602, 603, 603, 603, 603, 604, 604, 604, 605, 605, + 605, 606, 606, 606, 607, 607, 607, 607, 608, 608, + 608, 609, 609, 609, 610, 610, 610, 610, 611, 611, + 611, 612, 612, 612, 613, 613, 613, 613, 614, 614, + 614, 615, 615, 615, 616, 616, 616, 616, 617, 617, + 617, 618, 618, 618, 619, 619, 619, 619, 620, 620, + 620, 621, 621, 621, 622, 622, 622, 622, 623, 623, + 623, 624, 624, 624, 625, 625, 625, 625, 626, 626, + 626, 627, 627, 627, 628, 628, 628, 628, 629, 629, + 629, 630, 630, 630, 631, 631, 631, 631, 632, 632, + 632, 633, 633, 633, 634, 634, 634, 634, 635, 635, + 635, 636, 636, 636, 637, 637, 637, 638, 638, 638, + 638, 639, 639, 639, 640, 640, 640, 641, 641, 641, + 641, 642, 642, 642, 643, 643, 643, 644, 644, 644, + 644, 645, 645, 645, 646, 646, 646, 647, 647, 647, + 647, 648, 648, 648, 649, 649, 649, 650, 650 }; + static ULLong pfive[27] = { + 5ll, + 25ll, + 125ll, + 625ll, + 3125ll, + 15625ll, + 78125ll, + 390625ll, + 1953125ll, + 9765625ll, + 48828125ll, + 244140625ll, + 1220703125ll, + 6103515625ll, + 30517578125ll, + 152587890625ll, + 762939453125ll, + 3814697265625ll, + 19073486328125ll, + 95367431640625ll, + 476837158203125ll, + 2384185791015625ll, + 11920928955078125ll, + 59604644775390625ll, + 298023223876953125ll, + 1490116119384765625ll, + 7450580596923828125ll + }; + + static int pfivebits[25] = {3, 5, 7, 10, 12, 14, 17, 19, 21, 24, 26, 28, 31, + 33, 35, 38, 40, 42, 45, 47, 49, 52, 54, 56, 59}; +#endif /*}*/ +#endif /*}} NO_LONG_LONG */ + +typedef union { double d; ULong L[2]; +#ifdef USE_BF96 + ULLong LL; +#endif + } U; + +#ifdef IEEE_8087 +#define word0(x) (x)->L[1] +#define word1(x) (x)->L[0] +#else +#define word0(x) (x)->L[0] +#define word1(x) (x)->L[1] +#endif +#define dval(x) (x)->d +#define LLval(x) (x)->LL + +#ifndef STRTOD_DIGLIM +#define STRTOD_DIGLIM 40 +#endif + +#ifdef DIGLIM_DEBUG +extern int strtod_diglim; +#else +#define strtod_diglim STRTOD_DIGLIM +#endif + +/* The following definition of Storeinc is appropriate for MIPS processors. + * An alternative that might be better on some machines is + * #define Storeinc(a,b,c) (*a++ = b << 16 | c & 0xffff) + */ +#if defined(IEEE_8087) + defined(VAX) +#define Storeinc(a,b,c) (((unsigned short *)a)[1] = (unsigned short)b, \ +((unsigned short *)a)[0] = (unsigned short)c, a++) +#else +#define Storeinc(a,b,c) (((unsigned short *)a)[0] = (unsigned short)b, \ +((unsigned short *)a)[1] = (unsigned short)c, a++) +#endif + +/* #define P DBL_MANT_DIG */ +/* Ten_pmax = floor(P*log(2)/log(5)) */ +/* Bletch = (highest power of 2 < DBL_MAX_10_EXP) / 16 */ +/* Quick_max = floor((P-1)*log(FLT_RADIX)/log(10) - 1) */ +/* Int_max = floor(P*log(FLT_RADIX)/log(10) - 1) */ + +#ifdef IEEE_Arith +#define Exp_shift 20 +#define Exp_shift1 20 +#define Exp_msk1 0x100000 +#define Exp_msk11 0x100000 +#define Exp_mask 0x7ff00000 +#define P 53 +#define Nbits 53 +#define Bias 1023 +#define Emax 1023 +#define Emin (-1022) +#define Exp_1 0x3ff00000 +#define Exp_11 0x3ff00000 +#define Ebits 11 +#define Frac_mask 0xfffff +#define Frac_mask1 0xfffff +#define Ten_pmax 22 +#define Bletch 0x10 +#define Bndry_mask 0xfffff +#define Bndry_mask1 0xfffff +#define LSB 1 +#define Sign_bit 0x80000000 +#define Log2P 1 +#define Tiny0 0 +#define Tiny1 1 +#define Quick_max 14 +#define Int_max 14 +#ifndef NO_IEEE_Scale +#define Avoid_Underflow +#ifdef Flush_Denorm /* debugging option */ +#undef Sudden_Underflow +#endif +#endif + +#ifndef Flt_Rounds +#ifdef FLT_ROUNDS +#define Flt_Rounds FLT_ROUNDS +#else +#define Flt_Rounds 1 +#endif +#endif /*Flt_Rounds*/ + +#ifdef Honor_FLT_ROUNDS +#undef Check_FLT_ROUNDS +#define Check_FLT_ROUNDS +#else +#define Rounding Flt_Rounds +#endif + +#else /* ifndef IEEE_Arith */ +#undef Check_FLT_ROUNDS +#undef Honor_FLT_ROUNDS +#undef SET_INEXACT +#undef Sudden_Underflow +#define Sudden_Underflow +#ifdef IBM +#undef Flt_Rounds +#define Flt_Rounds 0 +#define Exp_shift 24 +#define Exp_shift1 24 +#define Exp_msk1 0x1000000 +#define Exp_msk11 0x1000000 +#define Exp_mask 0x7f000000 +#define P 14 +#define Nbits 56 +#define Bias 65 +#define Emax 248 +#define Emin (-260) +#define Exp_1 0x41000000 +#define Exp_11 0x41000000 +#define Ebits 8 /* exponent has 7 bits, but 8 is the right value in b2d */ +#define Frac_mask 0xffffff +#define Frac_mask1 0xffffff +#define Bletch 4 +#define Ten_pmax 22 +#define Bndry_mask 0xefffff +#define Bndry_mask1 0xffffff +#define LSB 1 +#define Sign_bit 0x80000000 +#define Log2P 4 +#define Tiny0 0x100000 +#define Tiny1 0 +#define Quick_max 14 +#define Int_max 15 +#else /* VAX */ +#undef Flt_Rounds +#define Flt_Rounds 1 +#define Exp_shift 23 +#define Exp_shift1 7 +#define Exp_msk1 0x80 +#define Exp_msk11 0x800000 +#define Exp_mask 0x7f80 +#define P 56 +#define Nbits 56 +#define Bias 129 +#define Emax 126 +#define Emin (-129) +#define Exp_1 0x40800000 +#define Exp_11 0x4080 +#define Ebits 8 +#define Frac_mask 0x7fffff +#define Frac_mask1 0xffff007f +#define Ten_pmax 24 +#define Bletch 2 +#define Bndry_mask 0xffff007f +#define Bndry_mask1 0xffff007f +#define LSB 0x10000 +#define Sign_bit 0x8000 +#define Log2P 1 +#define Tiny0 0x80 +#define Tiny1 0 +#define Quick_max 15 +#define Int_max 15 +#endif /* IBM, VAX */ +#endif /* IEEE_Arith */ + +#ifndef IEEE_Arith +#define ROUND_BIASED +#else +#ifdef ROUND_BIASED_without_Round_Up +#undef ROUND_BIASED +#define ROUND_BIASED +#endif +#endif + +#ifdef RND_PRODQUOT +#define rounded_product(a,b) a = rnd_prod(a, b) +#define rounded_quotient(a,b) a = rnd_quot(a, b) +extern double rnd_prod(double, double), rnd_quot(double, double); +#else +#define rounded_product(a,b) a *= b +#define rounded_quotient(a,b) a /= b +#endif + +#define Big0 (Frac_mask1 | Exp_msk1*(DBL_MAX_EXP+Bias-1)) +#define Big1 0xffffffff + +#ifndef Pack_32 +#define Pack_32 +#endif + +typedef struct BCinfo BCinfo; + struct +BCinfo { int dp0, dp1, dplen, dsign, e0, inexact, nd, nd0, rounding, scale, uflchk; }; + +#define FFFFFFFF 0xffffffffUL + +#ifdef MULTIPLE_THREADS +#define MTa , PTI +#define MTb , &TI +#define MTd , ThInfo **PTI +static unsigned int maxthreads = 0; +#else +#define MTa /*nothing*/ +#define MTb /*nothing*/ +#define MTd /*nothing*/ +#endif + +#define Kmax 7 + +#ifdef __cplusplus +extern "C" double strtod(const char *s00, char **se); +extern "C" char *dtoa(double d, int mode, int ndigits, + int *decpt, int *sign, char **rve); +#endif + + struct +Bigint { + struct Bigint *next; + int k, maxwds, sign, wds; + ULong x[1]; + }; + + typedef struct Bigint Bigint; + typedef struct +ThInfo { + Bigint *Freelist[Kmax+1]; + Bigint *P5s; + } ThInfo; + + static ThInfo TI0; + +#ifdef MULTIPLE_THREADS + static ThInfo *TI1; + static int TI0_used; + + void +set_max_dtoa_threads(unsigned int n) +{ + size_t L; + + if (n > maxthreads) { + L = n*sizeof(ThInfo); + if (TI1) { + TI1 = (ThInfo*)REALLOC(TI1, L); + memset(TI1 + maxthreads, 0, (n-maxthreads)*sizeof(ThInfo)); + } + else { + TI1 = (ThInfo*)MALLOC(L); + if (TI0_used) { + memcpy(TI1, &TI0, sizeof(ThInfo)); + if (n > 1) + memset(TI1 + 1, 0, L - sizeof(ThInfo)); + memset(&TI0, 0, sizeof(ThInfo)); + } + else + memset(TI1, 0, L); + } + maxthreads = n; + } + } + + static ThInfo* +get_TI(void) +{ + unsigned int thno = dtoa_get_threadno(); + if (thno < maxthreads) + return TI1 + thno; + if (thno == 0) + TI0_used = 1; + return &TI0; + } +#define freelist TI->Freelist +#define p5s TI->P5s +#else +#define freelist TI0.Freelist +#define p5s TI0.P5s +#endif + + static Bigint * +Balloc(int k MTd) +{ + int x; + Bigint *rv; +#ifndef Omit_Private_Memory + unsigned int len; +#endif +#ifdef MULTIPLE_THREADS + ThInfo *TI; + + if (!(TI = *PTI)) + *PTI = TI = get_TI(); + if (TI == &TI0) + ACQUIRE_DTOA_LOCK(0); +#endif + /* The k > Kmax case does not need ACQUIRE_DTOA_LOCK(0), */ + /* but this case seems very unlikely. */ + if (k <= Kmax && (rv = freelist[k])) + freelist[k] = rv->next; + else { + x = 1 << k; +#ifdef Omit_Private_Memory + rv = (Bigint *)MALLOC(sizeof(Bigint) + (x-1)*sizeof(ULong)); +#else + len = (sizeof(Bigint) + (x-1)*sizeof(ULong) + sizeof(double) - 1) + /sizeof(double); + if (k <= Kmax && pmem_next - private_mem + len <= PRIVATE_mem +#ifdef MULTIPLE_THREADS + && TI == TI1 +#endif + ) { + rv = (Bigint*)pmem_next; + pmem_next += len; + } + else + rv = (Bigint*)MALLOC(len*sizeof(double)); +#endif + rv->k = k; + rv->maxwds = x; + } +#ifdef MULTIPLE_THREADS + if (TI == &TI0) + FREE_DTOA_LOCK(0); +#endif + rv->sign = rv->wds = 0; + return rv; + } + + static void +Bfree(Bigint *v MTd) +{ +#ifdef MULTIPLE_THREADS + ThInfo *TI; +#endif + if (v) { + if (v->k > Kmax) + FREE((void*)v); + else { +#ifdef MULTIPLE_THREADS + if (!(TI = *PTI)) + *PTI = TI = get_TI(); + if (TI == &TI0) + ACQUIRE_DTOA_LOCK(0); +#endif + v->next = freelist[v->k]; + freelist[v->k] = v; +#ifdef MULTIPLE_THREADS + if (TI == &TI0) + FREE_DTOA_LOCK(0); +#endif + } + } + } + +#define Bcopy(x,y) memcpy((char *)&x->sign, (char *)&y->sign, \ +y->wds*sizeof(Long) + 2*sizeof(int)) + + static Bigint * +multadd(Bigint *b, int m, int a MTd) /* multiply by m and add a */ +{ + int i, wds; +#ifdef ULLong + ULong *x; + ULLong carry, y; +#else + ULong carry, *x, y; +#ifdef Pack_32 + ULong xi, z; +#endif +#endif + Bigint *b1; + + wds = b->wds; + x = b->x; + i = 0; + carry = a; + do { +#ifdef ULLong + y = *x * (ULLong)m + carry; + carry = y >> 32; + *x++ = y & FFFFFFFF; +#else +#ifdef Pack_32 + xi = *x; + y = (xi & 0xffff) * m + carry; + z = (xi >> 16) * m + (y >> 16); + carry = z >> 16; + *x++ = (z << 16) + (y & 0xffff); +#else + y = *x * m + carry; + carry = y >> 16; + *x++ = y & 0xffff; +#endif +#endif + } + while(++i < wds); + if (carry) { + if (wds >= b->maxwds) { + b1 = Balloc(b->k+1 MTa); + Bcopy(b1, b); + Bfree(b MTa); + b = b1; + } + b->x[wds++] = carry; + b->wds = wds; + } + return b; + } + + static Bigint * +s2b(const char *s, int nd0, int nd, ULong y9, int dplen MTd) +{ + Bigint *b; + int i, k; + Long x, y; + + x = (nd + 8) / 9; + for(k = 0, y = 1; x > y; y <<= 1, k++) ; +#ifdef Pack_32 + b = Balloc(k MTa); + b->x[0] = y9; + b->wds = 1; +#else + b = Balloc(k+1 MTa); + b->x[0] = y9 & 0xffff; + b->wds = (b->x[1] = y9 >> 16) ? 2 : 1; +#endif + + i = 9; + if (9 < nd0) { + s += 9; + do b = multadd(b, 10, *s++ - '0' MTa); + while(++i < nd0); + s += dplen; + } + else + s += dplen + 9; + for(; i < nd; i++) + b = multadd(b, 10, *s++ - '0' MTa); + return b; + } + + static int +hi0bits(ULong x) +{ + int k = 0; + + if (!(x & 0xffff0000)) { + k = 16; + x <<= 16; + } + if (!(x & 0xff000000)) { + k += 8; + x <<= 8; + } + if (!(x & 0xf0000000)) { + k += 4; + x <<= 4; + } + if (!(x & 0xc0000000)) { + k += 2; + x <<= 2; + } + if (!(x & 0x80000000)) { + k++; + if (!(x & 0x40000000)) + return 32; + } + return k; + } + + static int +lo0bits(ULong *y) +{ + int k; + ULong x = *y; + + if (x & 7) { + if (x & 1) + return 0; + if (x & 2) { + *y = x >> 1; + return 1; + } + *y = x >> 2; + return 2; + } + k = 0; + if (!(x & 0xffff)) { + k = 16; + x >>= 16; + } + if (!(x & 0xff)) { + k += 8; + x >>= 8; + } + if (!(x & 0xf)) { + k += 4; + x >>= 4; + } + if (!(x & 0x3)) { + k += 2; + x >>= 2; + } + if (!(x & 1)) { + k++; + x >>= 1; + if (!x) + return 32; + } + *y = x; + return k; + } + + static Bigint * +i2b(int i MTd) +{ + Bigint *b; + + b = Balloc(1 MTa); + b->x[0] = i; + b->wds = 1; + return b; + } + + static Bigint * +mult(Bigint *a, Bigint *b MTd) +{ + Bigint *c; + int k, wa, wb, wc; + ULong *x, *xa, *xae, *xb, *xbe, *xc, *xc0; + ULong y; +#ifdef ULLong + ULLong carry, z; +#else + ULong carry, z; +#ifdef Pack_32 + ULong z2; +#endif +#endif + + if (a->wds < b->wds) { + c = a; + a = b; + b = c; + } + k = a->k; + wa = a->wds; + wb = b->wds; + wc = wa + wb; + if (wc > a->maxwds) + k++; + c = Balloc(k MTa); + for(x = c->x, xa = x + wc; x < xa; x++) + *x = 0; + xa = a->x; + xae = xa + wa; + xb = b->x; + xbe = xb + wb; + xc0 = c->x; +#ifdef ULLong + for(; xb < xbe; xc0++) { + if ((y = *xb++)) { + x = xa; + xc = xc0; + carry = 0; + do { + z = *x++ * (ULLong)y + *xc + carry; + carry = z >> 32; + *xc++ = z & FFFFFFFF; + } + while(x < xae); + *xc = carry; + } + } +#else +#ifdef Pack_32 + for(; xb < xbe; xb++, xc0++) { + if (y = *xb & 0xffff) { + x = xa; + xc = xc0; + carry = 0; + do { + z = (*x & 0xffff) * y + (*xc & 0xffff) + carry; + carry = z >> 16; + z2 = (*x++ >> 16) * y + (*xc >> 16) + carry; + carry = z2 >> 16; + Storeinc(xc, z2, z); + } + while(x < xae); + *xc = carry; + } + if (y = *xb >> 16) { + x = xa; + xc = xc0; + carry = 0; + z2 = *xc; + do { + z = (*x & 0xffff) * y + (*xc >> 16) + carry; + carry = z >> 16; + Storeinc(xc, z, z2); + z2 = (*x++ >> 16) * y + (*xc & 0xffff) + carry; + carry = z2 >> 16; + } + while(x < xae); + *xc = z2; + } + } +#else + for(; xb < xbe; xc0++) { + if (y = *xb++) { + x = xa; + xc = xc0; + carry = 0; + do { + z = *x++ * y + *xc + carry; + carry = z >> 16; + *xc++ = z & 0xffff; + } + while(x < xae); + *xc = carry; + } + } +#endif +#endif + for(xc0 = c->x, xc = xc0 + wc; wc > 0 && !*--xc; --wc) ; + c->wds = wc; + return c; + } + + static Bigint * +pow5mult(Bigint *b, int k MTd) +{ + Bigint *b1, *p5, *p51; +#ifdef MULTIPLE_THREADS + ThInfo *TI; +#endif + int i; + static int p05[3] = { 5, 25, 125 }; + + if ((i = k & 3)) + b = multadd(b, p05[i-1], 0 MTa); + + if (!