[JoshuaMoelans]

Similar to how we do it in inproc, we re-assign to the event itself. The subsequent decref(event) becomes a no-op (since the return value of the on_crash on a discard is supposed to be sentry_value_new_null as per the contract), whereas before, we would have a double decref(event) (once by the user in the on_crash, and once in the !should_handle branch.

#skip-changelog

[JoshuaMoelans]

we still call decref(event) later on, but since this would now happen on a sentry_value_new_null it's a no-op (consistent with how inproc handles this). We could also remove the decrefs from both 🤔

[codecov]

Codecov Report

❌ Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 74.33%. Comparing base (efe5fff) to head (64749ae).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1990      +/-   ##
==========================================
+ Coverage   74.19%   74.33%   +0.14%     
==========================================
  Files         104      104              
  Lines       25634    25633       -1     
  Branches     4626     4626              
==========================================
+ Hits        19018    19054      +36     
+ Misses       5317     5276      -41     
- Partials     1299     1303       +4     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[jpnurmi]

How did it come up? Found by coincidence?

[JoshuaMoelans]

@jpnurmi yeah kind of, in the review of getsentry/sentry-unreal#1536 Claude found this 👀