table parser does not parse columns with only name and type. this leads to data loss over these column types during the copy phase/

[purzelrakete]

Lhm::Table::Parser is used to parse the strucutre of origin and destination tables in a migration. The resulting intersection determines which columns are copied in the chunking phase.

Lhm::Table::Parser currently parses the output of show create table. An error in the regular expression for columns causes the following table definition to parse incorrectly:

CREATE TABLE users (
id int(11) NOT NULL AUTO_INCREMENT,
reference int(11) DEFAULT NULL,
username varchar(255) DEFAULT NULL,
group varchar(255) DEFAULT NULL,
created_at datetime DEFAULT NULL,
comment varchar(20) DEFAULT NULL,
description text,
PRIMARY KEY (id),
UNIQUE KEY index_users_on_reference (reference),
KEY index_users_on_username_and_created_at (username,created_at)
) ENGINE=InnoDB DEFAULT CHARSET=utf8

Parsing this ddl will fail to add the description column to the parsed table.

A more robust Table parser using information_schema and no regular expressions fixes this problem.

[purzelrakete]

fixed in 1.0.2