(k >>= 2)) + return b; +#ifdef MULTIPLE_THREADS + if (!(TI = *PTI)) + *PTI = TI = get_TI(); +#endif + if (!(p5 = p5s)) { + /* first time */ +#ifdef MULTIPLE_THREADS + if (!(TI = *PTI)) + *PTI = TI = get_TI(); + if (TI == &TI0) + ACQUIRE_DTOA_LOCK(1); + if (!(p5 = p5s)) { + p5 = p5s = i2b(625 MTa); + p5->next = 0; + } + if (TI == &TI0) + FREE_DTOA_LOCK(1); +#else + p5 = p5s = i2b(625 MTa); + p5->next = 0; +#endif + } + for(;;) { + if (k & 1) { + b1 = mult(b, p5 MTa); + Bfree(b MTa); + b = b1; + } + if (!(k >>= 1)) + break; + if (!(p51 = p5->next)) { +#ifdef MULTIPLE_THREADS + if (!TI && !(TI = *PTI)) + *PTI = TI = get_TI(); + if (TI == &TI0) + ACQUIRE_DTOA_LOCK(1); + if (!(p51 = p5->next)) { + p51 = p5->next = mult(p5,p5 MTa); + p51->next = 0; + } + if (TI == &TI0) + FREE_DTOA_LOCK(1); +#else + p51 = p5->next = mult(p5,p5); + p51->next = 0; +#endif + } + p5 = p51; + } + return b; + } + + static Bigint * +lshift(Bigint *b, int k MTd) +{ + int i, k1, n, n1; + Bigint *b1; + ULong *x, *x1, *xe, z; + +#ifdef Pack_32 + n = k >> 5; +#else + n = k >> 4; +#endif + k1 = b->k; + n1 = n + b->wds + 1; + for(i = b->maxwds; n1 > i; i <<= 1) + k1++; + b1 = Balloc(k1 MTa); + x1 = b1->x; + for(i = 0; i < n; i++) + *x1++ = 0; + x = b->x; + xe = x + b->wds; +#ifdef Pack_32 + if (k &= 0x1f) { + k1 = 32 - k; + z = 0; + do { + *x1++ = *x << k | z; + z = *x++ >> k1; + } + while(x < xe); + if ((*x1 = z)) + ++n1; + } +#else + if (k &= 0xf) { + k1 = 16 - k; + z = 0; + do { + *x1++ = *x << k & 0xffff | z; + z = *x++ >> k1; + } + while(x < xe); + if (*x1 = z) + ++n1; + } +#endif + else do + *x1++ = *x++; + while(x < xe); + b1->wds = n1 - 1; + Bfree(b MTa); + return b1; + } + + static int +cmp(Bigint *a, Bigint *b) +{ + ULong *xa, *xa0, *xb, *xb0; + int i, j; + + i = a->wds; + j = b->wds; +#ifdef DEBUG + if (i > 1 && !a->x[i-1]) + Bug("cmp called with a->x[a->wds-1] == 0"); + if (j > 1 && !b->x[j-1]) + Bug("cmp called with b->x[b->wds-1] == 0"); +#endif + if (i -= j) + return i; + xa0 = a->x; + xa = xa0 + j; + xb0 = b->x; + xb = xb0 + j; + for(;;) { + if (*--xa != *--xb) + return *xa < *xb ? -1 : 1; + if (xa <= xa0) + break; + } + return 0; + } + + static Bigint * +diff(Bigint *a, Bigint *b MTd) +{ + Bigint *c; + int i, wa, wb; + ULong *xa, *xae, *xb, *xbe, *xc; +#ifdef ULLong + ULLong borrow, y; +#else + ULong borrow, y; +#ifdef Pack_32 + ULong z; +#endif +#endif + + i = cmp(a,b); + if (!i) { + c = Balloc(0 MTa); + c->wds = 1; + c->x[0] = 0; + return c; + } + if (i < 0) { + c = a; + a = b; + b = c; + i = 1; + } + else + i = 0; + c = Balloc(a->k MTa); + c->sign = i; + wa = a->wds; + xa = a->x; + xae = xa + wa; + wb = b->wds; + xb = b->x; + xbe = xb + wb; + xc = c->x; + borrow = 0; +#ifdef ULLong + do { + y = (ULLong)*xa++ - *xb++ - borrow; + borrow = y >> 32 & (ULong)1; + *xc++ = y & FFFFFFFF; + } + while(xb < xbe); + while(xa < xae) { + y = *xa++ - borrow; + borrow = y >> 32 & (ULong)1; + *xc++ = y & FFFFFFFF; + } +#else +#ifdef Pack_32 + do { + y = (*xa & 0xffff) - (*xb & 0xffff) - borrow; + borrow = (y & 0x10000) >> 16; + z = (*xa++ >> 16) - (*xb++ >> 16) - borrow; + borrow = (z & 0x10000) >> 16; + Storeinc(xc, z, y); + } + while(xb < xbe); + while(xa < xae) { + y = (*xa & 0xffff) - borrow; + borrow = (y & 0x10000) >> 16; + z = (*xa++ >> 16) - borrow; + borrow = (z & 0x10000) >> 16; + Storeinc(xc, z, y); + } +#else + do { + y = *xa++ - *xb++ - borrow; + borrow = (y & 0x10000) >> 16; + *xc++ = y & 0xffff; + } + while(xb < xbe); + while(xa < xae) { + y = *xa++ - borrow; + borrow = (y & 0x10000) >> 16; + *xc++ = y & 0xffff; + } +#endif +#endif + while(!*--xc) + wa--; + c->wds = wa; + return c; + } + + static double +ulp(U *x) +{ + Long L; + U u; + + L = (word0(x) & Exp_mask) - (P-1)*Exp_msk1; +#ifndef Avoid_Underflow +#ifndef Sudden_Underflow + if (L > 0) { +#endif +#endif +#ifdef IBM + L |= Exp_msk1 >> 4; +#endif + word0(&u) = L; + word1(&u) = 0; +#ifndef Avoid_Underflow +#ifndef Sudden_Underflow + } + else { + L = -L >> Exp_shift; + if (L < Exp_shift) { + word0(&u) = 0x80000 >> L; + word1(&u) = 0; + } + else { + word0(&u) = 0; + L -= Exp_shift; + word1(&u) = L >= 31 ? 1 : 1 << 31 - L; + } + } +#endif +#endif + return dval(&u); + } + + static double +b2d(Bigint *a, int *e) +{ + ULong *xa, *xa0, w, y, z; + int k; + U d; +#ifdef VAX + ULong d0, d1; +#else +#define d0 word0(&d) +#define d1 word1(&d) +#endif + + xa0 = a->x; + xa = xa0 + a->wds; + y = *--xa; +#ifdef DEBUG + if (!y) Bug("zero y in b2d"); +#endif + k = hi0bits(y); + *e = 32 - k; +#ifdef Pack_32 + if (k < Ebits) { + d0 = Exp_1 | y >> (Ebits - k); + w = xa > xa0 ? *--xa : 0; + d1 = y << ((32-Ebits) + k) | w >> (Ebits - k); + goto ret_d; + } + z = xa > xa0 ? *--xa : 0; + if (k -= Ebits) { + d0 = Exp_1 | y << k | z >> (32 - k); + y = xa > xa0 ? *--xa : 0; + d1 = z << k | y >> (32 - k); + } + else { + d0 = Exp_1 | y; + d1 = z; + } +#else + if (k < Ebits + 16) { + z = xa > xa0 ? *--xa : 0; + d0 = Exp_1 | y << k - Ebits | z >> Ebits + 16 - k; + w = xa > xa0 ? *--xa : 0; + y = xa > xa0 ? *--xa : 0; + d1 = z << k + 16 - Ebits | w << k - Ebits | y >> 16 + Ebits - k; + goto ret_d; + } + z = xa > xa0 ? *--xa : 0; + w = xa > xa0 ? *--xa : 0; + k -= Ebits + 16; + d0 = Exp_1 | y << k + 16 | z << k | w >> 16 - k; + y = xa > xa0 ? *--xa : 0; + d1 = w << k + 16 | y << k; +#endif + ret_d: +#ifdef VAX + word0(&d) = d0 >> 16 | d0 << 16; + word1(&d) = d1 >> 16 | d1 << 16; +#else +#undef d0 +#undef d1 +#endif + return dval(&d); + } + + static Bigint * +d2b(U *d, int *e, int *bits MTd) +{ + Bigint *b; + int de, k; + ULong *x, y, z; +#ifndef Sudden_Underflow + int i; +#endif +#ifdef VAX + ULong d0, d1; + d0 = word0(d) >> 16 | word0(d) << 16; + d1 = word1(d) >> 16 | word1(d) << 16; +#else +#define d0 word0(d) +#define d1 word1(d) +#endif + +#ifdef Pack_32 + b = Balloc(1 MTa); +#else + b = Balloc(2 MTa); +#endif + x = b->x; + + z = d0 & Frac_mask; + d0 &= 0x7fffffff; /* clear sign bit, which we ignore */ +#ifdef Sudden_Underflow + de = (int)(d0 >> Exp_shift); +#ifndef IBM + z |= Exp_msk11; +#endif +#else + if ((de = (int)(d0 >> Exp_shift))) + z |= Exp_msk1; +#endif +#ifdef Pack_32 + if ((y = d1)) { + if ((k = lo0bits(&y))) { + x[0] = y | z << (32 - k); + z >>= k; + } + else + x[0] = y; +#ifndef Sudden_Underflow + i = +#endif + b->wds = (x[1] = z) ? 2 : 1; + } + else { + k = lo0bits(&z); + x[0] = z; +#ifndef Sudden_Underflow + i = +#endif + b->wds = 1; + k += 32; + } +#else + if (y = d1) { + if (k = lo0bits(&y)) + if (k >= 16) { + x[0] = y | z << 32 - k & 0xffff; + x[1] = z >> k - 16 & 0xffff; + x[2] = z >> k; + i = 2; + } + else { + x[0] = y & 0xffff; + x[1] = y >> 16 | z << 16 - k & 0xffff; + x[2] = z >> k & 0xffff; + x[3] = z >> k+16; + i = 3; + } + else { + x[0] = y & 0xffff; + x[1] = y >> 16; + x[2] = z & 0xffff; + x[3] = z >> 16; + i = 3; + } + } + else { +#ifdef DEBUG + if (!z) + Bug("Zero passed to d2b"); +#endif + k = lo0bits(&z); + if (k >= 16) { + x[0] = z; + i = 0; + } + else { + x[0] = z & 0xffff; + x[1] = z >> 16; + i = 1; + } + k += 32; + } + while(!x[i]) + --i; + b->wds = i + 1; +#endif +#ifndef Sudden_Underflow + if (de) { +#endif +#ifdef IBM + *e = (de - Bias - (P-1) << 2) + k; + *bits = 4*P + 8 - k - hi0bits(word0(d) & Frac_mask); +#else + *e = de - Bias - (P-1) + k; + *bits = P - k; +#endif +#ifndef Sudden_Underflow + } + else { + *e = de - Bias - (P-1) + 1 + k; +#ifdef Pack_32 + *bits = 32*i - hi0bits(x[i-1]); +#else + *bits = (i+2)*16 - hi0bits(x[i]); +#endif + } +#endif + return b; + } +#undef d0 +#undef d1 + + static double +ratio(Bigint *a, Bigint *b) +{ + U da, db; + int k, ka, kb; + + dval(&da) = b2d(a, &ka); + dval(&db) = b2d(b, &kb); +#ifdef Pack_32 + k = ka - kb + 32*(a->wds - b->wds); +#else + k = ka - kb + 16*(a->wds - b->wds); +#endif +#ifdef IBM + if (k > 0) { + word0(&da) += (k >> 2)*Exp_msk1; + if (k &= 3) + dval(&da) *= 1 << k; + } + else { + k = -k; + word0(&db) += (k >> 2)*Exp_msk1; + if (k &= 3) + dval(&db) *= 1 << k; + } +#else + if (k > 0) + word0(&da) += k*Exp_msk1; + else { + k = -k; + word0(&db) += k*Exp_msk1; + } +#endif + return dval(&da) / dval(&db); + } + + static const double +tens[] = { + 1e0, 1e1, 1e2, 1e3, 1e4, 1e5, 1e6, 1e7, 1e8, 1e9, + 1e10, 1e11, 1e12, 1e13, 1e14, 1e15, 1e16, 1e17, 1e18, 1e19, + 1e20, 1e21, 1e22 +#ifdef VAX + , 1e23, 1e24 +#endif + }; + + static const double +#ifdef IEEE_Arith +bigtens[] = { 1e16, 1e32, 1e64, 1e128, 1e256 }; +static const double tinytens[] = { 1e-16, 1e-32, 1e-64, 1e-128, +#ifdef Avoid_Underflow + 9007199254740992.*9007199254740992.e-256 + /* = 2^106 * 1e-256 */ +#else + 1e-256 +#endif + }; +/* The factor of 2^53 in tinytens[4] helps us avoid setting the underflow */ +/* flag unnecessarily. It leads to a song and dance at the end of strtod. */ +#define Scale_Bit 0x10 +#define n_bigtens 5 +#else +#ifdef IBM +bigtens[] = { 1e16, 1e32, 1e64 }; +static const double tinytens[] = { 1e-16, 1e-32, 1e-64 }; +#define n_bigtens 3 +#else +bigtens[] = { 1e16, 1e32 }; +static const double tinytens[] = { 1e-16, 1e-32 }; +#define n_bigtens 2 +#endif +#endif + +#undef Need_Hexdig +#ifdef INFNAN_CHECK +#ifndef No_Hex_NaN +#define Need_Hexdig +#endif +#endif + +#ifndef Need_Hexdig +#ifndef NO_HEX_FP +#define Need_Hexdig +#endif +#endif + +#ifdef Need_Hexdig /*{*/ +#if 0 +static unsigned char hexdig[256]; + + static void +htinit(unsigned char *h, unsigned char *s, int inc) +{ + int i, j; + for(i = 0; (j = s[i]) !=0; i++) + h[j] = i + inc; + } + + static void +hexdig_init(void) /* Use of hexdig_init omitted 20121220 to avoid a */ + /* race condition when multiple threads are used. */ +{ +#define USC (unsigned char *) + htinit(hexdig, USC "0123456789", 0x10); + htinit(hexdig, USC "abcdef", 0x10 + 10); + htinit(hexdig, USC "ABCDEF", 0x10 + 10); + } +#else +static unsigned char hexdig[256] = { + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 16,17,18,19,20,21,22,23,24,25,0,0,0,0,0,0, + 0,26,27,28,29,30,31,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,26,27,28,29,30,31,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 + }; +#endif +#endif /* } Need_Hexdig */ + +#ifdef INFNAN_CHECK + +#ifndef NAN_WORD0 +#define NAN_WORD0 0x7ff80000 +#endif + +#ifndef NAN_WORD1 +#define NAN_WORD1 0 +#endif + + static int +match(const char **sp, const char *t) +{ + int c, d; + const char *s = *sp; + + while((d = *t++)) { + if ((c = *++s) >= 'A' && c <= 'Z') + c += 'a' - 'A'; + if (c != d) + return 0; + } + *sp = s + 1; + return 1; + } + +#ifndef No_Hex_NaN + static void +hexnan(U *rvp, const char **sp) +{ + ULong c, x[2]; + const char *s; + int c1, havedig, udx0, xshift; + + /**** if (!hexdig['0']) hexdig_init(); ****/ + x[0] = x[1] = 0; + havedig = xshift = 0; + udx0 = 1; + s = *sp; + /* allow optional initial 0x or 0X */ + while((c = *(const unsigned char*)(s+1)) && c <= ' ') + ++s; + if (s[1] == '0' && (s[2] == 'x' || s[2] == 'X')) + s += 2; + while((c = *(const unsigned char*)++s)) { + if ((c1 = hexdig[c])) + c = c1 & 0xf; + else if (c <= ' ') { + if (udx0 && havedig) { + udx0 = 0; + xshift = 1; + } + continue; + } +#ifdef GDTOA_NON_PEDANTIC_NANCHECK + else if (/*(*/ c == ')' && havedig) { + *sp = s + 1; + break; + } + else + return; /* invalid form: don't change *sp */ +#else + else { + do { + if (/*(*/ c == ')') { + *sp = s + 1; + break; + } + } while((c = *++s)); + break; + } +#endif + havedig = 1; + if (xshift) { + xshift = 0; + x[0] = x[1]; + x[1] = 0; + } + if (udx0) + x[0] = (x[0] << 4) | (x[1] >> 28); + x[1] = (x[1] << 4) | c; + } + if ((x[0] &= 0xfffff) || x[1]) { + word0(rvp) = Exp_mask | x[0]; + word1(rvp) = x[1]; + } + } +#endif /*No_Hex_NaN*/ +#endif /* INFNAN_CHECK */ + +#ifdef Pack_32 +#define ULbits 32 +#define kshift 5 +#define kmask 31 +#else +#define ULbits 16 +#define kshift 4 +#define kmask 15 +#endif + +#if !defined(NO_HEX_FP) || defined(Honor_FLT_ROUNDS) /*{*/ + static Bigint * +increment(Bigint *b MTd) +{ + ULong *x, *xe; + Bigint *b1; + + x = b->x; + xe = x + b->wds; + do { + if (*x < (ULong)0xffffffffL) { + ++*x; + return b; + } + *x++ = 0; + } while(x < xe); + { + if (b->wds >= b->maxwds) { + b1 = Balloc(b->k+1 MTa); + Bcopy(b1,b); + Bfree(b MTa); + b = b1; + } + b->x[b->wds++] = 1; + } + return b; + } + +#endif /*}*/ + +#ifndef NO_HEX_FP /*{*/ + + static void +rshift(Bigint *b, int k) +{ + ULong *x, *x1, *xe, y; + int n; + + x = x1 = b->x; + n = k >> kshift; + if (n < b->wds) { + xe = x + b->wds; + x += n; + if (k &= kmask) { + n = 32 - k; + y = *x++ >> k; + while(x < xe) { + *x1++ = (y | (*x << n)) & 0xffffffff; + y = *x++ >> k; + } + if ((*x1 = y) !