> > I was hoping this would be in for beta 3, but alas - can someone
> > *please* commit the win32 version patch at:
> > http://candle.pha.pa.us/mhonarc/patches/msg00000.html
> Personally I was holding off in hopes of seeing a cleaner solution.
> A patch that requires every executable-building Makefile to
> have a platform-specific wart isn't going to be very
> maintainable. You already missed pg_config, plus whichever
> of the contrib modules build executables...
IIRC, when this patch was posted, pg_config was a shellscript. Definitly
when the original version of the patch was posted. But sure, I should
probably have updated it when that changed.
Do you have any ideas on *how* to do this? Or specifically which parts
you don't like (or is it the whole concept)? I had no idea you were
holding off based on that. I've asked repeatedly for comments about what
is bad, and this is the first time I heard something about the new patch
(other than the location of the files and the inclusion of the icon
which Peter E commented on sevearal times), if I'm not mistaken.
Part of the versioninfo struct is file-specific, part is product
specific. I see no way around it other than having some parts of it in
every Makefile. Sure, we could have some kind of central mapping file
somewhere, but I would consider that *less* maintainable.
I specifically did not include contrib modules. That can be done once we
have a final verdict on exactly how it should be.
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match