~drizzle-trunk/drizzle/development : changes

~drizzle-trunk/drizzle/development » Changes from revision 896.4.9

Rev	Summary	Authors	Date
896.4.9	No longer write the FRM. just use proto. No longer write the FRM. just use proto.	Stewart Smith	16 years ago
896.4.8	Just read table proto. Remove open_binary_frm. FRM is not re... Just read table proto. Remove open_binary_frm. FRM is not read anymore.	Stewart Smith	16 years ago
896.4.7	last bit of bitmap logic from FRM path to proto path last bit of bitmap logic from FRM path to proto path	Stewart Smith	16 years ago
896.4.6	move fixing auto_inc and blobs from FRM code path to table_p... move fixing auto_inc and blobs from FRM code path to table_proto code path	Stewart Smith	16 years ago
896.4.5	share->null_fields computed from proto. assert that same as ... share->null_fields computed from proto. assert that same as FRM	Stewart Smith	16 years ago
896.4.4	for db_create_options: options HA_X can be set even if HA_US... for db_create_options: options HA_X can be set even if HA_USED_X isn't set in used. Shows up in ALTER TABLE paths (show_check.test)	Stewart Smith	16 years ago
896.4.3	fix table_reader for new table_options fix table_reader for new table_options	Stewart Smith	16 years ago
896.4.2	db_create_option is silently truncated to 16bits in FRM writ... db_create_option is silently truncated to 16bits in FRM write path. Emulate that behaviour here.	Stewart Smith	16 years ago
896.4.1	read db_create_options from proto (was already stored, just ... read db_create_options from proto (was already stored, just had to set the right bits) instead of FRM keep an assert to ensure we're generating the same from the proto as we are storing in FRM.	Stewart Smith	16 years ago
896.3.15	don't compute number of null fields twice. (this bug even go... don't compute number of null fields twice. (this bug even got hit by my asserts). Stewart copy&paste=fail	Stewart Smith	16 years ago
896.3.14	stored_rec_length should include data_offset. stored_rec_length should include data_offset. This was causing vcol failures in innodb.	Stewart Smith	16 years ago
896.3.13	compute key_part->offset instead of storing in proto. compute key_part->offset instead of storing in proto.	Stewart Smith	16 years ago
896.3.12	don't need to read the real column type for VIRTUAL column w... don't need to read the real column type for VIRTUAL column when computing record offsets.	Stewart Smith	16 years ago
896.3.11	properly calculate field offsets in record buffer in parse_t... properly calculate field offsets in record buffer in parse_table_proto. vcol fields come at end of record.	Stewart Smith	16 years ago
896.3.10	fix small memory leak in drizzle_read_table_proto: delete in... fix small memory leak in drizzle_read_table_proto: delete input stream	Stewart Smith	16 years ago
896.3.9	create proper null terminated string in fix_fields_vcol_func create proper null terminated string in fix_fields_vcol_func	Stewart Smith	16 years ago
896.3.8	correctly store virtual column expression in proto correctly store virtual column expression in proto	Stewart Smith	16 years ago
896.3.7	fix table_reader for virtual columns fix table_reader for virtual columns	Stewart Smith	16 years ago
896.3.6	Read Fields out of proto instead of FRM. Read Fields out of proto instead of FRM. As non-radioactive fallout: - no implicit defaults (see http://bugs.mysql.com/bug.php?id=43151) - no real defaults for BLOB yet... but this paves the way (and it shouldn't be too hard now) As radioactive fallout: - parse_table_proto is approaching the length of War and Peace - handling of corrupted proto is about non-existant. abort() is not ideal	Stewart Smith	16 years ago
896.3.5	fix indentation so I can remotely read the function. fix indentation so I can remotely read the function.	Stewart Smith	16 years ago

Older »