=0) + x1++; + } + else + while(x < xe) + *x1++ = *x++; + } + if ((b->wds = x1 - b->x) == 0) + b->x[0] = 0; + } + + static ULong +any_on(Bigint *b, int k) +{ + int n, nwds; + ULong *x, *x0, x1, x2; + + x = b->x; + nwds = b->wds; + n = k >> kshift; + if (n > nwds) + n = nwds; + else if (n < nwds && (k &= kmask)) { + x1 = x2 = x[n]; + x1 >>= k; + x1 <<= k; + if (x1 != x2) + return 1; + } + x0 = x; + x += n; + while(x > x0) + if (*--x) + return 1; + return 0; + } + +enum { /* rounding values: same as FLT_ROUNDS */ + Round_zero = 0, + Round_near = 1, + Round_up = 2, + Round_down = 3 + }; + +static void +gethex( const char **sp, U *rvp, int rounding, int sign MTd) +{ + Bigint *b; + const unsigned char *decpt, *s0, *s, *s1; + Long e, e1; + ULong L, lostbits, *x; + int big, denorm, esign, havedig, k, n, nbits, up, zret; +#ifdef IBM + int j; +#endif + enum { +#ifdef IEEE_Arith /*{{*/ + emax = 0x7fe - Bias - P + 1, + emin = Emin - P + 1 +#else /*}{*/ + emin = Emin - P, +#ifdef VAX + emax = 0x7ff - Bias - P + 1 +#endif +#ifdef IBM + emax = 0x7f - Bias - P +#endif +#endif /*}}*/ + }; +#ifdef USE_LOCALE + int i; +#ifdef NO_LOCALE_CACHE + const unsigned char *decimalpoint = (unsigned char*) + localeconv()->decimal_point; +#else + const unsigned char *decimalpoint; + static unsigned char *decimalpoint_cache; + if (!(s0 = decimalpoint_cache)) { + s0 = (unsigned char*)localeconv()->decimal_point; + if ((decimalpoint_cache = (unsigned char*) + MALLOC(strlen((const char*)s0) + 1))) { + strcpy((char*)decimalpoint_cache, (const char*)s0); + s0 = decimalpoint_cache; + } + } + decimalpoint = s0; +#endif +#endif + + /**** if (!hexdig['0']) hexdig_init(); ****/ + havedig = 0; + s0 = *(const unsigned char **)sp + 2; + while(s0[havedig] == '0') + havedig++; + s0 += havedig; + s = s0; + decpt = 0; + zret = 0; + e = 0; + if (hexdig[*s]) + havedig++; + else { + zret = 1; +#ifdef USE_LOCALE + for(i = 0; decimalpoint[i]; ++i) { + if (s[i] != decimalpoint[i]) + goto pcheck; + } + decpt = s += i; +#else + if (*s != '.') + goto pcheck; + decpt = ++s; +#endif + if (!hexdig[*s]) + goto pcheck; + while(*s == '0') + s++; + if (hexdig[*s]) + zret = 0; + havedig = 1; + s0 = s; + } + while(hexdig[*s]) + s++; +#ifdef USE_LOCALE + if (*s == *decimalpoint && !decpt) { + for(i = 1; decimalpoint[i]; ++i) { + if (s[i] != decimalpoint[i]) + goto pcheck; + } + decpt = s += i; +#else + if (*s == '.' && !decpt) { + decpt = ++s; +#endif + while(hexdig[*s]) + s++; + }/*}*/ + if (decpt) + e = -(((Long)(s-decpt)) << 2); + pcheck: + s1 = s; + big = esign = 0; + switch(*s) { + case 'p': + case 'P': + switch(*++s) { + case '-': + esign = 1; + /* no break */ + case '+': + s++; + } + if ((n = hexdig[*s]) == 0 || n > 0x19) { + s = s1; + break; + } + e1 = n - 0x10; + while((n = hexdig[*++s]) !=0 && n <= 0x19) { + if (e1 & 0xf8000000) + big = 1; + e1 = 10*e1 + n - 0x10; + } + if (esign) + e1 = -e1; + e += e1; + } + *sp = (char*)s; + if (!havedig) + *sp = (char*)s0 - 1; + if (zret) + goto retz1; + if (big) { + if (esign) { +#ifdef IEEE_Arith + switch(rounding) { + case Round_up: + if (sign) + break; + goto ret_tiny; + case Round_down: + if (!sign) + break; + goto ret_tiny; + } +#endif + goto retz; +#ifdef IEEE_Arith + ret_tinyf: + Bfree(b MTa); + ret_tiny: + Set_errno(ERANGE); + word0(rvp) = 0; + word1(rvp) = 1; + return; +#endif /* IEEE_Arith */ + } + switch(rounding) { + case Round_near: + goto ovfl1; + case Round_up: + if (!sign) + goto ovfl1; + goto ret_big; + case Round_down: + if (sign) + goto ovfl1; + goto ret_big; + } + ret_big: + word0(rvp) = Big0; + word1(rvp) = Big1; + return; + } + n = s1 - s0 - 1; + for(k = 0; n > (1 << (kshift-2)) - 1; n >>= 1) + k++; + b = Balloc(k MTa); + x = b->x; + n = 0; + L = 0; +#ifdef USE_LOCALE + for(i = 0; decimalpoint[i+1]; ++i); +#endif + while(s1 > s0) { +#ifdef USE_LOCALE + if (*--s1 == decimalpoint[i]) { + s1 -= i; + continue; + } +#else + if (*--s1 == '.') + continue; +#endif + if (n == ULbits) { + *x++ = L; + L = 0; + n = 0; + } + L |= (hexdig[*s1] & 0x0f) << n; + n += 4; + } + *x++ = L; + b->wds = n = x - b->x; + n = ULbits*n - hi0bits(L); + nbits = Nbits; + lostbits = 0; + x = b->x; + if (n > nbits) { + n -= nbits; + if (any_on(b,n)) { + lostbits = 1; + k = n - 1; + if (x[k>>kshift] & 1 << (k & kmask)) { + lostbits = 2; + if (k > 0 && any_on(b,k)) + lostbits = 3; + } + } + rshift(b, n); + e += n; + } + else if (n < nbits) { + n = nbits - n; + b = lshift(b, n MTa); + e -= n; + x = b->x; + } + if (e > emax) { + ovfl: + Bfree(b MTa); + ovfl1: + Set_errno(ERANGE); +#ifdef Honor_FLT_ROUNDS + switch (rounding) { + case Round_zero: + goto ret_big; + case Round_down: + if (!sign) + goto ret_big; + break; + case Round_up: + if (sign) + goto ret_big; + } +#endif + word0(rvp) = Exp_mask; + word1(rvp) = 0; + return; + } + denorm = 0; + if (e < emin) { + denorm = 1; + n = emin - e; + if (n >= nbits) { +#ifdef IEEE_Arith /*{*/ + switch (rounding) { + case Round_near: + if (n == nbits && (n < 2 || lostbits || any_on(b,n-1))) + goto ret_tinyf; + break; + case Round_up: + if (!sign) + goto ret_tinyf; + break; + case Round_down: + if (sign) + goto ret_tinyf; + } +#endif /* } IEEE_Arith */ + Bfree(b MTa); + retz: + Set_errno(ERANGE); + retz1: + rvp->d = 0.; + return; + } + k = n - 1; + if (lostbits) + lostbits = 1; + else if (k > 0) + lostbits = any_on(b,k); + if (x[k>>kshift] & 1 << (k & kmask)) + lostbits |= 2; + nbits -= n; + rshift(b,n); + e = emin; + } + if (lostbits) { + up = 0; + switch(rounding) { + case Round_zero: + break; + case Round_near: + if (lostbits & 2 + && (lostbits & 1) | (x[0] & 1)) + up = 1; + break; + case Round_up: + up = 1 - sign; + break; + case Round_down: + up = sign; + } + if (up) { + k = b->wds; + b = increment(b MTa); + x = b->x; + if (denorm) { +#if 0 + if (nbits == Nbits - 1 + && x[nbits >> kshift] & 1 << (nbits & kmask)) + denorm = 0; /* not currently used */ +#endif + } + else if (b->wds > k + || ((n = nbits & kmask) !=0 + && hi0bits(x[k-1]) < 32-n)) { + rshift(b,1); + if (++e > Emax) + goto ovfl; + } + } + } +#ifdef IEEE_Arith + if (denorm) + word0(rvp) = b->wds > 1 ? b->x[1] & ~0x100000 : 0; + else + word0(rvp) = (b->x[1] & ~0x100000) | ((e + 0x3ff + 52) << 20); + word1(rvp) = b->x[0]; +#endif +#ifdef IBM + if ((j = e & 3)) { + k = b->x[0] & ((1 << j) - 1); + rshift(b,j); + if (k) { + switch(rounding) { + case Round_up: + if (!sign) + increment(b); + break; + case Round_down: + if (sign) + increment(b); + break; + case Round_near: + j = 1 << (j-1); + if (k & j && ((k & (j-1)) | lostbits)) + increment(b); + } + } + } + e >>= 2; + word0(rvp) = b->x[1] | ((e + 65 + 13) << 24); + word1(rvp) = b->x[0]; +#endif +#ifdef VAX + /* The next two lines ignore swap of low- and high-order 2 bytes. */ + /* word0(rvp) = (b->x[1] & ~0x800000) | ((e + 129 + 55) << 23); */ + /* word1(rvp) = b->x[0]; */ + word0(rvp) = ((b->x[1] & ~0x800000) >> 16) | ((e + 129 + 55) << 7) | (b->x[1] << 16); + word1(rvp) = (b->x[0] >> 16) | (b->x[0] << 16); +#endif + Bfree(b MTa); + } +#endif /*!NO_HEX_FP}*/ + + static int +dshift(Bigint *b, int p2) +{ + int rv = hi0bits(b->x[b->wds-1]) - 4; + if (p2 > 0) + rv -= p2; + return rv & kmask; + } + + static int +quorem(Bigint *b, Bigint *S) +{ + int n; + ULong *bx, *bxe, q, *sx, *sxe; +#ifdef ULLong + ULLong borrow, carry, y, ys; +#else + ULong borrow, carry, y, ys; +#ifdef Pack_32 + ULong si, z, zs; +#endif +#endif + + n = S->wds; +#ifdef DEBUG + /*debug*/ if (b->wds > n) + /*debug*/ Bug("oversize b in quorem"); +#endif + if (b->wds < n) + return 0; + sx = S->x; + sxe = sx + --n; + bx = b->x; + bxe = bx + n; + q = *bxe / (*sxe + 1); /* ensure q <= true quotient */ +#ifdef DEBUG +#ifdef NO_STRTOD_BIGCOMP + /*debug*/ if (q > 9) +#else + /* An oversized q is possible when quorem is called from bigcomp and */ + /* the input is near, e.g., twice the smallest denormalized number. */ + /*debug*/ if (q > 15) +#endif + /*debug*/ Bug("oversized quotient in quorem"); +#endif + if (q) { + borrow = 0; + carry = 0; + do { +#ifdef ULLong + ys = *sx++ * (ULLong)q + carry; + carry = ys >> 32; + y = *bx - (ys & FFFFFFFF) - borrow; + borrow = y >> 32 & (ULong)1; + *bx++ = y & FFFFFFFF; +#else +#ifdef Pack_32 + si = *sx++; + ys = (si & 0xffff) * q + carry; + zs = (si >> 16) * q + (ys >> 16); + carry = zs >> 16; + y = (*bx & 0xffff) - (ys & 0xffff) - borrow; + borrow = (y & 0x10000) >> 16; + z = (*bx >> 16) - (zs & 0xffff) - borrow; + borrow = (z & 0x10000) >> 16; + Storeinc(bx, z, y); +#else + ys = *sx++ * q + carry; + carry = ys >> 16; + y = *bx - (ys & 0xffff) - borrow; + borrow = (y & 0x10000) >> 16; + *bx++ = y & 0xffff; +#endif +#endif + } + while(sx <= sxe); + if (!*bxe) { + bx = b->x; + while(--bxe > bx && !*bxe) + --n; + b->wds = n; + } + } + if (cmp(b, S) >= 0) { + q++; + borrow = 0; + carry = 0; + bx = b->x; + sx = S->x; + do { +#ifdef ULLong + ys = *sx++ + carry; + carry = ys >> 32; + y = *bx - (ys & FFFFFFFF) - borrow; + borrow = y >> 32 & (ULong)1; + *bx++ = y & FFFFFFFF; +#else +#ifdef Pack_32 + si = *sx++; + ys = (si & 0xffff) + carry; + zs = (si >> 16) + (ys >> 16); + carry = zs >> 16; + y = (*bx & 0xffff) - (ys & 0xffff) - borrow; + borrow = (y & 0x10000) >> 16; + z = (*bx >> 16) - (zs & 0xffff) - borrow; + borrow = (z & 0x10000) >> 16; + Storeinc(bx, z, y); +#else + ys = *sx++ + carry; + carry = ys >> 16; + y = *bx - (ys & 0xffff) - borrow; + borrow = (y & 0x10000) >> 16; + *bx++ = y & 0xffff; +#endif +#endif + } + while(sx <= sxe); + bx = b->x; + bxe = bx + n; + if (!*bxe) { + while(--bxe > bx && !*bxe) + --n; + b->wds = n; + } + } + return q; + } + +#if defined(Avoid_Underflow) || !defined(NO_STRTOD_BIGCOMP) /*{*/ + static double +sulp(U *x, BCinfo *bc) +{ + U u; + double rv; + int i; + + rv = ulp(x); + if (!bc->scale || (i = 2*P + 1 - ((word0(x) & Exp_mask) >> Exp_shift)) <= 0) + return rv; /* Is there an example where i <= 0 ? */ + word0(&u) = Exp_1 + (i << Exp_shift); + word1(&u) = 0; + return rv * u.d; + } +#endif /*}*/ + +#ifndef NO_STRTOD_BIGCOMP + static void +bigcomp(U *rv, const char *s0, BCinfo *bc MTd) +{ + Bigint *b, *d; + int b2, bbits, d2, dd, dig, dsign, i, j, nd, nd0, p2, p5, speccase; + + dsign = bc->dsign; + nd = bc->nd; + nd0 = bc->nd0; + p5 = nd + bc->e0 - 1; + speccase = 0; +#ifndef Sudden_Underflow + if (rv->d == 0.) { /* special case: value near underflow-to-zero */ + /* threshold was rounded to zero */ + b = i2b(1 MTa); + p2 = Emin - P + 1; + bbits = 1; +#ifdef Avoid_Underflow + word0(rv) = (P+2) << Exp_shift; +#else + word1(rv) = 1; +#endif + i = 0; +#ifdef Honor_FLT_ROUNDS + if (bc->rounding == 1) +#endif + { + speccase = 1; + --p2; + dsign = 0; + goto have_i; + } + } + else +#endif + b = d2b(rv, &p2, &bbits MTa); +#ifdef Avoid_Underflow + p2 -= bc->scale; +#endif + /* floor(log2(rv)) == bbits - 1 + p2 */ + /* Check for denormal case. */ + i = P - bbits; + if (i > (j = P - Emin - 1 + p2)) { +#ifdef Sudden_Underflow + Bfree(b MTa); + b = i2b(1 MTa); + p2 = Emin; + i = P - 1; +#ifdef Avoid_Underflow + word0(rv) = (1 + bc->scale) << Exp_shift; +#else + word0(rv) = Exp_msk1; +#endif + word1(rv) = 0; +#else + i = j; +#endif + } +#ifdef Honor_FLT_ROUNDS + if (bc->rounding != 1) { + if (i > 0) + b = lshift(b, i MTa); + if (dsign) + b = increment(b MTa); + } + else +#endif + { + b = lshift(b, ++i MTa); + b->x[0] |= 1; + } +#ifndef Sudden_Underflow + have_i: +#endif + p2 -= p5 + i; + d = i2b(1 MTa); + /* Arrange for convenient computation of quotients: + * shift left if necessary so divisor has 4 leading 0 bits. + */ + if (p5 > 0) + d = pow5mult(d, p5 MTa); + else if (p5 < 0) + b = pow5mult(b, -p5 MTa); + if (p2 > 0) { + b2 = p2; + d2 = 0; + } + else { + b2 = 0; + d2 = -p2; + } + i = dshift(d, d2); + if ((b2 += i) > 0) + b = lshift(b, b2 MTa); + if ((d2 += i) > 0) + d = lshift(d, d2 MTa); + + /* Now b/d = exactly half-way between the two floating-point values */ + /* on either side of the input string. Compute first digit of b/d. */ + + if (!(dig = quorem(b,d))) { + b = multadd(b, 10, 0 MTa); /* very unlikely */ + dig = quorem(b,d); + } + + /* Compare b/d with s0 */ + + for(i = 0; i < nd0; ) { + if ((dd = s0[i++] - '0' - dig)) + goto ret; + if (!b->x[0] && b->wds == 1) { + if (i < nd) + dd = 1; + goto ret; + } + b = multadd(b, 10, 0 MTa); + dig = quorem(b,d); + } + for(j = bc->dp1; i++ < nd;) { + if ((dd = s0[j++] - '0' - dig)) + goto ret; + if (!b->x[0] && b->wds == 1) { + if (i < nd) + dd = 1; + goto ret; + } + b = multadd(b, 10, 0 MTa); + dig = quorem(b,d); + } + if (dig > 0 || b->x[0] || b->wds > 1) + dd = -1; + ret: + Bfree(b MTa); + Bfree(d MTa); +#ifdef Honor_FLT_ROUNDS + if (bc->rounding != 1) { + if (dd < 0) { + if (bc->rounding == 0) { + if (!dsign) + goto retlow1; + } + else if (dsign) + goto rethi1; + } + else if (dd > 0) { + if (bc->rounding == 0) { + if (dsign) + goto rethi1; + goto ret1; + } + if (!dsign) + goto rethi1; + dval(rv) += 2.