[Broken] ldc-developers/ldc#157 (master - cf88418)

[Travis-CI]

The build was broken.

Repository	ldc-developers/ldc
Build #157	https://travis-ci.org/ldc-developers/ldc/builds/3533275
Changeset	https://github.com/ldc-developers/ldc/compare/e239bd6e8715...cf88418bd01a
Commit	cf88418 (master)
Message	Use shiny new debug libraries.
Author	David Nadlinger
Duration	28 minutes and 25 seconds

You can configure recipients for build notifications in your configuration file. Further documentation about Travis CI can be found here. For help please join our IRC channel irc.freenode.net#travis.

We need your help! Travis CI has run 406,714 tests for 5,442 OSS projects to date, including Ruby, Rails, Rubinius, Rubygems, Bundler, Node.js, Leiningen, Symfony ... If you use any of these then you benefit from Travis CI. Please donate so we can make Travis CI even better!

Nedap offers you the opportunity to work on technology that really matters: food, water, energy ...

Get up and running in minutes, and deploy instantly with git. Focus 100% on your code, and never think about servers, instances, or VMs again.

The delivery of this email was kindly sponsored by Postmark, an amazing service for transactional email delivery.

See all of our sponsors →

[David Nadlinger]

Interesting – none of these tests seem to fail on my LLVM 3.2/x86_64 box.

In any case, the failure were not really caused by these commits, they only switched the Phobos version used for release builds to be compiled with -O3 -release.

David

[David Nadlinger]

Okay, these failures seem to be caused by a issue in GC.mark() which
only occurs on x86, and seems to be fixed in LLVM 3.2. Not sure what
the best way going forward here is?

David

[David Nadlinger]

On Thu, Dec 6, 2012 at 5:34 PM, David Nadlinger <da...@klickverbot.at> wrote:
> Not sure what the best way going forward here is?

Just using -O2 instead of -O3 on pre-3.2 seemed to do the trick. But
while my local tests and the new multilib builds on ci.lycus.org pass,
the Travis builds are still failing.

Need to reactivate my actual 32 bit VMs for testing, I suppose.

David

[David Nadlinger]

I'm still on the go, but fired up an x86 EC2 instance running Ubuntu
12.04 LTS, and the problem also occurs there.

Backtraces look like stack corruption:
---
Program received signal SIGSEGV, Segmentation fault.
0xbffff0e0 in ?? ()
(gdb) bt
#0 0xbffff0e0 in ?? ()
#1 0x00441f0f in ?? ()
#2 0x0cec8300 in ?? ()
#3 0x07d588a1 in ?? ()
#4 0x24048908 in ?? ()
#5 0xfed3b0e8 in ?? ()
#6 0x8601b1ff in ?? ()
#7 0xc4832648 in ?? ()
---

No idea what is actually going on here (related to exception
unwinding?) Need to further investigate this later today. I still
wonder what we should do about this concerning the release… Package a
LLVM 3.2-based version?

David