I have committed fixes for all the so-far-reported regression test
discrepancies, with the exception that some of the 'geometry' expected
files look to be out of date and I have no proposed updates. Would
those folks who reported problems check again, using either a fresh CVS
update or a snapshot dated later than this message?
I also hacked libpq++'s Makefile.in so that it picks up CXXFLAGS
from the template file or configure calculations, rather than
automatically absorbing all of CFLAGS into CXXFLAGS. This should
make things better on platforms where the C++ compiler doesn't like
the same switch set as the C compiler. However, we might now need
to take steps to add stuff back into CXXFLAGS on some platforms.
Please check this if you've had trouble building libpq++ in the past.
regards, tom lane
pgsql-hackers by date
|Next:||From: Thomas Lockhart||Date: 2000-03-31 06:24:20|
|Subject: Re: BIT datatype|
|Previous:||From: Don Baccus||Date: 2000-03-31 03:08:56|
|Subject: Re: slow join on postgresql6.5|