*sulp(rv,bc); + } + else { + bc->inexact = 0; + if (dsign) + goto rethi1; + } + } + else +#endif + if (speccase) { + if (dd <= 0) + rv->d = 0.; + } + else if (dd < 0) { + if (!dsign) /* does not happen for round-near */ +retlow1: + dval(rv) -= sulp(rv,bc); + } + else if (dd > 0) { + if (dsign) { + rethi1: + dval(rv) += sulp(rv,bc); + } + } + else { + /* Exact half-way case: apply round-even rule. */ + if ((j = ((word0(rv) & Exp_mask) >> Exp_shift) - bc->scale) <= 0) { + i = 1 - j; + if (i <= 31) { + if (word1(rv) & (0x1 << i)) + goto odd; + } + else if (word0(rv) & (0x1 << (i-32))) + goto odd; + } + else if (word1(rv) & 1) { + odd: + if (dsign) + goto rethi1; + goto retlow1; + } + } + +#ifdef Honor_FLT_ROUNDS + ret1: +#endif + return; + } +#endif /* NO_STRTOD_BIGCOMP */ + +#if 0 + double +strtod(const char *s00, char **se) +{ + int bb2, bb5, bbe, bd2, bd5, bbbits, bs2, c, e, e1; + int esign, i, j, k, nd, nd0, nf, nz, nz0, nz1, sign; + const char *s, *s0, *s1; + double aadj, aadj1; + Long L; + U aadj2, adj, rv, rv0; + ULong y, z; + BCinfo bc; + Bigint *bb, *bb1, *bd, *bd0, *bs, *delta; +#ifdef USE_BF96 + ULLong bhi, blo, brv, t00, t01, t02, t10, t11, terv, tg, tlo, yz; + const BF96 *p10; + int bexact, erv; +#endif +#ifdef Avoid_Underflow + ULong Lsb, Lsb1; +#endif +#ifdef SET_INEXACT + int oldinexact; +#endif +#ifndef NO_STRTOD_BIGCOMP + int req_bigcomp = 0; +#endif +#ifdef MULTIPLE_THREADS + ThInfo *TI = 0; +#endif +#ifdef Honor_FLT_ROUNDS /*{*/ +#ifdef Trust_FLT_ROUNDS /*{{ only define this if FLT_ROUNDS really works! */ + bc.rounding = Flt_Rounds; +#else /*}{*/ + bc.rounding = 1; + switch(fegetround()) { + case FE_TOWARDZERO: bc.rounding = 0; break; + case FE_UPWARD: bc.rounding = 2; break; + case FE_DOWNWARD: bc.rounding = 3; + } +#endif /*}}*/ +#endif /*}*/ +#ifdef USE_LOCALE + const char *s2; +#endif + + sign = nz0 = nz1 = nz = bc.dplen = bc.uflchk = 0; + dval(&rv) = 0.; + for(s = s00;;s++) switch(*s) { + case '-': + sign = 1; + /* no break */ + case '+': + if (*++s) + goto break2; + /* no break */ + case 0: + goto ret0; + case '\t': + case '\n': + case '\v': + case '\f': + case '\r': + case ' ': + continue; + default: + goto break2; + } + break2: + if (*s == '0') { +#ifndef NO_HEX_FP /*{*/ + switch(s[1]) { + case 'x': + case 'X': +#ifdef Honor_FLT_ROUNDS + gethex(&s, &rv, bc.rounding, sign MTb); +#else + gethex(&s, &rv, 1, sign MTb); +#endif + goto ret; + } +#endif /*}*/ + nz0 = 1; + while(*++s == '0') ; + if (!*s) + goto ret; + } + s0 = s; + nd = nf = 0; +#ifdef USE_BF96 + yz = 0; + for(; (c = *s) >= '0' && c <= '9'; nd++, s++) + if (nd < 19) + yz = 10*yz + c - '0'; +#else + y = z = 0; + for(; (c = *s) >= '0' && c <= '9'; nd++, s++) + if (nd < 9) + y = 10*y + c - '0'; + else if (nd < DBL_DIG + 2) + z = 10*z + c - '0'; +#endif + nd0 = nd; + bc.dp0 = bc.dp1 = s - s0; + for(s1 = s; s1 > s0 && *--s1 == '0'; ) + ++nz1; +#ifdef USE_LOCALE + s1 = localeconv()->decimal_point; + if (c == *s1) { + c = '.'; + if (*++s1) { + s2 = s; + for(;;) { + if (*++s2 != *s1) { + c = 0; + break; + } + if (!*++s1) { + s = s2; + break; + } + } + } + } +#endif + if (c == '.') { + c = *++s; + bc.dp1 = s - s0; + bc.dplen = bc.dp1 - bc.dp0; + if (!nd) { + for(; c == '0'; c = *++s) + nz++; + if (c > '0' && c <= '9') { + bc.dp0 = s0 - s; + bc.dp1 = bc.dp0 + bc.dplen; + s0 = s; + nf += nz; + nz = 0; + goto have_dig; + } + goto dig_done; + } + for(; c >= '0' && c <= '9'; c = *++s) { + have_dig: + nz++; + if (c -= '0') { + nf += nz; + i = 1; +#ifdef USE_BF96 + for(; i < nz; ++i) { + if (++nd <= 19) + yz *= 10; + } + if (++nd <= 19) + yz = 10*yz + c; +#else + for(; i < nz; ++i) { + if (nd++ < 9) + y *= 10; + else if (nd <= DBL_DIG + 2) + z *= 10; + } + if (nd++ < 9) + y = 10*y + c; + else if (nd <= DBL_DIG + 2) + z = 10*z + c; +#endif + nz = nz1 = 0; + } + } + } + dig_done: + e = 0; + if (c == 'e' || c == 'E') { + if (!nd && !nz && !nz0) { + goto ret0; + } + s00 = s; + esign = 0; + switch(c = *++s) { + case '-': + esign = 1; + case '+': + c = *++s; + } + if (c >= '0' && c <= '9') { + while(c == '0') + c = *++s; + if (c > '0' && c <= '9') { + L = c - '0'; + s1 = s; + while((c = *++s) >= '0' && c <= '9') + L = 10*L + c - '0'; + if (s - s1 > 8 || L > 19999) + /* Avoid confusion from exponents + * so large that e might overflow. + */ + e = 19999; /* safe for 16 bit ints */ + else + e = (int)L; + if (esign) + e = -e; + } + else + e = 0; + } + else + s = s00; + } + if (!nd) { + if (!nz && !nz0) { +#ifdef INFNAN_CHECK /*{*/ + /* Check for Nan and Infinity */ + if (!bc.dplen) + switch(c) { + case 'i': + case 'I': + if (match(&s,"nf")) { + --s; + if (!match(&s,"inity")) + ++s; + word0(&rv) = 0x7ff00000; + word1(&rv) = 0; + goto ret; + } + break; + case 'n': + case 'N': + if (match(&s, "an")) { + word0(&rv) = NAN_WORD0; + word1(&rv) = NAN_WORD1; +#ifndef No_Hex_NaN + if (*s == '(') /*)*/ + hexnan(&rv, &s); +#endif + goto ret; + } + } +#endif /*} INFNAN_CHECK */ + ret0: + s = s00; + sign = 0; + } + goto ret; + } + bc.e0 = e1 = e -= nf; + + /* Now we have nd0 digits, starting at s0, followed by a + * decimal point, followed by nd-nd0 digits. The number we're + * after is the integer represented by those digits times + * 10**e */ + + if (!nd0) + nd0 = nd; +#ifndef USE_BF96 + k = nd < DBL_DIG + 2 ? nd : DBL_DIG + 2; + dval(&rv) = y; + if (k > 9) { +#ifdef SET_INEXACT + if (k > DBL_DIG) + oldinexact = get_inexact(); +#endif + dval(&rv) = tens[k - 9] * dval(&rv) + z; + } +#endif + bd0 = 0; + if (nd <= DBL_DIG +#ifndef RND_PRODQUOT +#ifndef Honor_FLT_ROUNDS + && Flt_Rounds == 1 +#endif +#endif + ) { +#ifdef USE_BF96 + dval(&rv) = yz; +#endif + if (!e) + goto ret; +#ifndef ROUND_BIASED_without_Round_Up + if (e > 0) { + if (e <= Ten_pmax) { +#ifdef SET_INEXACT + bc.inexact = 0; + oldinexact = 1; +#endif +#ifdef VAX + goto vax_ovfl_check; +#else +#ifdef Honor_FLT_ROUNDS + /* round correctly FLT_ROUNDS = 2 or 3 */ + if (sign) { + rv.d = -rv.d; + sign = 0; + } +#endif + /* rv = */ rounded_product(dval(&rv), tens[e]); + goto ret; +#endif + } + i = DBL_DIG - nd; + if (e <= Ten_pmax + i) { + /* A fancier test would sometimes let us do + * this for larger i values. + */ +#ifdef SET_INEXACT + bc.inexact = 0; + oldinexact = 1; +#endif +#ifdef Honor_FLT_ROUNDS + /* round correctly FLT_ROUNDS = 2 or 3 */ + if (sign) { + rv.d = -rv.d; + sign = 0; + } +#endif + e -= i; + dval(&rv) *= tens[i]; +#ifdef VAX + /* VAX exponent range is so narrow we must + * worry about overflow here... + */ + vax_ovfl_check: + word0(&rv) -= P*Exp_msk1; + /* rv = */ rounded_product(dval(&rv), tens[e]); + if ((word0(&rv) & Exp_mask) + > Exp_msk1*(DBL_MAX_EXP+Bias-1-P)) + goto ovfl; + word0(&rv) += P*Exp_msk1; +#else + /* rv = */ rounded_product(dval(&rv), tens[e]); +#endif + goto ret; + } + } +#ifndef Inaccurate_Divide + else if (e >= -Ten_pmax) { +#ifdef SET_INEXACT + bc.inexact = 0; + oldinexact = 1; +#endif +#ifdef Honor_FLT_ROUNDS + /* round correctly FLT_ROUNDS = 2 or 3 */ + if (sign) { + rv.d = -rv.d; + sign = 0; + } +#endif + /* rv = */ rounded_quotient(dval(&rv), tens[-e]); + goto ret; + } +#endif +#endif /* ROUND_BIASED_without_Round_Up */ + } +#ifdef USE_BF96 + k = nd < 19 ? nd : 19; +#endif + e1 += nd - k; /* scale factor = 10^e1 */ + +#ifdef IEEE_Arith +#ifdef SET_INEXACT + bc.inexact = 1; +#ifndef USE_BF96 + if (k <= DBL_DIG) +#endif + oldinexact = get_inexact(); +#endif +#ifdef Honor_FLT_ROUNDS + if (bc.rounding >= 2) { + if (sign) + bc.rounding = bc.rounding == 2 ? 0 : 2; + else + if (bc.rounding != 2) + bc.rounding = 0; + } +#endif +#endif /*IEEE_Arith*/ + +#ifdef USE_BF96 /*{*/ + Debug(++dtoa_stats[0]); + i = e1 + 342; + if (i < 0) + goto undfl; + if (i > 650) + goto ovfl; + p10 = &pten[i]; + brv = yz; + /* shift brv left, with i = number of bits shifted */ + i = 0; + if (!(brv & 0xffffffff00000000ull)) { + i = 32; + brv <<= 32; + } + if (!(brv & 0xffff000000000000ull)) { + i += 16; + brv <<= 16; + } + if (!(brv & 0xff00000000000000ull)) { + i += 8; + brv <<= 8; + } + if (!(brv & 0xf000000000000000ull)) { + i += 4; + brv <<= 4; + } + if (!(brv & 0xc000000000000000ull)) { + i += 2; + brv <<= 2; + } + if (!(brv & 0x8000000000000000ull)) { + i += 1; + brv <<= 1; + } + erv = (64 + 0x3fe) + p10->e - i; + if (erv <= 0 && nd > 19) + goto many_digits; /* denormal: may need to look at all digits */ + bhi = brv >> 32; + blo = brv & 0xffffffffull; + /* Unsigned 32-bit ints lie in [0,2^32-1] and */ + /* unsigned 64-bit ints lie in [0, 2^64-1]. The product of two unsigned */ + /* 32-bit ints is <= 2^64 - 2*2^32-1 + 1 = 2^64 - 1 - 2*(2^32 - 1), so */ + /* we can add two unsigned 32-bit ints to the product of two such ints, */ + /* and 64 bits suffice to contain the result. */ + t01 = bhi * p10->b1; + t10 = blo * p10->b0 + (t01 & 0xffffffffull); + t00 = bhi * p10->b0 + (t01 >> 32) + (t10 >> 32); + if (t00 & 0x8000000000000000ull) { + if ((t00 & 0x3ff) && (~t00 & 0x3fe)) { /* unambiguous result? */ + if (nd > 19 && ((t00 + (1< 19 && ((t00 + (1<b2; + t11 = blo * p10->b1 + (t02 & 0xffffffffull); + bexact = 1; + if (e1 < 0 || e1 > 41 || (t10 | t11) & 0xffffffffull || nd > 19) + bexact = 0; + tlo = (t10 & 0xffffffffull) + (t02 >> 32) + (t11 >> 32); + if (!bexact && (tlo + 0x10) >> 32 > tlo >> 32) + goto many_digits; + t00 += tlo >> 32; + if (t00 & 0x8000000000000000ull) { + if (erv <= 0) { /* denormal result */ + if (nd >= 20 || !((tlo & 0xfffffff0) | (t00 & 0x3ff))) + goto many_digits; + denormal: + if (erv <= -52) { +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 0: goto undfl; + case 2: goto tiniest; + } +#endif + if (erv < -52 || !(t00 & 0x7fffffffffffffffull)) + goto undfl; + goto tiniest; + } + tg = 1ull << (11 - erv); + t00 &= ~(tg - 1); /* clear low bits */ +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 0: goto noround_den; + case 2: goto roundup_den; + } +#endif + if (t00 & tg) { +#ifdef Honor_FLT_ROUNDS + roundup_den: +#endif + t00 += tg << 1; + if (!(t00 & 0x8000000000000000ull)) { + if (++erv > 0) + goto smallest_normal; + t00 = 0x8000000000000000ull; + } + } +#ifdef Honor_FLT_ROUNDS + noround_den: +#endif + LLval(&rv) = t00 >> (12 - erv); + Set_errno(ERANGE); + goto ret; + } + if (bexact) { +#ifdef SET_INEXACT + if (!(t00 & 0x7ff) && !(tlo & 0xffffffffull)) { + bc.inexact = 0; + goto noround; + } +#endif +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 2: + if (t00 & 0x7ff) + goto roundup; + case 0: goto noround; + } +#endif + if (t00 & 0x400 && (tlo & 0xffffffff) | (t00 & 0xbff)) + goto roundup; + goto noround; + } + if ((tlo & 0xfffffff0) | (t00 & 0x3ff) + && (nd <= 19 || ((t00 + (1ull << i)) & 0xfffffffffffffc00ull) + == (t00 & 0xfffffffffffffc00ull))) { + /* Unambiguous result. */ + /* If nd > 19, then incrementing the 19th digit */ + /* does not affect rv. */ +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 0: goto noround; + case 2: goto roundup; + } +#endif + if (t00 & 0x400) { /* round up */ + roundup: + t00 += 0x800; + if (!(t00 & 0x8000000000000000ull)) { + /* rounded up to a power of 2 */ + if (erv >= 0x7fe) + goto ovfl; + terv = erv + 1; + LLval(&rv) = terv << 52; + goto ret; + } + } + noround: + if (erv >= 0x7ff) + goto ovfl; + terv = erv; + LLval(&rv) = (terv << 52) | ((t00 & 0x7ffffffffffff800ull) >> 11); + goto ret; + } + } + else { + if (erv <= 1) { /* denormal result */ + if (nd >= 20 || !((tlo & 0xfffffff0) | (t00 & 0x1ff))) + goto many_digits; + denormal1: + if (erv <= -51) { +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 0: goto undfl; + case 2: goto tiniest; + } +#endif + if (erv < -51 || !(t00 & 0x3fffffffffffffffull)) + goto undfl; + tiniest: + LLval(&rv) = 1; + Set_errno(ERANGE); + goto ret; + } + tg = 1ull << (11 - erv); +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 0: goto noround1_den; + case 2: goto roundup1_den; + } +#endif + if (t00 & tg) { +#ifdef Honor_FLT_ROUNDS + roundup1_den: +#endif + if (0x8000000000000000ull & (t00 += (tg<<1)) && erv == 1) { + + smallest_normal: + LLval(&rv) = 0x0010000000000000ull; + goto ret; + } + } +#ifdef Honor_FLT_ROUNDS + noround1_den: +#endif + if (erv <= -52) + goto undfl; + LLval(&rv) = t00 >> (12 - erv); + Set_errno(ERANGE); + goto ret; + } + if (bexact) { +#ifdef SET_INEXACT + if (!(t00 & 0x3ff) && !(tlo & 0xffffffffull)) { + bc.inexact = 0; + goto noround1; + } +#endif +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 2: + if (t00 & 0x3ff) + goto roundup1; + case 0: goto noround1; + } +#endif + if (t00 & 0x200 && (t00 & 0x5ff || tlo)) + goto roundup1; + goto noround1; + } + if ((tlo & 0xfffffff0) | (t00 & 0x1ff) + && (nd <= 19 || ((t00 + (1ull << i)) & 0x7ffffffffffffe00ull) + == (t00 & 0x7ffffffffffffe00ull))) { + /* Unambiguous result. */ +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 0: goto noround1; + case 2: goto roundup1; + } +#endif + if (t00 & 0x200) { /* round up */ + roundup1: + t00 += 0x400; + if (!