glibc: pull a bug work-around from Arch (i686-only ATM) #1446
Conversation
This should work around some of our build problems, e.g. http://hydra.nixos.org/build/7575893/nixlog/1/tail-reload
|
Are you sure this only affects i686? The patch applies to sysdeps/x86_64... I would just drop the conditional. |
|
On Thu, Jan 02, 2014 at 01:29:57AM -0800, Eelco Dolstra wrote:
I think they have some code shared for i686/x86_64, especially that of SIMD. :) |
|
I haven't noticed affecting x86_64, and the Arch bug is also only filed against 32-bit versions (although they patch it unconditionally). Certainly feel free to drop the conditional -- it was just to limit a rebuild ATM, as it seemed to nave no effect elsewhere (I planned e.g. to port it to stdenv-updates without the condition). Yes, it looks like using SIMD even on i686, based on runtime CPU detection. I wonder if the problem is even reported upstream. |
This should work around some of our build problems, e.g. http://hydra.nixos.org/build/7575893/nixlog/1/tail-reload Cf. #1446
This should work around some of our build problems, e.g. http://hydra.nixos.org/build/7575893/nixlog/1/tail-reload Cf. #1446; unconditional on stdenv-updates: c6b9205
|
Pushed to master with i686 only. I believe in stabilizing stdenv-updates soon where it's for all platforms (finally!). |
This should work around some of our build problems,
e.g. http://hydra.nixos.org/build/7575893/nixlog/1/tail-reload
I tested the builds of jdk and xbmc were broken on my build machine, and now they work. It's a stdenv change for i686, but to me it seems worth it.
If noone protests, I plan to merge in about a day or two? (I don't see any Hydra branch suitable for pre-building this, for nixos:glibc-2.18 is stopped.)