Conversation
|
Looks good. Thank you 👍 In this repository and in most repositories in ruby org, you don't need to create a short-lived issue just to pair with a bug fix PR. Of course a standalone bug report issue or an issue meant for bigger discussion is always welcome 😄 |
Oh, Ok Thank you so much for this info. 😄 |
* Match absolute paths in the irb_require already-loaded check $LOADED_FEATURES holds absolute paths for anything the real require loaded, so anchoring the check with \A made it unable to match them. require 'x' followed by irb_require 'x' re-evaluates x.rb in the session. Anchor on a path boundary instead, which still rejects the foo/foo2 substring match #1253 was fixing. * Initialize the counter before requiring in the new test The probe file read an uninitialized global, which warns on stderr under -w and tripped the test's own assert_empty(err) on the CI rows that enable warnings.
2 participants
closes #1252
Previously, requiring foo could be skipped if foo2.rb was already loaded. The filename matching now requires an exact match, and a regression test covers this case.