(t00 & 0x4000000000000000ull)) { + /* rounded up to a power of 2 */ + if (erv >= 0x7ff) + goto ovfl; + terv = erv; + LLval(&rv) = terv << 52; + goto ret; + } + } + noround1: + if (erv >= 0x800) + goto ovfl; + terv = erv - 1; + LLval(&rv) = (terv << 52) | ((t00 & 0x3ffffffffffffc00ull) >> 10); + goto ret; + } + } + many_digits: + Debug(++dtoa_stats[2]); + if (nd > 17) { + if (nd > 18) { + yz /= 100; + e1 += 2; + } + else { + yz /= 10; + e1 += 1; + } + y = yz / 100000000; + } + else if (nd > 9) { + i = nd - 9; + y = (yz >> i) / pfive[i-1]; + } + else + y = yz; + dval(&rv) = yz; +#endif /*}*/ + +#ifdef IEEE_Arith +#ifdef Avoid_Underflow + bc.scale = 0; +#endif +#endif /*IEEE_Arith*/ + + /* Get starting approximation = rv * 10**e1 */ + + if (e1 > 0) { + if ((i = e1 & 15)) + dval(&rv) *= tens[i]; + if (e1 &= ~15) { + if (e1 > DBL_MAX_10_EXP) { + ovfl: + /* Can't trust HUGE_VAL */ +#ifdef IEEE_Arith +#ifdef Honor_FLT_ROUNDS + switch(bc.rounding) { + case 0: /* toward 0 */ + case 3: /* toward -infinity */ + word0(&rv) = Big0; + word1(&rv) = Big1; + break; + default: + word0(&rv) = Exp_mask; + word1(&rv) = 0; + } +#else /*Honor_FLT_ROUNDS*/ + word0(&rv) = Exp_mask; + word1(&rv) = 0; +#endif /*Honor_FLT_ROUNDS*/ +#ifdef SET_INEXACT + /* set overflow bit */ + dval(&rv0) = 1e300; + dval(&rv0) *= dval(&rv0); +#endif +#else /*IEEE_Arith*/ + word0(&rv) = Big0; + word1(&rv) = Big1; +#endif /*IEEE_Arith*/ + range_err: + if (bd0) { + Bfree(bb MTb); + Bfree(bd MTb); + Bfree(bs MTb); + Bfree(bd0 MTb); + Bfree(delta MTb); + } + Set_errno(ERANGE); + goto ret; + } + e1 >>= 4; + for(j = 0; e1 > 1; j++, e1 >>= 1) + if (e1 & 1) + dval(&rv) *= bigtens[j]; + /* The last multiplication could overflow. */ + word0(&rv) -= P*Exp_msk1; + dval(&rv) *= bigtens[j]; + if ((z = word0(&rv) & Exp_mask) + > Exp_msk1*(DBL_MAX_EXP+Bias-P)) + goto ovfl; + if (z > Exp_msk1*(DBL_MAX_EXP+Bias-1-P)) { + /* set to largest number */ + /* (Can't trust DBL_MAX) */ + word0(&rv) = Big0; + word1(&rv) = Big1; + } + else + word0(&rv) += P*Exp_msk1; + } + } + else if (e1 < 0) { + e1 = -e1; + if ((i = e1 & 15)) + dval(&rv) /= tens[i]; + if (e1 >>= 4) { + if (e1 >= 1 << n_bigtens) + goto undfl; +#ifdef Avoid_Underflow + if (e1 & Scale_Bit) + bc.scale = 2*P; + for(j = 0; e1 > 0; j++, e1 >>= 1) + if (e1 & 1) + dval(&rv) *= tinytens[j]; + if (bc.scale && (j = 2*P + 1 - ((word0(&rv) & Exp_mask) + >> Exp_shift)) > 0) { + /* scaled rv is denormal; clear j low bits */ + if (j >= 32) { + if (j > 54) + goto undfl; + word1(&rv) = 0; + if (j >= 53) + word0(&rv) = (P+2)*Exp_msk1; + else + word0(&rv) &= 0xffffffff << (j-32); + } + else + word1(&rv) &= 0xffffffff << j; + } +#else + for(j = 0; e1 > 1; j++, e1 >>= 1) + if (e1 & 1) + dval(&rv) *= tinytens[j]; + /* The last multiplication could underflow. */ + dval(&rv0) = dval(&rv); + dval(&rv) *= tinytens[j]; + if (!dval(&rv)) { + dval(&rv) = 2.*dval(&rv0); + dval(&rv) *= tinytens[j]; +#endif + if (!dval(&rv)) { + undfl: + dval(&rv) = 0.; +#ifdef Honor_FLT_ROUNDS + if (bc.rounding == 2) + word1(&rv) = 1; +#endif + goto range_err; + } +#ifndef Avoid_Underflow + word0(&rv) = Tiny0; + word1(&rv) = Tiny1; + /* The refinement below will clean + * this approximation up. + */ + } +#endif + } + } + + /* Now the hard part -- adjusting rv to the correct value.*/ + + /* Put digits into bd: true value = bd * 10^e */ + + bc.nd = nd - nz1; +#ifndef NO_STRTOD_BIGCOMP + bc.nd0 = nd0; /* Only needed if nd > strtod_diglim, but done here */ + /* to silence an erroneous warning about bc.nd0 */ + /* possibly not being initialized. */ + if (nd > strtod_diglim) { + /* ASSERT(strtod_diglim >= 18); 18 == one more than the */ + /* minimum number of decimal digits to distinguish double values */ + /* in IEEE arithmetic. */ + i = j = 18; + if (i > nd0) + j += bc.dplen; + for(;;) { + if (--j < bc.dp1 && j >= bc.dp0) + j = bc.dp0 - 1; + if (s0[j] != '0') + break; + --i; + } + e += nd - i; + nd = i; + if (nd0 > nd) + nd0 = nd; + if (nd < 9) { /* must recompute y */ + y = 0; + for(i = 0; i < nd0; ++i) + y = 10*y + s0[i] - '0'; + for(j = bc.dp1; i < nd; ++i) + y = 10*y + s0[j++] - '0'; + } + } +#endif + bd0 = s2b(s0, nd0, nd, y, bc.dplen MTb); + + for(;;) { + bd = Balloc(bd0->k MTb); + Bcopy(bd, bd0); + bb = d2b(&rv, &bbe, &bbbits MTb); /* rv = bb * 2^bbe */ + bs = i2b(1 MTb); + + if (e >= 0) { + bb2 = bb5 = 0; + bd2 = bd5 = e; + } + else { + bb2 = bb5 = -e; + bd2 = bd5 = 0; + } + if (bbe >= 0) + bb2 += bbe; + else + bd2 -= bbe; + bs2 = bb2; +#ifdef Honor_FLT_ROUNDS + if (bc.rounding != 1) + bs2++; +#endif +#ifdef Avoid_Underflow + Lsb = LSB; + Lsb1 = 0; + j = bbe - bc.scale; + i = j + bbbits - 1; /* logb(rv) */ + j = P + 1 - bbbits; + if (i < Emin) { /* denormal */ + i = Emin - i; + j -= i; + if (i < 32) + Lsb <<= i; + else if (i < 52) + Lsb1 = Lsb << (i-32); + else + Lsb1 = Exp_mask; + } +#else /*Avoid_Underflow*/ +#ifdef Sudden_Underflow +#ifdef IBM + j = 1 + 4*P - 3 - bbbits + ((bbe + bbbits - 1) & 3); +#else + j = P + 1 - bbbits; +#endif +#else /*Sudden_Underflow*/ + j = bbe; + i = j + bbbits - 1; /* logb(rv) */ + if (i < Emin) /* denormal */ + j += P - Emin; + else + j = P + 1 - bbbits; +#endif /*Sudden_Underflow*/ +#endif /*Avoid_Underflow*/ + bb2 += j; + bd2 += j; +#ifdef Avoid_Underflow + bd2 += bc.scale; +#endif + i = bb2 < bd2 ? bb2 : bd2; + if (i > bs2) + i = bs2; + if (i > 0) { + bb2 -= i; + bd2 -= i; + bs2 -= i; + } + if (bb5 > 0) { + bs = pow5mult(bs, bb5 MTb); + bb1 = mult(bs, bb MTb); + Bfree(bb MTb); + bb = bb1; + } + if (bb2 > 0) + bb = lshift(bb, bb2 MTb); + if (bd5 > 0) + bd = pow5mult(bd, bd5 MTb); + if (bd2 > 0) + bd = lshift(bd, bd2 MTb); + if (bs2 > 0) + bs = lshift(bs, bs2 MTb); + delta = diff(bb, bd MTb); + bc.dsign = delta->sign; + delta->sign = 0; + i = cmp(delta, bs); +#ifndef NO_STRTOD_BIGCOMP /*{*/ + if (bc.nd > nd && i <= 0) { + if (bc.dsign) { + /* Must use bigcomp(). */ + req_bigcomp = 1; + break; + } +#ifdef Honor_FLT_ROUNDS + if (bc.rounding != 1) { + if (i < 0) { + req_bigcomp = 1; + break; + } + } + else +#endif + i = -1; /* Discarded digits make delta smaller. */ + } +#endif /*}*/ +#ifdef Honor_FLT_ROUNDS /*{*/ + if (bc.rounding != 1) { + if (i < 0) { + /* Error is less than an ulp */ + if (!delta->x[0] && delta->wds <= 1) { + /* exact */ +#ifdef SET_INEXACT + bc.inexact = 0; +#endif + break; + } + if (bc.rounding) { + if (bc.dsign) { + adj.d = 1.; + goto apply_adj; + } + } + else if (!bc.dsign) { + adj.d = -1.; + if (!word1(&rv) + && !(word0(&rv) & Frac_mask)) { + y = word0(&rv) & Exp_mask; +#ifdef Avoid_Underflow + if (!bc.scale || y > 2*P*Exp_msk1) +#else + if (y) +#endif + { + delta = lshift(delta,Log2P MTb); + if (cmp(delta, bs) <= 0) + adj.d = -0.5; + } + } + apply_adj: +#ifdef Avoid_Underflow /*{*/ + if (bc.scale && (y = word0(&rv) & Exp_mask) + <= 2*P*Exp_msk1) + word0(&adj) += (2*P+1)*Exp_msk1 - y; +#else +#ifdef Sudden_Underflow + if ((word0(&rv) & Exp_mask) <= + P*Exp_msk1) { + word0(&rv) += P*Exp_msk1; + dval(&rv) += adj.d*ulp(dval(&rv)); + word0(&rv) -= P*Exp_msk1; + } + else +#endif /*Sudden_Underflow*/ +#endif /*Avoid_Underflow}*/ + dval(&rv) += adj.d*ulp(&rv); + } + break; + } + adj.d = ratio(delta, bs); + if (adj.d < 1.) + adj.d = 1.; + if (adj.d <= 0x7ffffffe) { + /* adj = rounding ? ceil(adj) : floor(adj); */ + y = adj.d; + if (y != adj.d) { + if (!((bc.rounding>>1) ^ bc.dsign)) + y++; + adj.d = y; + } + } +#ifdef Avoid_Underflow /*{*/ + if (bc.scale && (y = word0(&rv) & Exp_mask) <= 2*P*Exp_msk1) + word0(&adj) += (2*P+1)*Exp_msk1 - y; +#else +#ifdef Sudden_Underflow + if ((word0(&rv) & Exp_mask) <= P*Exp_msk1) { + word0(&rv) += P*Exp_msk1; + adj.d *= ulp(dval(&rv)); + if (bc.dsign) + dval(&rv) += adj.d; + else + dval(&rv) -= adj.d; + word0(&rv) -= P*Exp_msk1; + goto cont; + } +#endif /*Sudden_Underflow*/ +#endif /*Avoid_Underflow}*/ + adj.d *= ulp(&rv); + if (bc.dsign) { + if (word0(&rv) == Big0 && word1(&rv) == Big1) + goto ovfl; + dval(&rv) += adj.d; + } + else + dval(&rv) -= adj.d; + goto cont; + } +#endif /*}Honor_FLT_ROUNDS*/ + + if (i < 0) { + /* Error is less than half an ulp -- check for + * special case of mantissa a power of two. + */ + if (bc.dsign || word1(&rv) || word0(&rv) & Bndry_mask +#ifdef IEEE_Arith /*{*/ +#ifdef Avoid_Underflow + || (word0(&rv) & Exp_mask) <= (2*P+1)*Exp_msk1 +#else + || (word0(&rv) & Exp_mask) <= Exp_msk1 +#endif +#endif /*}*/ + ) { +#ifdef SET_INEXACT + if (!delta->x[0] && delta->wds <= 1) + bc.inexact = 0; +#endif + break; + } + if (!delta->x[0] && delta->wds <= 1) { + /* exact result */ +#ifdef SET_INEXACT + bc.inexact = 0; +#endif + break; + } + delta = lshift(delta,Log2P MTb); + if (cmp(delta, bs) > 0) + goto drop_down; + break; + } + if (i == 0) { + /* exactly half-way between */ + if (bc.dsign) { + if ((word0(&rv) & Bndry_mask1) == Bndry_mask1 + && word1(&rv) == ( +#ifdef Avoid_Underflow + (bc.scale && (y = word0(&rv) & Exp_mask) <= 2*P*Exp_msk1) + ? (0xffffffff & (0xffffffff << (2*P+1-(y>>Exp_shift)))) : +#endif + 0xffffffff)) { + /*boundary case -- increment exponent*/ + if (word0(&rv) == Big0 && word1(&rv) == Big1) + goto ovfl; + word0(&rv) = (word0(&rv) & Exp_mask) + + Exp_msk1 +#ifdef IBM + | Exp_msk1 >> 4 +#endif + ; + word1(&rv) = 0; +#ifdef Avoid_Underflow + bc.dsign = 0; +#endif + break; + } + } + else if (!(word0(&rv) & Bndry_mask) && !word1(&rv)) { + drop_down: + /* boundary case -- decrement exponent */ +#ifdef Sudden_Underflow /*{{*/ + L = word0(&rv) & Exp_mask; +#ifdef IBM + if (L < Exp_msk1) +#else +#ifdef Avoid_Underflow + if (L <= (bc.scale ? (2*P+1)*Exp_msk1 : Exp_msk1)) +#else + if (L <= Exp_msk1) +#endif /*Avoid_Underflow*/ +#endif /*IBM*/ + { + if (bc.nd >nd) { + bc.uflchk = 1; + break; + } + goto undfl; + } + L -= Exp_msk1; +#else /*Sudden_Underflow}{*/ +#ifdef Avoid_Underflow + if (bc.scale) { + L = word0(&rv) & Exp_mask; + if (L <= (2*P+1)*Exp_msk1) { + if (L > (P+2)*Exp_msk1) + /* round even ==> */ + /* accept rv */ + break; + /* rv = smallest denormal */ + if (bc.nd >nd) { + bc.uflchk = 1; + break; + } + goto undfl; + } + } +#endif /*Avoid_Underflow*/ + L = (word0(&rv) & Exp_mask) - Exp_msk1; +#endif /*Sudden_Underflow}}*/ + word0(&rv) = L | Bndry_mask1; + word1(&rv) = 0xffffffff; +#ifdef IBM + goto cont; +#else +#ifndef NO_STRTOD_BIGCOMP + if (bc.nd > nd) + goto cont; +#endif + break; +#endif + } +#ifndef ROUND_BIASED +#ifdef Avoid_Underflow + if (Lsb1) { + if (!(word0(&rv) & Lsb1)) + break; + } + else if (!(word1(&rv) & Lsb)) + break; +#else + if (!(word1(&rv) & LSB)) + break; +#endif +#endif + if (bc.dsign) +#ifdef Avoid_Underflow + dval(&rv) += sulp(&rv, &bc); +#else + dval(&rv) += ulp(&rv); +#endif +#ifndef ROUND_BIASED + else { +#ifdef Avoid_Underflow + dval(&rv) -= sulp(&rv, &bc); +#else + dval(&rv) -= ulp(&rv); +#endif +#ifndef Sudden_Underflow + if (!dval(&rv)) { + if (bc.nd >nd) { + bc.uflchk = 1; + break; + } + goto undfl; + } +#endif + } +#ifdef Avoid_Underflow + bc.dsign = 1 - bc.dsign; +#endif +#endif + break; + } + if ((aadj = ratio(delta, bs)) <= 2.) { + if (bc.dsign) + aadj = aadj1 = 1.; + else if (word1(&rv) || word0(&rv) & Bndry_mask) { +#ifndef Sudden_Underflow + if (word1(&rv) == Tiny1 && !word0(&rv)) { + if (bc.nd >nd) { + bc.uflchk = 1; + break; + } + goto undfl; + } +#endif + aadj = 1.; + aadj1 = -1.; + } + else { + /* special case -- power of FLT_RADIX to be */ + /* rounded down... */ + + if (aadj < 2./FLT_RADIX) + aadj = 1./FLT_RADIX; + else + aadj *= 0.5; + aadj1 = -aadj; + } + } + else { + aadj *= 0.5; + aadj1 = bc.dsign ? aadj : -aadj; +#ifdef Check_FLT_ROUNDS + switch(bc.rounding) { + case 2: /* towards +infinity */ + aadj1 -= 0.5; + break; + case 0: /* towards 0 */ + case 3: /* towards -infinity */ + aadj1 += 0.5; + } +#else + if (Flt_Rounds == 0) + aadj1 += 0.5; +#endif /*Check_FLT_ROUNDS*/ + } + y = word0(&rv) & Exp_mask; + + /* Check for overflow */ + + if (y == Exp_msk1*(DBL_MAX_EXP+Bias-1)) { + dval(&rv0) = dval(&rv); + word0(&rv) -= P*Exp_msk1; + adj.d = aadj1 * ulp(&rv); + dval(&rv) += adj.d; + if ((word0(&rv) & Exp_mask) >= + Exp_msk1*(DBL_MAX_EXP+Bias-P)) { + if (word0(&rv0) == Big0 && word1(&rv0) == Big1) + goto ovfl; + word0(&rv) = Big0; + word1(&rv) = Big1; + goto cont; + } + else + word0(&rv) += P*Exp_msk1; + } + else { +#ifdef Avoid_Underflow + if (bc.scale && y <= 2*P*Exp_msk1) { + if (aadj <= 0x7fffffff) { + if ((z = aadj) <= 0) + z = 1; + aadj = z; + aadj1 = bc.dsign ? aadj : -aadj; + } + dval(&aadj2) = aadj1; + word0(&aadj2) += (2*P+1)*Exp_msk1 - y; + aadj1 = dval(&aadj2); + adj.d = aadj1 * ulp(&rv); + dval(&rv) += adj.d; + if (rv.d == 0.) +#ifdef NO_STRTOD_BIGCOMP + goto undfl; +#else + { + req_bigcomp = 1; + break; + } +#endif + } + else { + adj.d = aadj1 * ulp(&rv); + dval(&rv) += adj.d; + } +#else +#ifdef Sudden_Underflow + if ((word0(&rv) & Exp_mask) <= P*Exp_msk1) { + dval(&rv0) = dval(&rv); + word0(&rv) += P*Exp_msk1; + adj.d = aadj1 * ulp(&rv); + dval(&rv) += adj.d; +#ifdef IBM + if ((word0(&rv) & Exp_mask) < P*Exp_msk1) +#else + if ((word0(&rv) & Exp_mask) <= P*Exp_msk1) +#endif + { + if (word0(&rv0) == Tiny0 + && word1(&rv0) == Tiny1) { + if (bc.nd >nd) { + bc.uflchk = 1; + break; + } + goto undfl; + } + word0(&rv) = Tiny0; + word1(&rv) = Tiny1; + goto cont; + } + else + word0(&rv) -= P*Exp_msk1; + } + else { + adj.d = aadj1 * ulp(&rv); + dval(&rv) += adj.d; + } +#else /*Sudden_Underflow*/ + /* Compute adj so that the IEEE rounding rules will + * correctly round rv + adj in some half-way cases. + * If rv * ulp(rv) is denormalized (i.e., + * y <= (P-1)*Exp_msk1), we must adjust aadj to avoid + * trouble from bits lost to denormalization; + * example: 1.2e-307 . + */ + if (y <= (P-1)*Exp_msk1 && aadj > 1.) { + aadj1 = (double)(int)(aadj + 0.5); + if (!bc.dsign) + aadj1 = -aadj1; + } + adj.d = aadj1 * ulp(&rv); + dval(&rv) += adj.d; +#endif /*Sudden_Underflow*/ +#endif /*Avoid_Underflow*/ + } + z = word0(&rv) & Exp_mask; +#ifndef SET_INEXACT + if (bc.nd == nd) { +#ifdef Avoid_Underflow + if (!bc.scale) +#endif + if (y == z) { + /* Can we stop now? */ + L = (Long)aadj; + aadj -= L; + /* The tolerances below are conservative. */ + if (bc.dsign || word1(&rv) || word0(&rv) & Bndry_mask) { + if (aadj < .4999999 || aadj > .5000001) + break; + } + else if (aadj < .4999999/FLT_RADIX) + break; + } + } +#endif + cont: + Bfree(bb MTb); + Bfree(bd MTb); + Bfree(bs MTb); + Bfree(delta MTb); + } + Bfree(bb MTb); + Bfree(bd MTb); + Bfree(bs MTb); + Bfree(bd0 MTb); + Bfree(delta MTb); +#ifndef NO_STRTOD_BIGCOMP + if (req_bigcomp) { + bd0 = 0; + bc.e0 += nz1; + bigcomp(&rv, s0, &bc MTb); + y = word0(&rv) & Exp_mask; + if (y == Exp_mask) + goto ovfl; + if (y == 0 && rv.d == 0.) + goto undfl; + } +#endif +#ifdef Avoid_Underflow + if (bc.scale) { + word0(&rv0) = Exp_1 - 2*P*Exp_msk1; + word1(&rv0) = 0; + dval(&rv) *= dval(&rv0); +#ifndef NO_ERRNO + /* try to avoid the bug of testing an 8087 register value */ +#ifdef IEEE_Arith + if (!