On 2022-01-08, at 17:05:37 +0000, Jeremy Sowden wrote: > On 2022-01-06, at 23:15:31 +0100, Jan Engelhardt wrote: > > On Thursday 2022-01-06 22:09, Jeremy Sowden wrote: > > >+ dnl The [MYSQL_CLIENT] macro calls [_MYSQL_CONFIG] to locate mysql_config. > > >+ > > >+ _MYSQL_CONFIG > > > > One caveat of m4 macros is that they may be left unexpanded if not > > found, and it is up to the tarball producer to ensure the m4 macro is > > expanded. Over the years, I built the opinion that this is not always > > a nice experience to have. > > > > I would do away with _MYSQL_CONFIG and just attempt to run > > `mysql_config` out the blue. sh failing to execute mysql_config, or a > > compiler failing to find mysql.h as part of AC_CHECK_HEADER is a nicer > > experience than _MYSQL_CONFIG being left accidentally unexpanded. > > I'll use `m4_ifdef` to add a fall-back. I took another look at what the macros in mysql.m4 give us, and all we need is what we implement for libpcap and libpq, so I've updated the mysql patch to match those. J.
Attachment:
signature.asc
Description: PGP signature
