-
Notifications
You must be signed in to change notification settings - Fork 30.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revert "path: fix bugs and inconsistencies" #55414
Revert "path: fix bugs and inconsistencies" #55414
Conversation
Review requested:
|
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #55414 +/- ##
==========================================
- Coverage 88.40% 88.40% -0.01%
==========================================
Files 653 653
Lines 187600 187518 -82
Branches 36117 36089 -28
==========================================
- Hits 165854 165770 -84
+ Misses 14974 14971 -3
- Partials 6772 6777 +5
|
FWIW the npm issue was already reported in the comments of the original PR, see #54224 (comment). |
I wonder why no one (besides you) followed up on that. It should never have landed. Still, we should have more tests on CITGM to verify these changes. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Since my PR to fix inconsistencies between functions broke a lot of things in other dependencies, it would be better to revert the PR rather than fix all the other dependencies. Thank you for the PR to revert the changes. |
We used to test npm in CITGM, but it often failed (nodejs/citgm#897). I think it was removed as part of one of the clear outs of persistently failing modules. There's a (stalled?) PR to add it back: nodejs/citgm#979 |
fixes #55424 |
This also broke a ton of Yarn's e2e tests, did CITGM miss that too? 🤔 |
For discussion of the CITGM, see nodejs/citgm#1067 |
I'm not sure if our tests cover it :) https://github.com/nodejs/citgm/tree/main/test/yarn |
Oh, yeah, good point. |
Hey, can this land with the passing CI and approvals? |
Looking forward to see 👀 this in the next semver minor or patch release asap ⏰ coz current Node 23 version a little bit broken, atm 🫠 |
Landed in ee46d22 |
Great! Should this go out in a patch, or in the next minor? |
great question, when will this fix land? |
npm is very broken; it should go out in a patch asap if possible. |
FWIW I just asked about a potential v23.0.1 patch in slack https://openjs-foundation.slack.com/archives/C019MGJQ8RH/p1729695807771819 |
This reverts commit efbba60. PR-URL: nodejs#55414 Reviewed-By: Claudio Wunder <cwunder@gnome.org> Reviewed-By: Richard Lau <rlau@redhat.com> Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
This reverts commit efbba60. PR-URL: nodejs#55414 Reviewed-By: Claudio Wunder <cwunder@gnome.org> Reviewed-By: Richard Lau <rlau@redhat.com> Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
This (uncleanly) reverts commit efbba60.
Fixes #55410
The commit efbba60 introduced a change where paths would retain trailing
/
characters. This change led to issues innpm
and likely other libraries, because trailing/
characters are not typically expected to be preserved.The purpose of using
resolve()
on a path is to fully resolve it to its canonical form, regardless of how it is written (e.g., both/hello/../world
and/hello/../hello/../world
should resolve to the same path). However, this commit altered that behavior, making/hello/
resolve differently from/hello
. That made the resolution of identical paths (one with/
and one without) return different results. This reverts that behavior