Re: CFLAGS

2004-11-02 Thread David Brodbeck
On Mon, 1 Nov 2004 14:20:35 -0500, Michael Barnes wrote > I was able to change the default CFLAGS by putting the CCFLAGS and > the CFLAGS values in my environment before running "perl Makefile.PL". > > I would guess that this could be considered a bug, because its not >

Re: CFLAGS

2004-11-01 Thread David Brodbeck
On Mon, 01 Nov 2004 11:54:50 -0800, Justin Mason wrote > BTW the default CFLAGS are coming from whatever perl was built with; > so I'd be worried about bugs in your perl accordingly ;) That's not where spamc is getting them. They're hard-coded in the configure script: ri

Re: CFLAGS

2004-11-01 Thread Justin Mason
chitecture. I've just been editing the configure script before > > > > building, but it'd be nice if there was an easier way. > > > > > > yep: > > > > > > perl Makefile.PL CCFLAGS="-O" > > > > Hmm...just tried that

Re: CFLAGS

2004-11-01 Thread Michael Barnes
f there was an easier way. > > > > yep: > > > > perl Makefile.PL CCFLAGS="-O" > > Hmm...just tried that with SpamAssassin 3.0.1 and it *still* built > spamc with -O2. I had to edit spamc/configure again to force it to > use -O. I was able to change the

Re: CFLAGS

2004-11-01 Thread David Brodbeck
On Wed, 22 Sep 2004 09:15:48 -0700, Justin Mason wrote > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > David Brodbeck writes: > > Is there a way to get the SpamAssassin build process to use -O instead of > > -O2 > > while building spamc? I run FreeBSD on a DEC Alpha, and -O2 triggers > > o

Re: CFLAGS

2004-09-22 Thread Justin Mason
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 David Brodbeck writes: > Is there a way to get the SpamAssassin build process to use -O instead of -O2 > while building spamc? I run FreeBSD on a DEC Alpha, and -O2 triggers > optimizer bugs in gcc on that architecture. I've just been editing the >

CFLAGS

2004-09-22 Thread David Brodbeck
Is there a way to get the SpamAssassin build process to use -O instead of -O2 while building spamc? I run FreeBSD on a DEC Alpha, and -O2 triggers optimizer bugs in gcc on that architecture. I've just been editing the configure script before building, but it'd be nice if there was an easier way.