Public SourceMage Discussion List

[Arwed von Merkatz <v.merkatz AT gmx.net>]

On Tue, Dec 05, 2006 at 07:28:48PM -0600, morfic wrote:
> On Tue, 05 Dec 2006 17:07:40 -0800
> seth AT swoolley.homeip.net wrote:
> 
> > On Wed, Dec 06, 2006 at 01:02:26AM +0100, "Andra?? 'ruskie' Levstik" 
> > wrote:
> > > IMHO
> > > 
> > > Drop all extra flags and only leave a minimal set... the rest can be 
> > > setup
> > > through the custom ones or the extra ones.
> > 
> > That's not how things work.  If we do drop them we need to QA the spells 
> > between
> > for change with and without it and fix them before we do it.
> > 
> > Let's go forwards, not backwards.
> > 
> > Seth
> 
> I probably miss something, how does leaving out REDUNDANT flags require QA? 
> even if it's not commandline, gcc still uses it all the same.
> Which is why i don't understand how things require QA on that.
> 
> - -O2 -march=k8 -pipe is the same as -O2 -march=k8 -pipe -mmmx -m3dnow 
> -msse -msse2 -mfpmath=sse -fomit-frame-pointer and perhaps something i 
> don't even think of right now
> All you need to do for QA is look at what flags gcc sets for k8, and you 
> can find that in the gcc source, then you take those out, and you know you 
> lose or gain nothing.
> 
> What am i missing?

That's almost correct. The problem is that there are spells checking for
certain flags in $CFLAGS and enabling/disabling parts of the build based
on that. Not too many spells do that, but it needs to be checked.

-- 
Arwed v. Merkatz                              Source Mage GNU/Linux developer
                                                    http://www.sourcemage.org