Skip to content

Commit

Permalink
Remove passing ruby versions from travis allowed failures
Browse files Browse the repository at this point in the history
  • Loading branch information
@dkubb
dkubb committed Dec 19, 2013
1 parent 2874b51 commit 4263f13
Showing 1 changed file with 0 additions and 4 deletions.
4 changes: 0 additions & 4 deletions .travis.yml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -20,10 +20,6 @@ matrix:
env: JRUBY_OPTS="$JRUBY_OPTS --debug" # for simplecov
- rvm: jruby-head
env: JRUBY_OPTS="$JRUBY_OPTS --debug" # for simplecov
allow_failures:
- rvm: ruby-head
- rvm: jruby-head
- rvm: rbx

This comment has been minimized.

Sign in to view

Copy link
@sferik

sferik Dec 19, 2013

Collaborator

IMHO, we should allow failures for all unstable Ruby versions. jruby-head may be passing today, but if it breaks tomorrow, I don’t think that should affect the overall build status for this project.

I added Rubinius to this list because, in my experience, even its “stable” versions have been buggy. If you disagree or believe this has changed, I’d be willing to remove rbx from the list but I believe -head versions should remain.

This comment has been minimized.

Sign in to view

Copy link
@sferik

sferik Dec 23, 2013

Collaborator

Here’s an example of rbx being flaky and why I think it should be included in allowed_failures: https://travis-ci.org/dkubb/memoizable/builds/15863756

This comment has been minimized.

Sign in to view

Copy link
@dkubb

dkubb Dec 23, 2013

Author Owner

@sferik that's a good point. I had never thought of it that way.

I had always been using it as an early warning system when the build starts to fail with a previously passing version of some implementation. When this happens I would mark it as an allowed failure, work on a repro and submit it to the tracker. When it started passing again, I would remove it from the matrix.

I guess I had been thinking more about using travis to help ruby implementors test my code, than I have about actually testing for flaws in the code itself.

This comment has been minimized.

Sign in to view

Copy link
@dkubb

dkubb Dec 23, 2013

Author Owner

I don't have any strong feelings about this, but if you do I can start to mark the *-head versions as being allowed to fail I can start doing that.
fast_finish: true
notifications:
irc:
Expand Down

0 comments on commit 4263f13

Please sign in to comment.