[EMAIL PROTECTED] (Peter Eisentraut - PostgreSQL) writes:
> Log message:
> Fix CFLAGS selection to actually work. Add test to detect whether gcc's
> option -fno-strict-aliasing is available.
A couple comments on this patch ...
The immediately previous code would default to CFLAGS="-g" if
--enable-debug is specified and the compiler is not gcc. As it
now stands, we default to CFLAGS="-O -g" instead. I'm not convinced
this is a good idea. On most non-gcc compilers, this combination
gets you "-g" and possibly a ton of warnings.
I think you changed the behavior for the AIX port. Diffing
src/template/aix against the old contents:
diff -c -r1.11 -r1.14
*** pgsql-server/src/template/aix 2002/09/04 22:54:18 1.11
--- pgsql-server/src/template/aix 2003/10/25 15:32:11 1.14
***************
*** 1,9 ****
! if test "$GCC" = yes ; then
! CFLAGS='-O2 -pipe'
! else # not GCC
! CFLAGS='-O2 -qmaxmem=16384 -qsrcmsg -qlonglong'
case $host_os in
! aix3.2.5 | aix4.1*)
! CFLAGS='-qmaxmem=16384 -qsrcmsg' ;;
esac
! fi # not GCC
--- 1,7 ----
! if test "$GCC" != yes ; then
case $host_os in
! aix3.2.5 | aix4.1*) ;;
! *) CFLAGS="-O2 -qlonglong";;
esac
! CFLAGS="-O -qmaxmem=16384 -qsrcmsg"
! fi
The "case" is now useless because CFLAGS will always end up with the
later setting. Don't we want something like
if test "$GCC" != yes ; then
case $host_os in
aix3.2.5 | aix4.1*) CFLAGS='-qmaxmem=16384 -qsrcmsg' ;;
*) CFLAGS='-O2 -qmaxmem=16384 -qsrcmsg -qlonglong' ;;
esac
fi
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so that your
message can get through to the mailing list cleanly
