Default ISA/tuning flags for GCC, --enable-kernel= level for glibc

[Matthew Woehlke]

Michael Cronenworth wrote:
> -------- Original Message --------

Can you please not quote full e-mail headers? Aside from feeding 
spammers, it's a waste of bandwidth.

> Gerd Hoffmann wrote:
>> Can we *please* wait a bit until raising the bar *that* high?  Breaking
>> mock builds for rawhide on F9/10 releases is a really bad idea IMHO.
>> Which meas supporting at least 2.6.27.  Maybe in a year when F9 is EOL
>> and F10 has a 2.6.29 kernel update we can reconsider 2.6.29.
> 
> So when 2010 rolls around you'll be saying "pleaseeeee wait until F12 
> has reached EOL"?

If at that point you're calling for 2.6.34 or whatever is current then, 
then of course you'll get the same complaining.

I fail to see how supporting live upgrade from an in-support version is 
unreasonable.

> The pain has to occur sometime.

Why? Drop support for not-supported kernels. Currently that seems to 
mean 'earlier than 2.6.18'. In the future, that will be something 
higher, and we can bump accordingly. There is a lag built into this 
system, but that's much different than stagnation.

> How about rolling up to 2.6.18 right now? That shouldn't hurt anything, 
> should it?

Unless I missed something, Gerd didn't say anything to indicate a 
problem with that. I see rather 2.6.27 was mentioned, which is later 
even than 2.6.18 (although I'd have to agree with Jesse about supporting 
2.6.18).

-- 
Matthew
Please do not quote my e-mail address unobfuscated in message bodies.
-- 
find / -user your -name base -print0 | xargs -0 chown us:cats -- Unknown