I’m getting regular build failures where the rvm get stable step ends in:
Downloading https://github.com/rvm/rvm/archive/.tar.gz
curl: (22) The requested URL returned error: 404 Not Found
Could not download 'https://github.com/rvm/rvm/archive/.tar.gz'.
curl returned status '22'.
Downloading https://bitbucket.org/mpapis/rvm/get/.tar.gz
curl: (22) The requested URL returned error: 404 Not Found
Could not download 'https://bitbucket.org/mpapis/rvm/get/.tar.gz'.
curl returned status '22'.
I’m seeing this a lot too. The strange thing is that it almost always happens with the JRuby build, but not with other Rubies. Though, I have a hard time seeing the association.
We had issues with JRuby not working with the pre-installed version of RVM, so we injected the logic to always run rvm get stable. This was earlier this year. At the moment, with rvm get stable we end up with faulty RVM, and hence the build failure.
This is still a major problem and, unfortunately, it’s having the very negative effect of discouraging testing on JRuby.
What I can’t understand is why does the RVM download fail only for JRuby. Can’t RVM just be put in the container image anyway? Why does it have to download RVM every time when this is already the Ruby container?
@pkuczynski It would be really helpful if in the part I changed in https://github.com/rvm/rvm/pull/4722, it would print response headers upon error as well as the error code because in GitHub API, they contain additional information about the error. curl seems to only support displaying headers unconditionally though AFAICS.