Re: unknown conversion %m

Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> Done with that name. FYI, here is the complete set of warnings now
> generated on pitta:

The "unused variable" is flex's fault, not much we can do about that.
Seems like most of the others could be removed with some explicit casting.

> c:/mingw/msys/1.0/home/pgrunner/bf/root/HEAD/pgsql.3240/../pgsql/src/interfaces/ecpg/pgtypeslib/timestamp.c:505:6:
> warning: unknown conversion type character 'G' in format
> c:/mingw/msys/1.0/home/pgrunner/bf/root/HEAD/pgsql.3240/../pgsql/src/interfaces/ecpg/pgtypeslib/timestamp.c:685:6:
> warning: unknown conversion type character 'V' in format

These are a bit interesting. According to the Single Unix Spec, %V
has been standard for strftime since at least 1997, so it's damn odd
if MS' version doesn't support that. OTOH, %G is *not* in that standard
... should we try to avoid using that?

But it looks like all those cases are only reached if the ecpg-using
application tries to use those formats, so maybe any portability risks
there aren't our problem. Maybe a reasonable fix is the one that's
already there for %g, ie, just prevent the gcc check from occurring.

regards, tom lane