(word0(&rv) & Exp_mask)) +#else + if (word0(&rv) == 0 && word1(&rv) == 0) +#endif + Set_errno(ERANGE); +#endif + } +#endif /* Avoid_Underflow */ + ret: +#ifdef SET_INEXACT + if (bc.inexact) { + if (!(word0(&rv) & Exp_mask)) { + /* set underflow and inexact bits */ + dval(&rv0) = 1e-300; + dval(&rv0) *= dval(&rv0); + } + else if (!oldinexact) { + word0(&rv0) = Exp_1 + (70 << Exp_shift); + word1(&rv0) = 0; + dval(&rv0) += 1.; + } + } + else if (!oldinexact) + clear_inexact(); +#endif + if (se) + *se = (char *)s; + return sign ? -dval(&rv) : dval(&rv); + } + +#endif + +#ifndef MULTIPLE_THREADS + static char *dtoa_result; +#endif + + static char * +rv_alloc(int i MTd) +{ + int j, k, *r; + + j = sizeof(ULong); + for(k = 0; + sizeof(Bigint) - sizeof(ULong) - sizeof(int) + j <= i; + j <<= 1) + k++; + r = (int*)Balloc(k MTa); + *r = k; + return +#ifndef MULTIPLE_THREADS + dtoa_result = +#endif + (char *)(r+1); + } + + static char * +nrv_alloc(const char *s, char *s0, size_t s0len, char **rve, int n MTd) +{ + char *rv, *t; + + if (!s0) + s0 = rv_alloc(n MTa); + else if (s0len <= n) { + rv = 0; + t = rv + n; + goto rve_chk; + } + t = rv = s0; + while((*t = *s++)) + ++t; + rve_chk: + if (rve) + *rve = t; + return rv; + } + +/* freedtoa(s) must be used to free values s returned by dtoa + * when MULTIPLE_THREADS is #defined. It should be used in all cases, + * but for consistency with earlier versions of dtoa, it is optional + * when MULTIPLE_THREADS is not defined. + */ + +static void +freedtoa(char *s) +{ +#ifdef MULTIPLE_THREADS + ThInfo *TI = 0; +#endif + Bigint *b = (Bigint *)((int *)s - 1); + b->maxwds = 1 << (b->k = *(int*)b); + Bfree(b MTb); +#ifndef MULTIPLE_THREADS + if (s == dtoa_result) + dtoa_result = 0; +#endif + } + +/* dtoa for IEEE arithmetic (dmg): convert double to ASCII string. + * + * Inspired by "How to Print Floating-Point Numbers Accurately" by + * Guy L. Steele, Jr. and Jon L. White [Proc. ACM SIGPLAN '90, pp. 112-126]. + * + * Modifications: + * 1. Rather than iterating, we use a simple numeric overestimate + * to determine k = floor(log10(d)). We scale relevant + * quantities using O(log2(k)) rather than O(k) multiplications. + * 2. For some modes > 2 (corresponding to ecvt and fcvt), we don't + * try to generate digits strictly left to right. Instead, we + * compute with fewer bits and propagate the carry if necessary + * when rounding the final digit up. This is often faster. + * 3. Under the assumption that input will be rounded nearest, + * mode 0 renders 1e23 as 1e23 rather than 9.999999999999999e22. + * That is, we allow equality in stopping tests when the + * round-nearest rule will give the same floating-point value + * as would satisfaction of the stopping test with strict + * inequality. + * 4. We remove common factors of powers of 2 from relevant + * quantities. + * 5. When converting floating-point integers less than 1e16, + * we use floating-point arithmetic rather than resorting + * to multiple-precision integers. + * 6. When asked to produce fewer than 15 digits, we first try + * to get by with floating-point arithmetic; we resort to + * multiple-precision integer arithmetic only if we cannot + * guarantee that the floating-point calculation has given + * the correctly rounded result. For k requested digits and + * "uniformly" distributed input, the probability is + * something like 10^(k-15) that we must resort to the Long + * calculation. + */ + + char * +dtoa_r_dmgay(double dd, int mode, int ndigits, int *decpt, int *sign, char **rve, char *buf, size_t blen) +{ + /* Arguments ndigits, decpt, sign are similar to those + of ecvt and fcvt; trailing zeros are suppressed from + the returned string. If not null, *rve is set to point + to the end of the return value. If d is +-Infinity or NaN, + then *decpt is set to 9999. + + mode: + 0 ==> shortest string that yields d when read in + and rounded to nearest. + 1 ==> like 0, but with Steele & White stopping rule; + e.g. with IEEE P754 arithmetic , mode 0 gives + 1e23 whereas mode 1 gives 9.999999999999999e22. + 2 ==> max(1,ndigits) significant digits. This gives a + return value similar to that of ecvt, except + that trailing zeros are suppressed. + 3 ==> through ndigits past the decimal point. This + gives a return value similar to that from fcvt, + except that trailing zeros are suppressed, and + ndigits can be negative. + 4,5 ==> similar to 2 and 3, respectively, but (in + round-nearest mode) with the tests of mode 0 to + possibly return a shorter string that rounds to d. + With IEEE arithmetic and compilation with + -DHonor_FLT_ROUNDS, modes 4 and 5 behave the same + as modes 2 and 3 when FLT_ROUNDS != 1. + 6-9 ==> Debugging modes similar to mode - 4: don't try + fast floating-point estimate (if applicable). + + Values of mode other than 0-9 are treated as mode 0. + + When not NULL, buf is an output buffer of length blen, which must + be large enough to accommodate suppressed trailing zeros and a trailing + null byte. If blen is too small, rv = NULL is returned, in which case + if rve is not NULL, a subsequent call with blen >= (*rve - rv) + 1 + should succeed in returning buf. + + When buf is NULL, sufficient space is allocated for the return value, + which, when done using, the caller should pass to freedtoa(). + + USE_BF is automatically defined when neither NO_LONG_LONG nor NO_BF96 + is defined. + */ + +#ifdef MULTIPLE_THREADS + ThInfo *TI = 0; +#endif + int bbits, b2, b5, be, dig, i, ilim, ilim1, + j, j1, k, leftright, m2, m5, s2, s5, spec_case; +#if !defined(Sudden_Underflow) || defined(USE_BF96) + int denorm; +#endif + Bigint *b, *b1, *delta, *mlo, *mhi, *S; + U u; + char *s; +#ifdef SET_INEXACT + int inexact, oldinexact; +#endif +#ifdef USE_BF96 /*{{*/ + BF96 *p10; + ULLong dbhi, dbits, dblo, den, hb, rb, rblo, res, res0, res3, reslo, sres, + sulp, tv0, tv1, tv2, tv3, ulp, ulplo, ulpmask, ures, ureslo, zb; + int eulp, k1, n2, ulpadj, ulpshift; +#else /*}{*/ +#ifndef Sudden_Underflow + ULong x; +#endif + Long L; + U d2, eps; + double ds; + int ieps, ilim0, k0, k_check, try_quick; +#ifndef No_leftright +#ifdef IEEE_Arith + U eps1; +#endif +#endif +#endif /*}}*/ +#ifdef Honor_FLT_ROUNDS /*{*/ + int Rounding; +#ifdef Trust_FLT_ROUNDS /*{{ only define this if FLT_ROUNDS really works! */ + Rounding = Flt_Rounds; +#else /*}{*/ + Rounding = 1; + switch(fegetround()) { + case FE_TOWARDZERO: Rounding = 0; break; + case FE_UPWARD: Rounding = 2; break; + case FE_DOWNWARD: Rounding = 3; + } +#endif /*}}*/ +#endif /*}*/ + + u.d = dd; + if (word0(&u) & Sign_bit) { + /* set sign for everything, including 0's and NaNs */ + *sign = 1; + word0(&u) &= ~Sign_bit; /* clear sign bit */ + } + else + *sign = 0; + +#if defined(IEEE_Arith) + defined(VAX) +#ifdef IEEE_Arith + if ((word0(&u) & Exp_mask) == Exp_mask) +#else + if (word0(&u) == 0x8000) +#endif + { + /* Infinity or NaN */ + *decpt = 9999; +#ifdef IEEE_Arith + if (!word1(&u) && !(word0(&u) & 0xfffff)) + return nrv_alloc("Infinity", buf, blen, rve, 8 MTb); +#endif + return nrv_alloc("NaN", buf, blen, rve, 3 MTb); + } +#endif +#ifdef IBM + dval(&u) += 0; /* normalize */ +#endif + if (!dval(&u)) { + *decpt = 1; + return nrv_alloc("0", buf, blen, rve, 1 MTb); + } + +#ifdef SET_INEXACT +#ifndef USE_BF96 + try_quick = +#endif + oldinexact = get_inexact(); + inexact = 1; +#endif +#ifdef Honor_FLT_ROUNDS + if (Rounding >= 2) { + if (*sign) + Rounding = Rounding == 2 ? 0 : 2; + else + if (Rounding != 2) + Rounding = 0; + } +#endif +#ifdef USE_BF96 /*{{*/ + dbits = (u.LL & 0xfffffffffffffull) << 11; /* fraction bits */ + if ((be = u.LL >> 52)) /* biased exponent; nonzero ==> normal */ { + dbits |= 0x8000000000000000ull; + denorm = ulpadj = 0; + } + else { + denorm = 1; + ulpadj = be + 1; + dbits <<= 1; + if (!(dbits & 0xffffffff00000000ull)) { + dbits <<= 32; + be -= 32; + } + if (!(dbits & 0xffff000000000000ull)) { + dbits <<= 16; + be -= 16; + } + if (!(dbits & 0xff00000000000000ull)) { + dbits <<= 8; + be -= 8; + } + if (!(dbits & 0xf000000000000000ull)) { + dbits <<= 4; + be -= 4; + } + if (!(dbits & 0xc000000000000000ull)) { + dbits <<= 2; + be -= 2; + } + if (!(dbits & 0x8000000000000000ull)) { + dbits <<= 1; + be -= 1; + } + assert(be >= -51); + ulpadj -= be; + } + j = Lhint[be + 51]; + p10 = &pten[j]; + dbhi = dbits >> 32; + dblo = dbits & 0xffffffffull; + i = be - 0x3fe; + if (i < p10->e + || (i == p10->e && (dbhi < p10->b0 || (dbhi == p10->b0 && dblo < p10->b1)))) + --j; + k = j - 342; + + /* now 10^k <= dd < 10^(k+1) */ + +#else /*}{*/ + + b = d2b(&u, &be, &bbits MTb); +#ifdef Sudden_Underflow + i = (int)(word0(&u) >> Exp_shift1 & (Exp_mask>>Exp_shift1)); +#else + if ((i = (int)(word0(&u) >> Exp_shift1 & (Exp_mask>>Exp_shift1)))) { +#endif + dval(&d2) = dval(&u); + word0(&d2) &= Frac_mask1; + word0(&d2) |= Exp_11; +#ifdef IBM + if (j = 11 - hi0bits(word0(&d2) & Frac_mask)) + dval(&d2) /= 1 << j; +#endif + + /* log(x) ~=~ log(1.5) + (x-1.5)/1.5 + * log10(x) = log(x) / log(10) + * ~=~ log(1.5)/log(10) + (x-1.5)/(1.5*log(10)) + * log10(d) = (i-Bias)*log(2)/log(10) + log10(d2) + * + * This suggests computing an approximation k to log10(d) by + * + * k = (i - Bias)*0.301029995663981 + * + ( (d2-1.5)*0.289529654602168 + 0.176091259055681 ); + * + * We want k to be too large rather than too small. + * The error in the first-order Taylor series approximation + * is in our favor, so we just round up the constant enough + * to compensate for any error in the multiplication of + * (i - Bias) by 0.301029995663981; since |i - Bias| <= 1077, + * and 1077 * 0.30103 * 2^-52 ~=~ 7.2e-14, + * adding 1e-13 to the constant term more than suffices. + * Hence we adjust the constant term to 0.1760912590558. + * (We could get a more accurate k by invoking log10, + * but this is probably not worthwhile.) + */ + + i -= Bias; +#ifdef IBM + i <<= 2; + i += j; +#endif +#ifndef Sudden_Underflow + denorm = 0; + } + else { + /* d is denormalized */ + + i = bbits + be + (Bias + (P-1) - 1); + x = i > 32 ? word0(&u) << (64 - i) | word1(&u) >> (i - 32) + : word1(&u) << (32 - i); + dval(&d2) = x; + word0(&d2) -= 31*Exp_msk1; /* adjust exponent */ + i -= (Bias + (P-1) - 1) + 1; + denorm = 1; + } +#endif + ds = (dval(&d2)-1.5)*0.289529654602168 + 0.1760912590558 + i*0.301029995663981; + k = (int)ds; + if (ds < 0. && ds != k) + k--; /* want k = floor(ds) */ + k_check = 1; + if (k >= 0 && k <= Ten_pmax) { + if (dval(&u) < tens[k]) + k--; + k_check = 0; + } + j = bbits - i - 1; + if (j >= 0) { + b2 = 0; + s2 = j; + } + else { + b2 = -j; + s2 = 0; + } + if (k >= 0) { + b5 = 0; + s5 = k; + s2 += k; + } + else { + b2 -= k; + b5 = -k; + s5 = 0; + } +#endif /*}}*/ + if (mode < 0 || mode > 9) + mode = 0; + +#ifndef USE_BF96 +#ifndef SET_INEXACT +#ifdef Check_FLT_ROUNDS + try_quick = Rounding == 1; +#endif +#endif /*SET_INEXACT*/ +#endif + + if (mode > 5) { + mode -= 4; +#ifndef USE_BF96 + try_quick = 0; +#endif + } + leftright = 1; + ilim = ilim1 = -1; /* Values for cases 0 and 1; done here to */ + /* silence erroneous "gcc -Wall" warning. */ + switch(mode) { + case 0: + case 1: + i = 18; + ndigits = 0; + break; + case 2: + leftright = 0; + /* no break */ + case 4: + if (ndigits <= 0) + ndigits = 1; + ilim = ilim1 = i = ndigits; + break; + case 3: + leftright = 0; + /* no break */ + case 5: + i = ndigits + k + 1; + ilim = i; + ilim1 = i - 1; + if (i <= 0) + i = 1; + } + if (!buf) { + buf = rv_alloc(i MTb); + blen = sizeof(Bigint) + ((1 << ((int*)buf)[-1]) - 1)*sizeof(ULong) - sizeof(int); + } + else if (blen <= i) { + buf = 0; + if (rve) + *rve = buf + i; + return buf; + } + s = buf; + + /* Check for special case that d is a normalized power of 2. */ + + spec_case = 0; + if (mode < 2 || (leftright +#ifdef Honor_FLT_ROUNDS + && Rounding == 1 +#endif + )) { + if (!word1(&u) && !(word0(&u) & Bndry_mask) +#ifndef Sudden_Underflow + && word0(&u) & (Exp_mask & ~Exp_msk1) +#endif + ) { + /* The special case */ + spec_case = 1; + } + } + +#ifdef USE_BF96 /*{*/ + b = 0; + if (ilim < 0 && (mode == 3 || mode == 5)) { + S = mhi = 0; + goto no_digits; + } + i = 1; + j = 52 + 0x3ff - be; + ulpshift = 0; + ulplo = 0; + /* Can we do an exact computation with 64-bit integer arithmetic? */ + if (k < 0) { + if (k < -25) + goto toobig; + res = dbits >> 11; + n2 = pfivebits[k1 = -(k + 1)] + 53; + j1 = j; + if (n2 > 61) { + ulpshift = n2 - 61; + if (res & (ulpmask = (1ull << ulpshift) - 1)) + goto toobig; + j -= ulpshift; + res >>= ulpshift; + } + /* Yes. */ + res *= ulp = pfive[k1]; + if (ulpshift) { + ulplo = ulp; + ulp >>= ulpshift; + } + j += k; + if (ilim == 0) { + S = mhi = 0; + if (res > (5ull << j)) + goto one_digit; + goto no_digits; + } + goto no_div; + } + if (ilim == 0 && j + k >= 0) { + S = mhi = 0; + if ((dbits >> 11) > (pfive[k-1] << j)) + goto one_digit; + goto no_digits; + } + if (k <= dtoa_divmax && j + k >= 0) { + /* Another "yes" case -- we will use exact integer arithmetic. */ + use_exact: + Debug(++dtoa_stats[3]); + res = dbits >> 11; /* residual */ + ulp = 1; + if (k <= 0) + goto no_div; + j1 = j + k + 1; + den = pfive[k-i] << (j1 - i); + for(;;) { + dig = res / den; + *s++ = '0' + dig; + if (!