On Thu, Feb 02, 2006 at 01:49:02PM +0100, Waba wrote:
> Ok, I understand. Here are the three patches, expected to be applied in
> this order:
> - custom01.patch makes custom.pl work in strict+warning mode.
> - custom12.patch empties $set after each pass, preventing unnecessary
> code from being generated.
These look good to me. Good catch.
> - custom23.patch makes custom.pl generate sql_cmp() functions for long
One reason for this could be that on 32 bit architectures, long is 4 bytes,
and not enough to hold the assumed BIGINT type, according to:
Each data type in an SSQLS struct must be large enough to hold its
corresponding SQL field type. On 32 bit machines, long is redundant.
For your upcoming changes you mentioned, if you happen to have any test
code for your new features, please toss it in examples/ and include it in
your patch set too.