Bundler gets stuck in a resolution loop #2090

sferik opened this Issue Sep 7, 2012 · 2 comments


None yet
3 participants

sferik commented Sep 7, 2012

Here's a Gist with my Gemfile, gemspec, and a snippet of the output of DEBUG_RESOLVER=true bundle install: https://gist.github.com/d0cd1ef943e47ad85d9b

This may be a dupe of #2030 but @indirect seems to think its a different issue and asked me to document it here. Here's the IRC log:

[19:08:49] <@indirect> your Gemfile has a conflict
[19:08:53] <@indirect> not sure why Bundler isn't catching on
[19:08:55] <@indirect> but
[19:09:02] <@indirect> Builder 3.1.1 is getting activated
[19:09:19] <@indirect> and then one of your reqirements is builder (~> 3.0.0)
[19:09:23] <@indirect> that's… not resolvable
[19:09:41] <@indirect> so yes, the bug is that bundler should die with a useful error message
[19:09:43] <@indirect> sorry :(
[19:10:08] <@indirect> you can fix it by editing your gemspec to require builder ~> 3.0.0
[19:10:20] <@indirect> looks like one of your deps requires builder ~> 3.0.0
[19:10:23] <@indirect> and you only require ~> 3.0
[19:10:32] <@indirect> this is a bug in the gemspec integration code, I bet
[19:11:14] <@indirect> bet you $10 that if you moved all the gemspec deps into the gemfile, it would error out and finger the conflicting deps
[19:11:26] <@indirect> probably the conflict is a child dep of a development dep
[19:11:35] <@indirect> which is where the weird gemspec code lives
[19:11:47] <@indirect> sferik: ↑
[19:12:32] <sferik> indirect: cool. should I open a new issue?
[19:14:07] <@indirect> yeah, that would be helpful

dblock commented Jan 16, 2013

carlhuda#2248 may be related, smaller repro


indirect commented Feb 18, 2013

This should be fixed in 1.3.0.pre.8 and above (including master).

@indirect indirect closed this Feb 18, 2013

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment