Closed aaronjensen closed 1 month ago
The comment on the offending lines says:
# Need to be specific down to the lib path so queries generated in specs don't get ignored
What does this mean? Is this referring to specs in this project? Does this mean that explicit code was added to the applicative library work around problems in this projects tests? It seems like the tests should be able to control what they need to control in order to work and this code shouldn't need to resort to something so uncommon. I've seen projects use Gem::Specification to figure out if a gem is present (Rails does this), but I've never seen anything actually get the source location. That seems like a bridge too far.
ok this is caused by the changes in #2313, please pin to an older version for now till I release a fix. cc @st0012
@st0012 I agree with the comment above that if we're using Gem::Specification
only for tests, that's unjustified. Why not simply match on sentry-ruby/sentry-rails
as a simple string match in the backtrace_cleaner
?
Sorry for the trouble, I wasn't aware that invoking Gem::Specification
would cause issues in some setup. In the projects I worked with it's all been available so I wanted to get the most accurate source result with it. I'll open a PR to address it.
released in 5.18.1
Issue Description
https://github.com/getsentry/sentry-ruby/commit/7b00b250968164cb4bb6ee7ac4a777dc2e868252 introduced invocations of
Gem::Specification.find_by_name
, which doesn't work when using bundler standalone. This means that this release will cause production issues for people that use bundler standalone.Reproduction Steps
No time right now, having to deal with the issues created by this new release.
Expected Behavior
It doesn't crash when using bundler standalone
Actual Behavior
It crashes
Ruby Version
3.3.1
SDK Version
5.18.0
Integration and Its Version
No response
Sentry Config
No response