Unexpected CRS inequality

[djhoese]

Example of problem

#include <proj.h>
#include <stdio.h>

int main(void) {
    PJ_CONTEXT* context = proj_context_create();
    PJ* pj1 = proj_create(
        context,
        "+proj=geos +lon_0=0 +h=35786400 +a=6378137.0 +rf=298.257223563 +units=m +sweep=y +no_defs +type=crs"
    );
    PJ* pj2 = proj_create(
        context,
        "+proj=geos +lon_0=0 +h=35786400 +ellps=WGS84 +units=m +sweep=y +no_defs +type=crs"
    );
    printf("%d\n", proj_is_equivalent_to_with_ctx(
        context, pj1, pj2, PJ_COMP_EQUIVALENT
    ));
    return 0;
}

Problem description

I'm using pyproj and end up comparing two CRS objects. One is created from a and rf parameters taken from a satellite instrument data file (pj1 above). The other is taken from a configuration for the expected/documented/ideal version of this CRS (pj2 above). The instrument/data file is documented as being on an WGS84 ellipsoid (my understanding at least). I expected that these two sets of parameters would be equivalent. I haven't had the time yet to determine if they were in older versions of PROJ.

pyproj4/pyproj#1284

Expected Output

1 meaning the CRSes are equivalent.

Environment Information

• PROJ version (proj): Rel. 9.1.1, December 1st, 2022
• Operation System Information: PopOS/Ubuntu

Installation method

• conda via conda-forge package

[djhoese]

For the record, this seems to work as expected in PROJ 9.0.x but not 9.1.x.