[06:15] <fabbione> morning
[10:43] <elmo> NO WAY
[10:43] <elmo> fabbione: dude, now you've killed crested :-P
[10:46] <fabbione> no shit :)
[10:46] <fabbione> i have no access to it :P
[06:18] <lamont__> elmo: ping
[09:35] <jbailey-gcc> lamont, lamont__: ping?
[09:49] <lamont__> jbailey-gcc: yo
[09:53] <jbailey-gcc> lamont__: Was chatting with an IBM ppc guy, and he recommended that we do -mtune=power4 on ppc.
[09:53] <lamont__> iz certainly possible
[09:54] <lamont__> prolly wants to be an overridable tune, thouhg.
[09:54] <jbailey-gcc> Are we able to do that?
[09:55] <jbailey-gcc> I don't know how important allowing the package to override tune is, mind you.
[09:56] <lamont__> overrides come in 'if they don't say anything', always clear, and always force varieties.
[09:56] <jbailey-gcc> -mcpu is far more important to let the package override.
[09:57] <lamont__> yeah, but when the, oh say video/audio app, specifically builds things tuned for each of 3 processor families, overriding all of them to one specific family is, um, not so good.
[09:58] <jbailey-gcc> Right. =)
[10:05] <jbailey-gcc> Hmm, just chatting with James Morrison who hacks on the gcc sparc backend, and he's suggested that we might want to give similar treatment to sparc.
[10:06] <jbailey-gcc> although in that case we should do both mcpu and mtune, and possible mvis
[10:06] <lamont__> happy to take tunes - if you could email me the lot of them, I'll probably push new bits sometime soonish
[10:06] <lamont__> like after the weekend
[10:07] <lamont__> as it sits, the only thing we force on !i386 is -pipe
[10:07] <jbailey-gcc> I remember that, so I've been trying to find out interesting settings. =)
[10:08] <lamont__> please.  and thanks
[10:10] <jbailey-gcc> Will do.
[10:21] <cfan>  hi.. i want to shift to debian from fedora core 3.. can people tell me what flavour i should use in debian