(res -= dig*den)) { +#ifdef SET_INEXACT + inexact = 0; + oldinexact = 1; +#endif + goto retc; + } + if (ilim < 0) { + ures = den - res; + if (2*res <= ulp + && (spec_case ? 4*res <= ulp : (2*res < ulp || dig & 1))) + goto ulp_reached; + if (2*ures < ulp) + goto Roundup; + } + else if (i == ilim) { + switch(Rounding) { + case 0: goto retc; + case 2: goto Roundup; + } + ures = 2*res; + if (ures > den + || (ures == den && dig & 1) + || (spec_case && res <= ulp && 2*res >= ulp)) + goto Roundup; + goto retc; + } + if (j1 < ++i) { + res *= 10; + ulp *= 10; + } + else { + if (i > k) + break; + den = pfive[k-i] << (j1 - i); + } + } + no_div: + for(;;) { + dig = den = res >> j; + *s++ = '0' + dig; + if (!(res -= den << j)) { +#ifdef SET_INEXACT + inexact = 0; + oldinexact = 1; +#endif + goto retc; + } + if (ilim < 0) { + ures = (1ull << j) - res; + if (2*res <= ulp + && (spec_case ? 4*res <= ulp : (2*res < ulp || dig & 1))) { + ulp_reached: + if (ures < res + || (ures == res && dig & 1)) + goto Roundup; + goto retc; + } + if (2*ures < ulp) + goto Roundup; + } + --j; + if (i == ilim) { +#ifdef Honor_FLT_ROUNDS + switch(Rounding) { + case 0: goto retc; + case 2: goto Roundup; + } +#endif + hb = 1ull << j; + if (res & hb && (dig & 1 || res & (hb-1))) + goto Roundup; + if (spec_case && res <= ulp && 2*res >= ulp) { + Roundup: + while(*--s == '9') + if (s == buf) { + ++k; + *s++ = '1'; + goto ret1; + } + ++*s++; + goto ret1; + } + goto retc; + } + ++i; + res *= 5; + if (ulpshift) { + ulplo = 5*(ulplo & ulpmask); + ulp = 5*ulp + (ulplo >> ulpshift); + } + else + ulp *= 5; + } + } + toobig: + if (ilim > 28) + goto Fast_failed1; + /* Scale by 10^-k */ + p10 = &pten[342-k]; + tv0 = p10->b2 * dblo; /* rarely matters, but does, e.g., for 9.862818194192001e18 */ + tv1 = p10->b1 * dblo + (tv0 >> 32); + tv2 = p10->b2 * dbhi + (tv1 & 0xffffffffull); + tv3 = p10->b0 * dblo + (tv1>>32) + (tv2>>32); + res3 = p10->b1 * dbhi + (tv3 & 0xffffffffull); + res = p10->b0 * dbhi + (tv3>>32) + (res3>>32); + be += p10->e - 0x3fe; + eulp = j1 = be - 54 + ulpadj; + if (!(res & 0x8000000000000000ull)) { + --be; + res3 <<= 1; + res = (res << 1) | ((res3 & 0x100000000ull) >> 32); + } + res0 = res; /* save for Fast_failed */ +#if !defined(SET_INEXACT) && !defined(NO_DTOA_64) /*{*/ + if (ilim > 19) + goto Fast_failed; + Debug(++dtoa_stats[4]); + assert(be >= 0 && be <= 4); /* be = 0 is rare, but possible, e.g., for 1e20 */ + res >>= 4 - be; + ulp = p10->b0; /* ulp */ + ulp = (ulp << 29) | (p10->b1 >> 3); + /* scaled ulp = ulp * 2^(eulp - 60) */ + /* We maintain 61 bits of the scaled ulp. */ + if (ilim == 0) { + if (!(res & 0x7fffffffffffffeull) + || !((~res) & 0x7fffffffffffffeull)) + goto Fast_failed1; + S = mhi = 0; + if (res >= 0x5000000000000000ull) + goto one_digit; + goto no_digits; + } + rb = 1; /* upper bound on rounding error */ + for(;;++i) { + dig = res >> 60; + *s++ = '0' + dig; + res &= 0xfffffffffffffffull; + if (ilim < 0) { + ures = 0x1000000000000000ull - res; + if (eulp > 0) { + assert(eulp <= 4); + sulp = ulp << (eulp - 1); + if (res <= ures) { + if (res + rb > ures - rb) + goto Fast_failed; + if (res < sulp) + goto retc; + } + else { + if (res - rb <= ures + rb) + goto Fast_failed; + if (ures < sulp) + goto Roundup; + } + } + else { + zb = -(1ull << (eulp + 63)); + if (!(zb & res)) { + sres = res << (1 - eulp); + if (sres < ulp && (!spec_case || 2*sres < ulp)) { + if ((res+rb) << (1 - eulp) >= ulp) + goto Fast_failed; + if (ures < res) { + if (ures + rb >= res - rb) + goto Fast_failed; + goto Roundup; + } + if (ures - rb < res + rb) + goto Fast_failed; + goto retc; + } + } + if (!(zb & ures) && ures << -eulp < ulp) { + if (ures << (1 - eulp) < ulp) + goto Roundup; + goto Fast_failed; + } + } + } + else if (i == ilim) { + ures = 0x1000000000000000ull - res; + if (ures < res) { + if (ures <= rb || res - rb <= ures + rb) { + if (j + k >= 0 && k >= 0 && k <= 27) + goto use_exact1; + goto Fast_failed; + } +#ifdef Honor_FLT_ROUNDS + if (Rounding == 0) + goto retc; +#endif + goto Roundup; + } + if (res <= rb || ures - rb <= res + rb) { + if (j + k >= 0 && k >= 0 && k <= 27) { + use_exact1: + s = buf; + i = 1; + goto use_exact; + } + goto Fast_failed; + } +#ifdef Honor_FLT_ROUNDS + if (Rounding == 2) + goto Roundup; +#endif + goto retc; + } + rb *= 10; + if (rb >= 0x1000000000000000ull) + goto Fast_failed; + res *= 10; + ulp *= 5; + if (ulp & 0x8000000000000000ull) { + eulp += 4; + ulp >>= 3; + } + else { + eulp += 3; + ulp >>= 2; + } + } +#endif /*}*/ +#ifndef NO_BF96 + Fast_failed: +#endif + Debug(++dtoa_stats[5]); + s = buf; + i = 4 - be; + res = res0 >> i; + reslo = 0xffffffffull & res3; + if (i) + reslo = (res0 << (64 - i)) >> 32 | (reslo >> i); + rb = 0; + rblo = 4; /* roundoff bound */ + ulp = p10->b0; /* ulp */ + ulp = (ulp << 29) | (p10->b1 >> 3); + eulp = j1; + for(i = 1;;++i) { + dig = res >> 60; + *s++ = '0' + dig; + res &= 0xfffffffffffffffull; +#ifdef SET_INEXACT + if (!res && !reslo) { + if (!(res3 & 0xffffffffull)) { + inexact = 0; + oldinexact = 1; + } + goto retc; + } +#endif + if (ilim < 0) { + ures = 0x1000000000000000ull - res; + ureslo = 0; + if (reslo) { + ureslo = 0x100000000ull - reslo; + --ures; + } + if (eulp > 0) { + assert(eulp <= 4); + sulp = (ulp << (eulp - 1)) - rb; + if (res <= ures) { + if (res < sulp) { + if (res+rb < ures-rb) + goto retc; + } + } + else if (ures < sulp) { + if (res-rb > ures+rb) + goto Roundup; + } + goto Fast_failed1; + } + else { + zb = -(1ull << (eulp + 60)); + if (!(zb & (res + rb))) { + sres = (res - rb) << (1 - eulp); + if (sres < ulp && (!spec_case || 2*sres < ulp)) { + sres = res << (1 - eulp); + if ((j = eulp + 31) > 0) + sres += (rblo + reslo) >> j; + else + sres += (rblo + reslo) << -j; + if (sres + (rb << (1 - eulp)) >= ulp) + goto Fast_failed1; + if (sres >= ulp) + goto more96; + if (ures < res + || (ures == res && ureslo < reslo)) { + if (ures + rb >= res - rb) + goto Fast_failed1; + goto Roundup; + } + if (ures - rb <= res + rb) + goto Fast_failed1; + goto retc; + } + } + if (!(zb & ures) && (ures-rb) << (1 - eulp) < ulp) { + if ((ures + rb) << (1 - eulp) < ulp) + goto Roundup; + goto Fast_failed1; + } + } + } + else if (i == ilim) { + ures = 0x1000000000000000ull - res; + sres = ureslo = 0; + if (reslo) { + ureslo = 0x100000000ull - reslo; + --ures; + sres = (reslo + rblo) >> 31; + } + sres += 2*rb; + if (ures <= res) { + if (ures <=sres || res - ures <= sres) + goto Fast_failed1; +#ifdef Honor_FLT_ROUNDS + if (Rounding == 0) + goto retc; +#endif + goto Roundup; + } + if (res <= sres || ures - res <= sres) + goto Fast_failed1; +#ifdef Honor_FLT_ROUNDS + if (Rounding == 2) + goto Roundup; +#endif + goto retc; + } + more96: + rblo *= 10; + rb = 10*rb + (rblo >> 32); + rblo &= 0xffffffffull; + if (rb >= 0x1000000000000000ull) + goto Fast_failed1; + reslo *= 10; + res = 10*res + (reslo >> 32); + reslo &= 0xffffffffull; + ulp *= 5; + if (ulp & 0x8000000000000000ull) { + eulp += 4; + ulp >>= 3; + } + else { + eulp += 3; + ulp >>= 2; + } + } + Fast_failed1: + Debug(++dtoa_stats[6]); + S = mhi = mlo = 0; +#ifdef USE_BF96 + b = d2b(&u, &be, &bbits MTb); +#endif + s = buf; + i = (int)(word0(&u) >> Exp_shift1 & (Exp_mask>>Exp_shift1)); + i -= Bias; + if (ulpadj) + i -= ulpadj - 1; + j = bbits - i - 1; + if (j >= 0) { + b2 = 0; + s2 = j; + } + else { + b2 = -j; + s2 = 0; + } + if (k >= 0) { + b5 = 0; + s5 = k; + s2 += k; + } + else { + b2 -= k; + b5 = -k; + s5 = 0; + } +#endif /*}*/ + +#ifdef Honor_FLT_ROUNDS + if (mode > 1 && Rounding != 1) + leftright = 0; +#endif + +#ifndef USE_BF96 /*{*/ + if (ilim >= 0 && ilim <= Quick_max && try_quick) { + + /* Try to get by with floating-point arithmetic. */ + + i = 0; + dval(&d2) = dval(&u); + j1 = -(k0 = k); + ilim0 = ilim; + ieps = 2; /* conservative */ + if (k > 0) { + ds = tens[k&0xf]; + j = k >> 4; + if (j & Bletch) { + /* prevent overflows */ + j &= Bletch - 1; + dval(&u) /= bigtens[n_bigtens-1]; + ieps++; + } + for(; j; j >>= 1, i++) + if (j & 1) { + ieps++; + ds *= bigtens[i]; + } + dval(&u) /= ds; + } + else if (j1 > 0) { + dval(&u) *= tens[j1 & 0xf]; + for(j = j1 >> 4; j; j >>= 1, i++) + if (j & 1) { + ieps++; + dval(&u) *= bigtens[i]; + } + } + if (k_check && dval(&u) < 1. && ilim > 0) { + if (ilim1 <= 0) + goto fast_failed; + ilim = ilim1; + k--; + dval(&u) *= 10.; + ieps++; + } + dval(&eps) = ieps*dval(&u) + 7.; + word0(&eps) -= (P-1)*Exp_msk1; + if (ilim == 0) { + S = mhi = 0; + dval(&u) -= 5.; + if (dval(&u) > dval(&eps)) + goto one_digit; + if (dval(&u) < -dval(&eps)) + goto no_digits; + goto fast_failed; + } +#ifndef No_leftright + if (leftright) { + /* Use Steele & White method of only + * generating digits needed. + */ + dval(&eps) = 0.5/tens[ilim-1] - dval(&eps); +#ifdef IEEE_Arith + if (j1 >= 307) { + eps1.d = 1.01e256; /* 1.01 allows roundoff in the next few lines */ + word0(&eps1) -= Exp_msk1 * (Bias+P-1); + dval(&eps1) *= tens[j1 & 0xf]; + for(i = 0, j = (j1-256) >> 4; j; j >>= 1, i++) + if (j & 1) + dval(&eps1) *= bigtens[i]; + if (eps.d < eps1.d) + eps.d = eps1.d; + if (10. - u.d < 10.*eps.d && eps.d < 1.) { + /* eps.d < 1. excludes trouble with the tiniest denormal */ + *s++ = '1'; + ++k; + goto ret1; + } + } +#endif + for(i = 0;;) { + L = dval(&u); + dval(&u) -= L; + *s++ = '0' + (int)L; + if (1. - dval(&u) < dval(&eps)) + goto bump_up; + if (dval(&u) < dval(&eps)) + goto retc; + if (++i >= ilim) + break; + dval(&eps) *= 10.; + dval(&u) *= 10.; + } + } + else { +#endif + /* Generate ilim digits, then fix them up. */ + dval(&eps) *= tens[ilim-1]; + for(i = 1;; i++, dval(&u) *= 10.) { + L = (Long)(dval(&u)); + if (!(dval(&u) -= L)) + ilim = i; + *s++ = '0' + (int)L; + if (i == ilim) { + if (dval(&u) > 0.5 + dval(&eps)) + goto bump_up; + else if (dval(&u) < 0.5 - dval(&eps)) + goto retc; + break; + } + } +#ifndef No_leftright + } +#endif + fast_failed: + s = buf; + dval(&u) = dval(&d2); + k = k0; + ilim = ilim0; + } + + /* Do we have a "small" integer? */ + + if (be >= 0 && k <= Int_max) { + /* Yes. */ + ds = tens[k]; + if (ndigits < 0 && ilim <= 0) { + S = mhi = 0; + if (ilim < 0 || dval(&u) <= 5*ds) + goto no_digits; + goto one_digit; + } + for(i = 1;; i++, dval(&u) *= 10.) { + L = (Long)(dval(&u) / ds); + dval(&u) -= L*ds; +#ifdef Check_FLT_ROUNDS + /* If FLT_ROUNDS == 2, L will usually be high by 1 */ + if (dval(&u) < 0) { + L--; + dval(&u) += ds; + } +#endif + *s++ = '0' + (int)L; + if (!dval(&u)) { +#ifdef SET_INEXACT + inexact = 0; +#endif + break; + } + if (i == ilim) { +#ifdef Honor_FLT_ROUNDS + if (mode > 1) + switch(Rounding) { + case 0: goto retc; + case 2: goto bump_up; + } +#endif + dval(&u) += dval(&u); +#ifdef ROUND_BIASED + if (dval(&u) >= ds) +#else + if (dval(&u) > ds || (dval(&u) == ds && L & 1)) +#endif + { + bump_up: + while(*--s == '9') + if (s == buf) { + k++; + *s = '0'; + break; + } + ++*s++; + } + break; + } + } + goto retc; + } + +#endif /*}*/ + m2 = b2; + m5 = b5; + mhi = mlo = 0; + if (leftright) { + i = +#ifndef Sudden_Underflow + denorm ? be + (Bias + (P-1) - 1 + 1) : +#endif +#ifdef IBM + 1 + 4*P - 3 - bbits + ((bbits + be - 1) & 3); +#else + 1 + P - bbits; +#endif + b2 += i; + s2 += i; + mhi = i2b(1 MTb); + } + if (m2 > 0 && s2 > 0) { + i = m2 < s2 ? m2 : s2; + b2 -= i; + m2 -= i; + s2 -= i; + } + if (b5 > 0) { + if (leftright) { + if (m5 > 0) { + mhi = pow5mult(mhi, m5 MTb); + b1 = mult(mhi, b MTb); + Bfree(b MTb); + b = b1; + } + if ((j = b5 - m5)) + b = pow5mult(b, j MTb); + } + else + b = pow5mult(b, b5 MTb); + } + S = i2b(1 MTb); + if (s5 > 0) + S = pow5mult(S, s5 MTb); + + if (spec_case) { + b2 += Log2P; + s2 += Log2P; + } + + /* Arrange for convenient computation of quotients: + * shift left if necessary so divisor has 4 leading 0 bits. + * + * Perhaps we should just compute leading 28 bits of S once + * and for all and pass them and a shift to quorem, so it + * can do shifts and ors to compute the numerator for q. + */ + i = dshift(S, s2); + b2 += i; + m2 += i; + s2 += i; + if (b2 > 0) + b = lshift(b, b2 MTb); + if (s2 > 0) + S = lshift(S, s2 MTb); +#ifndef USE_BF96 + if (k_check) { + if (cmp(b,S) < 0) { + k--; + b = multadd(b, 10, 0 MTb); /* we botched the k estimate */ + if (leftright) + mhi = multadd(mhi, 10, 0 MTb); + ilim = ilim1; + } + } +#endif + if (ilim <= 0 && (mode == 3 || mode == 5)) { + if (ilim < 0 || cmp(b,S = multadd(S,5,0 MTb)) <= 0) { + /* no digits, fcvt style */ + no_digits: + k = -1 - ndigits; + goto ret; + } + one_digit: + *s++ = '1'; + ++k; + goto ret; + } + if (leftright) { + if (m2 > 0) + mhi = lshift(mhi, m2 MTb); + + /* Compute mlo -- check for special case + * that d is a normalized power of 2. + */ + + mlo = mhi; + if (spec_case) { + mhi = Balloc(mhi->k MTb); + Bcopy(mhi, mlo); + mhi = lshift(mhi, Log2P MTb); + } + + for(i = 1;;i++) { + dig = quorem(b,S) + '0'; + /* Do we yet have the shortest decimal string + * that will round to d? + */ + j = cmp(b, mlo); + delta = diff(S, mhi MTb); + j1 = delta->sign ? 