On 10/12/14 11:05, Rich Freeman wrote:
On Sun, Oct 12, 2014 at 8:23 AM, Anthony G. Basile <[email protected]> wrote:
I'm working on that. I'm not sure that #499996 is hardened i686 specific
right now. I'm hitting it on vm with even the vanilla gcc so something else
might be going on here.
VLC built fine on stable amd64 with near-default make.conf in my
testing. It may be an x86 issue though.
Roughly, there are two classes of bugs, those that are in gcc itself because
of bad implementation (usually arch specific like the situation with the
poor R10000 mips), and those that gcc exposes in other packages. I don't
see any of the former category right now.
True. Actually, the issue might be more that x86 in general isn't
getting the kind of love it used to.
Don't get me wrong - I'm not suggesting we should demote it or
anything. I'm just wondering if this is a portent of things to come.
Just as most amd64 bugs were the result of mismatched types and the
like, the same can work in reverse, and it is good to eliminate these
bugs so that the next time some arch shift comes along we're not
starting over from ground zero again.
--
Rich
Yeah we're not demoting anything, we're just trying to figure out what
we can live with and what we can't.
I think there's only one more issue we really can't live with and that's
the c++11 issue. 4.8 introduces -std=c++11 which breaks abi backwards
compatibility. You can see this in, for example, the webkit-gtk bugs.
If we stabilize gcc-4.8.3 now, then people using 4.8 to upgrade
consumers of webkit-gtk built with 4.7 and going to hit unresolved
symbols --- they'll have the characteristic c++ mangled name. We talked
about this before. I don't remember what we concluded, but thinking
about it now, the easiest approach would be to have some reverse
dependency contingent on a c++11 USE flag trigger the rebuilding of
webkit-gtk. Then we'll have a tracker for other packages which have the
same issue.
Comments?
--
Anthony G. Basile, Ph.D.
Gentoo Linux Developer [Hardened]
E-Mail : [email protected]
GnuPG FP : 1FED FAD9 D82C 52A5 3BAB DC79 9384 FA6E F52D 4BBA
GnuPG ID : F52D4BBA
