accelio: fix i686-linux build

[dfoxfranke]

• Compile with gcc5 to avoid the compiler bug described in https://gcc.gnu.org/ml/gcc-patches/2014-05/msg02560.html
• Add a patch to fix the many incorrect printf format specifiers and other sloppy type conversions that gcc5 catches and warns on (erroring out due to -Werror).

cc @wkennington

[wkennington]

Is this patch not included upstream? I'm building accelio fine on gcc5 so
I'm not sure why all of this is needed.

On Sun, Aug 30, 2015, 20:55 Daniel Fox Franke notifications@github.com
wrote:

Compile with gcc5 to avoid the compiler bug described in
https://gcc.gnu.org/ml/gcc-patches/2014-05/msg02560.html

Add a patch to fix the many incorrect printf format specifiers and
other sloppy type conversions that gcc5 catches and warns on (erroring out
due to -Werror).

cc @wkennington https://github.com/wkennington

You can view, comment on, or merge this pull request online at:

#9570
Commit Summary

• accelio: fix i686-linux build

File Changes

• M pkgs/development/libraries/accelio/default.nix
  https://github.com/NixOS/nixpkgs/pull/9570/files#diff-0 (2)
• A pkgs/development/libraries/accelio/fix-printfs.patch
  https://github.com/NixOS/nixpkgs/pull/9570/files#diff-1 (599)
• M pkgs/top-level/all-packages.nix
  https://github.com/NixOS/nixpkgs/pull/9570/files#diff-2 (2)

Patch Links:

• https://github.com/NixOS/nixpkgs/pull/9570.patch
• https://github.com/NixOS/nixpkgs/pull/9570.diff

—
Reply to this email directly or view it on GitHub
#9570.

[dfoxfranke]

It errors out for me. Are you turning off -Werror or some such?

...and I just noticed it still does on x86_64-linux because there's a conversion I missed, so hold off on merging this until I fix that.

[wkennington]

Nope, but I'm using gcc5 in the stdenv so maybe that changes things?

On Sun, Aug 30, 2015, 21:00 Daniel Fox Franke notifications@github.com
wrote:

It errors out for me. Are you turning off -Werror or some such?

...and I just noticed it still does on x86_64-linux because there's a
conversion I missed, so hold off on merging this until I fix that.

—
Reply to this email directly or view it on GitHub
#9570 (comment).

[dfoxfranke]

Most of those things I changed from %lu to %llu actually need to be changed to PRIu64 (a macro from <inttypes.h>) to be portable. I'll fix it tomorrow.

[wkennington]

You should try and get this upstreamed if not already.

On Sun, Aug 30, 2015, 21:48 Daniel Fox Franke notifications@github.com
wrote:

Most of those things I changed from %lu to %llu actually need to be
changed to PRIu64 (a macro from <inttypes.h>) to be portable. I'll fix it
tomorrow.

—
Reply to this email directly or view it on GitHub
#9570 (comment).

[dfoxfranke]

I will, once I get it right.
On Aug 31, 2015 12:51 AM, "William A. Kennington III" <
notifications@github.com> wrote:

You should try and get this upstreamed if not already.

On Sun, Aug 30, 2015, 21:48 Daniel Fox Franke notifications@github.com
wrote:

Most of those things I changed from %lu to %llu actually need to be
changed to PRIu64 (a macro from <inttypes.h>) to be portable. I'll fix it
tomorrow.

—
Reply to this email directly or view it on GitHub
#9570 (comment).

—
Reply to this email directly or view it on GitHub
#9570 (comment).

[lucabrunox]

Beware that %llu may break the x64 version. The correct format for size_t is %zu.

[dfoxfranke]

@lethalman The offending arguments are (mostly) uint64_t, not size_t.

[garbas]

@dfoxfranke this PR fails for me on x64

[dfoxfranke]

@garbas yes, this is currently WIP until I make the %llu -> PRIu64 fix I described above.

[domenkozar]

cool, let me know when it's ready

[dfoxfranke]

I've revised the patch and this should now be ready to go. The patch also fixes all the tests and examples, so in a separate commit I've re-enabled building them and set doCheck = true.

I'll submit these changes upstream once I get back from vacation next week.

[dfoxfranke]

Bah... hold off one more moment. I just want to make a cosmetic change to the patch file.

[dfoxfranke]

Done.

[domenkozar]

it breaks kernel module(s) http://hydra.nixos.org/eval/1219660#tabs-now-fail

[dfoxfranke]

@domenkozar Revert 8b66350 for the time being.