Closed trhodeos closed 11 years ago
hmm - I made a clean nexus-2.2-01 install, configured the plugins as you did and made the following request
$ wget http://localhost:4711/nexus/content/gems/gems/api/v1/dependencies?gems=ha-internal-services,rake,redcarpet,yard,vcr,webmock,rspec,ha-gem-release,octokit,httparty,cachetastic"
and all went OK. dito with the same request on ruby-host and ruby-proxy directly.
the NPE happens in nested loop over all repos and all the "versions' found for each gemname.
was the storage for those repos all empty ?
Ah, now that I think about it, the gem I was looking for wasn't in either repo... Is error handling still a work-in-progress for bundler api support?
the biggest thing for me was "dead locks". since I added
sisu-resource-locks=local
into conf/nexus.properties I haven't seem any - yet. and once there are some corrupted files bundler needs to fall back on full index - see #11
if you still can reproduce that NPE - maybe you can narrow it down to a minimal Gemfile. the url from stacktrace did not trigger any errors with me.
Hmmm, I'm pretty sure that was the issue. I haven't seen it since.. Closing this issue.
Nexus version 2.2-01
The group repository contains one hosted and one proxy repo.
The hosted repo is listed before the proxy repo
This is the stacktrace I get in logs/wrapper.log: