Menu
Search

Forums

Forum Index

September 17, 2013
Gravatar of David NadlingerPosted by David NadlingerReply
David Nadlinger
Gravatar of David Nadlinger


Reply
The Travis CI tests for std.process on Linux x86 have started to fail sporadically, see e.g. https://travis-ci.org/ldc-developers/ldc/jobs/11404239 and some of the builds before that.

It seems to be an OOM problem on the CI slave, but it's something we should investigate at some point anyway as it renders the build status indicator unreliable.

David
September 17, 2013
Gravatar of Kai NackePosted by Kai Nacke
in reply to David Nadlinger		Reply
Kai Nacke
Gravatar of Kai Nacke
Posted in reply to David Nadlinger


Reply
On Tuesday, 17 September 2013 at 17:19:53 UTC, David Nadlinger wrote:
> The Travis CI tests for std.process on Linux x86 have started to fail sporadically, see e.g. https://travis-ci.org/ldc-developers/ldc/jobs/11404239 and some of the builds before that.
>
> It seems to be an OOM problem on the CI slave, but it's something we should investigate at some point anyway as it renders the build status indicator unreliable.
>
> David

As it is related to the new multilib build:
We could move the multilib build to the "Allowed failures" section and create a normal LLVM 3.3 build.
This failure is really annoying. With LLVM 3.1 it occurred with every build.

Kai