1 : cmp(b, delta); + Bfree(delta MTb); +#ifndef ROUND_BIASED + if (j1 == 0 && mode != 1 && !(word1(&u) & 1) +#ifdef Honor_FLT_ROUNDS + && (mode <= 1 || Rounding >= 1) +#endif + ) { + if (dig == '9') + goto round_9_up; + if (j > 0) + dig++; +#ifdef SET_INEXACT + else if (!b->x[0] && b->wds <= 1) + inexact = 0; +#endif + *s++ = dig; + goto ret; + } +#endif + if (j < 0 || (j == 0 && mode != 1 +#ifndef ROUND_BIASED + && !(word1(&u) & 1) +#endif + )) { + if (!b->x[0] && b->wds <= 1) { +#ifdef SET_INEXACT + inexact = 0; +#endif + goto accept_dig; + } +#ifdef Honor_FLT_ROUNDS + if (mode > 1) + switch(Rounding) { + case 0: goto accept_dig; + case 2: goto keep_dig; + } +#endif /*Honor_FLT_ROUNDS*/ + if (j1 > 0) { + b = lshift(b, 1 MTb); + j1 = cmp(b, S); +#ifdef ROUND_BIASED + if (j1 >= 0 /*)*/ +#else + if ((j1 > 0 || (j1 == 0 && dig & 1)) +#endif + && dig++ == '9') + goto round_9_up; + } + accept_dig: + *s++ = dig; + goto ret; + } + if (j1 > 0) { +#ifdef Honor_FLT_ROUNDS + if (!Rounding && mode > 1) + goto accept_dig; +#endif + if (dig == '9') { /* possible if i == 1 */ + round_9_up: + *s++ = '9'; + goto roundoff; + } + *s++ = dig + 1; + goto ret; + } +#ifdef Honor_FLT_ROUNDS + keep_dig: +#endif + *s++ = dig; + if (i == ilim) + break; + b = multadd(b, 10, 0 MTb); + if (mlo == mhi) + mlo = mhi = multadd(mhi, 10, 0 MTb); + else { + mlo = multadd(mlo, 10, 0 MTb); + mhi = multadd(mhi, 10, 0 MTb); + } + } + } + else + for(i = 1;; i++) { + dig = quorem(b,S) + '0'; + *s++ = dig; + if (!b->x[0] && b->wds <= 1) { +#ifdef SET_INEXACT + inexact = 0; +#endif + goto ret; + } + if (i >= ilim) + break; + b = multadd(b, 10, 0 MTb); + } + + /* Round off last digit */ + +#ifdef Honor_FLT_ROUNDS + if (mode > 1) + switch(Rounding) { + case 0: goto ret; + case 2: goto roundoff; + } +#endif + b = lshift(b, 1 MTb); + j = cmp(b, S); +#ifdef ROUND_BIASED + if (j >= 0) +#else + if (j > 0 || (j == 0 && dig & 1)) +#endif + { + roundoff: + while(*--s == '9') + if (s == buf) { + k++; + *s++ = '1'; + goto ret; + } + ++*s++; + } + ret: + Bfree(S MTb); + if (mhi) { + if (mlo && mlo != mhi) + Bfree(mlo MTb); + Bfree(mhi MTb); + } + retc: + while(s > buf && s[-1] == '0') + --s; + ret1: + if (b) + Bfree(b MTb); + *s = 0; + *decpt = k + 1; + if (rve) + *rve = s; +#ifdef SET_INEXACT + if (inexact) { + if (!oldinexact) { + word0(&u) = Exp_1 + (70 << Exp_shift); + word1(&u) = 0; + dval(&u) += 1.; + } + } + else if (!oldinexact) + clear_inexact(); +#endif + return buf; + } + +static char * +dtoa(double dd, int mode, int ndigits, int *decpt, int *sign, char **rve) +{ + /* Sufficient space is allocated to the return value + to hold the suppressed trailing zeros. + See dtoa_r() above for details on the other arguments. + */ +#ifndef MULTIPLE_THREADS + if (dtoa_result) + freedtoa(dtoa_result); +#endif + return dtoa_r_dmgay(dd, mode, ndigits, decpt, sign, rve, 0, 0); + } + +#ifdef __cplusplus +} +#endif + diff --git a/nl-writer2/src/nl-utils2.cc b/nl-writer2/src/nl-utils2.cc new file mode 100644 index 000000000..488e506f5 --- /dev/null +++ b/nl-writer2/src/nl-utils2.cc @@ -0,0 +1,47 @@ +#include "mp/nl-utils2.h" + +/////////////////////// TECHNICAL ///////////////////////// +namespace mp { + +void File::Printf(const char *format, ...) { + va_list args; + va_start (args, format); + std::vfprintf (f_, format, args); + va_end (args); +} + +File NLUtils:: +openf(const std::string& fname, int Close, const char *mode) +{ + File f; + if (Close) { + std::remove(fname.c_str()); + return f; + } + f.Open(fname.c_str(), mode); + if (!f) { + log_warning("can't open %s", fname.c_str()); + return f; + } + if (if_show_filenames()) + log_message("File %s\n", fname.c_str()); + return f; +} + +void NLUtils::log_message(const char *format, ...) { + va_list args; + va_start (args, format); + std::vprintf (format, args); + va_end (args); +} + +void NLUtils::log_warning(const char *format, ...) { + std::fprintf(stderr, "WARNING: "); + va_list args; + va_start (args, format); + std::vfprintf (stderr, format, args); + va_end (args); + std::fprintf(stderr, "\n"); +} + +} // namespace mp diff --git a/nl-writer2/src/nl-writer2.cc b/nl-writer2/src/nl-writer2.cc new file mode 100644 index 000000000..ebe2e7b7f --- /dev/null +++ b/nl-writer2/src/nl-writer2.cc @@ -0,0 +1,459 @@ +/** + NL writer + + NL is a format for representing optimization problems such as linear, + quadratic, nonlinear, complementarity and constraint programming problems + in discrete or continuous variables. It is described in the technical report + "Writing .nl Files" (http://www.cs.sandia.gov/~dmgay/nlwrite.pdf). + + This is a complete reusable C++ implementation of an NL writer. + + Usage: + /// Write an NL file: + WriteNLFile(filename, feeder); + + /// Write an NL string: + WriteNLString(string, feeder); + + where feeder is an object that provides information on model + components. See NLFeeder for an example of a feeder class. + + See also NLReader and NLHandler classes. + + Copyright (C) 2023 AMPL Optimization Inc. + + Permission to use, copy, modify, and distribute this software and its + documentation for any purpose and without fee is hereby granted, + provided that the above copyright notice appear in all copies and that + both that the copyright notice and this permission notice and warranty + disclaimer appear in supporting documentation. + + The author and AMPL Optimization Inc disclaim all warranties with + regard to this software, including all implied warranties of + merchantability and fitness. In no event shall the author be liable + for any special, indirect or consequential damages or any damages + whatsoever resulting from loss of use, data or profits, whether in an + action of contract, negligence or other tortious action, arising out + of or in connection with the use or performance of this software. + + Author: Gleb Belov + */ + +#include +#include +#include +#include + +#include "mp/nl-writer2.h" +#include "mp/nl-writer2.hpp" + +#ifndef ssize_t // signed size_t (POSIX) + #define ssize_t int +#endif + +/// Check that the stub NLFeeder2 +/// compiles with NLWriter2 +static void test_nlfeeder2_with_nlwriter2() { + using Expr = void*; + class TestNLF : public mp::NLFeeder2 { + public: + }; + TestNLF nlfeeder; + mp::NLUtils utils; + mp::WriteNLFile("test.nl", nlfeeder, utils); +} + + +namespace DAVID_GAY_GFMT { +char* gfmt(double x, int prec); +} + +namespace NL_LIB_GFMT { +void gfmt(char*, size_t, double, int); +} + + +//////////////////////////////////////////////////////// +////////////////////// NLW2 UTILS ////////////////////// +//////////////////////////////////////////////////////// +namespace mp { + +/// \brief For printing to .nl or auxiliary files. +int TextFormatter::apr(File& f, const char *fmt, ...) +{ + auto fd = f.GetHandle(); + char buf[32], *s; + double x; + int rc; + ssize_t i, j; + va_list ap; + + rc = 0; + + va_start(ap, fmt); + if (*fmt != '%') + rc++; + for(;;) { + for(;;) { + switch(i = *fmt++) { + case 0: goto done; + case '%': break; + case '\t': + if (!nl_comments) { + putc('\n', fd); + va_end(ap); + return rc; + } + /* no break */ + default: + putc(i, fd); + continue; + } + break; + } + rc++; + switch(*fmt++) { + case 'c': + i = va_arg(ap, int); + putc(i, fd); + continue; + case 'z': + i = va_arg(ap, size_t); + goto have_i; + case 'd': + i = va_arg(ap, int); +have_i: + if (i < 0) { + putc('-',fd); + i = -i; + } + s = buf; + do { + j = i / 10; + *s++ = i - 10*j + '0'; + } + while((i = j)); + do { + i = *--s; + putc(i,fd); + } while (s > buf); + continue; + case '.': + while (*fmt++ != 'g'); + case 'g': + x = va_arg(ap, double); +#ifdef NL_LIB_USE_SPRINTF + snprintf(s = buf, + sizeof(buf), "%.*g", output_prec, x); +#elif NL_LIB_USE_OWN_GFMT + NL_LIB_GFMT::gfmt(s = buf, sizeof(buf), x, output_prec); +#else + s = DAVID_GAY_GFMT::gfmt(x, output_prec); +#endif + goto have_s; + case 's': + s = va_arg(ap, char*); +have_s: + while ((i = *s++)) + putc(i,fd); + continue; + default: + Utils().myexit("aprintf bug: unexpected fmt: " + + std::string(fmt-1)); + } + } +done: + va_end(ap); + return rc; +} + +void TextFormatter::nput(File& nm, double r) +{ + apr(nm, "n%g\n", r); + /* formerly fprintf(nm, "n%.16g\n", *r); */ +} + +void BinaryFormatter::nput(File& nm, double r) +{ + double x; + long L; + short sh; + + if ((x = r) <= 2147483647. + && x >= -2147483648. + && (L = (long)x, (double)x == L)) { + sh = (short)L; + if (sh == L) + apr(nm, "s%h", sh); + else + apr(nm, "l%l", L); + } + else + apr(nm, "n%g", x); +} + +namespace { +typedef int Long; +typedef Long Int; +/// For printing to .nl/.mps. used in #outmps. +typedef union U { double x; short sh; Long L; Int i; char c; size_t z; Long LL[2]; + #ifndef NO_LONG_LONG + long long ll; + #endif + } U; +} + +int BinaryFormatter:: +apr(File& f, const char *fmt, ...) +{ + auto fd = f.GetHandle(); + U u; + char *s; + int i, rc; + size_t len, x; + va_list ap; + + rc = 0; + va_start(ap, fmt); + + if ((i = *fmt) != '%') { + fmt++; +#define DMG +#ifdef DMG + if (i != 'i') +#endif + { + u.c = i; + fwrite(&u.c, 1, 1, fd); + rc++; + }} + + for(;;) { + while(*fmt == ' ') + fmt++; + if (*fmt++ != '%') + break; + switch(*fmt++) { + case 'c': + u.c = va_arg(ap, int); + len = 1; + break; + case 'd': + u.i = va_arg(ap, int); + len = sizeof(Int); + break; + case '.': + while(*fmt++ != 'g'); + case 'g': + u.x = va_arg(ap, double); + len = sizeof(double); + break; + case 'h': + u.sh = va_arg(ap, int); + len = sizeof(short); + if (*fmt == 'd') + fmt++; + break; + case 'l': + u.L = (Long)va_arg(ap, long); + len = sizeof(Long); + if (*fmt == 'd') + fmt++; + break; + case 's': + s = va_arg(ap, char*); + u.i = std::strlen(s); + fwrite((char *)&u.i, sizeof(Int), 1, fd); + fwrite(s, u.i, 1, fd); + goto s_written; + case 'z': + x = va_arg(ap, size_t); + u.L = (Long)x; + len = sizeof(Long); + break; + default: + Utils().myexit("bprintf bug: unexpected fmt: " + + std::string(fmt-1)); +#ifdef AVOID_BOGUS_WARNINGS + len = 0; +#endif + } + fwrite((char *)&u.L, len, 1, fd); +s_written: + rc++; + if (*fmt == ':') { + /* special Hollerith */ + s = va_arg(ap, char*); + fwrite(s, u.i, 1, fd); + fmt += 3; + rc++; + } + } + va_end(ap); + return rc; +} + +} // namespace mp + + +extern "C" { +char * +dtoa_r_dmgay(double dd, int mode, int ndigits, + int *decpt, int *sign, char **rve, + char *buf, size_t blen); +} + +namespace DAVID_GAY_GFMT { + +typedef int Long; + +/// Adjusts precision of floating point numbers. +int +g_fmt(char *b, double x, int prec) +{ + int i, k; + char *s; + int decpt, j, sign; + char *b0, sbuf[400], *se; + + b0 = b; +#if defined(_WIN32) || defined(_WIN64) +#ifdef IEEE_8087 + if (!(((Long*)&x)[1] & 0x7fffffff | ((Long*)&x)[0])) { +#else + if (!(((Long*)&x)[0] & 0x7fffffff | ((Long*)&x)[1])) { +#endif /* IEEE_8087 */ +#else /* _WIN32 || _WIN64 */ + if (! x) { +#endif /*_WIN32 || _WIN64 */ + *b++ = '0'; + *b = 0; + goto done; + } + s = dtoa_r_dmgay(x, prec ? 2 : 0, prec, + &decpt, &sign, &se, sbuf, sizeof(sbuf)); + if (sign) + *b++ = '-'; + if (decpt == 9999) /* Infinity or Nan */ { + if (*s == 'N') /* NaN */ + b = b0; + while((*b = *s++)) + b++; + goto done; + } + if (decpt <= -4 || decpt > se - s + (s[1] ? 5 : 4)) { + *b++ = *s++; + if (*s) { + *b++ = '.'; + while((*b = *s++)) + b++; + } + *b++ = 'e'; + /* sprintf(b, "%+.2d", decpt - 1); */ + if (--decpt < 0) { + *b++ = '-'; + decpt = -decpt; + } + else + *b++ = '+'; + for(j = 2, k = 10; 10*k <= decpt; j++, k *= 10); + for(;;) { + i = decpt / k; + *b++ = i + '0'; + if (--j <= 0) + break; + decpt -= i*k; + decpt *= 10; + } + *b = 0; + } + else if (decpt <= 0) { + *b++ = '0'; + *b++ = '.'; + for(; decpt < 0; decpt++) + *b++ = '0'; + while((*b = *s++)) + b++; + } + else { + while((*b = *s++)) { + b++; + if (--decpt == 0 && *s) + *b++ = '.'; + } + for(; decpt > 0; decpt--) + *b++ = '0'; + *b = 0; + } +done: + return b - b0; +} + +char Bf[400]; + +char * +gfmt(double x, int prec) +{ + g_fmt(Bf, x, prec); + return Bf; +} + +} // namespace DAVID_GAY_GFMT + + +namespace NL_LIB_GFMT { + +/// Adjusts precision of floating point numbers. +/// If too slow, should be replaced by fmt::format +/// or std::to_chars (C++ 17). +void gfmt(char *b, size_t sz, double x, int prec) { + switch (std::fpclassify(x)) + { + case FP_INFINITE: + if (x>=0) + std::strncpy(b, "Infinity", sz); + else + std::strncpy(b, "-Infinity", sz); + return; + case FP_NAN: + std::strncpy(b, "NaN", sz); + return; + // case FP_NORMAL: + // return "normal"; + // case FP_SUBNORMAL: + // return "subnormal"; + case FP_ZERO: + std::strncpy(b, "0", sz); + return; + default: + if (x == std::floor(x)) { // integer + auto x_by_1e5 = x / 1e5; + if (x_by_1e5 != std::round(x_by_1e5)) { + snprintf(b, sz, "%.*g", prec, x); + return; + } + } + auto l10 = std::log10(std::fabs(x)); + auto p0 = int(-std::floor(l10)); + assert(std::fabs(x * pow(10.0, p0)) > 0.9999); + assert(std::fabs(x * pow(10.0, p0)) < 10.0001); + --p0; + int i=1; // we already have 1 digit + for (; i=prec) + break; + auto x1 = std::stod(b); + if (x==x1) + return; + ++i; + } while (i<=prec); + } +} + +} // namespace NL_LIB_GFMT