You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We're seeing some issues with bundler/inline activating an older bundler version picked up from the Gemfile.lock.
At best this is confusing, at worst we ran into some odd gem activation errors that I'm, so far, unable to reproduce in a concise way, but I can promise you they were awful to debug ;)
Did you try upgrading rubygems & bundler?
Yes.
Post steps to reproduce the problem
First, setup a non-inline Gemfile for the application:
Describe the problem as clearly as you can
We're seeing some issues with
bundler/inline
activating an older bundler version picked up from theGemfile.lock
.At best this is confusing, at worst we ran into some odd gem activation errors that I'm, so far, unable to reproduce in a concise way, but I can promise you they were awful to debug ;)
Did you try upgrading rubygems & bundler?
Yes.
Post steps to reproduce the problem
First, setup a non-inline Gemfile for the application:
and activate an older ruby version:
and check the bundle version:
looks good.
Now add our binstub helper, we want this using an independent version of ruby+bundler:
and test it
okay, that's fine, but let's switch to a newer ruby, and 'accidentally' install the same older version of bundler:
which gives us
this shouldn't matter if
bundler/inline
is independent, butWe unexpectedly activate the old version from the
Gemfile.lock
in the parent directory.The current workaround for us is to set
BUNDLE_GEMFILE
to nothing:What were you expecting to happen?
bundler/inline
to ignore anyGemfile.lock
and use the most current version of bundler available.What actually happened?
bundler/inline
still used the 'BUNDLED WITH' version from the Gemfile.lock in the parent directory.The current workaround for us is to set
BUNDLE_GEMFILE
I'm not clear how engrained the
Gemfile.lock
version activation is, but it seems less surprising to disable that forbundler/inline
.The text was updated successfully, but these errors were encountered: