| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
There was a problem hiding this comment.
lint issues otherwise LGTM
Sorry, something went wrong.
|
Yep, I'll fix those in the morning. My editor always seems to want to inject those spaces where they shouldn't be. 🙈 |
Sorry, something went wrong.
|
Fast-track has been requested by @Qard. Please 👍 to approve. |
Sorry, something went wrong.
There was a problem hiding this comment.
RSLGTM
Sorry, something went wrong.
Sorry, something went wrong.
Sorry, something went wrong.
PR-URL: #47520 Reviewed-By: Robert Nagy <ronagy@icloud.com> Reviewed-By: Rafael Gonzaga <rafael.nunu@hotmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>
PR-URL: #47520 Reviewed-By: Robert Nagy <ronagy@icloud.com> Reviewed-By: Rafael Gonzaga <rafael.nunu@hotmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>
PR-URL: nodejs/node#47520 Reviewed-By: Robert Nagy <ronagy@icloud.com> Reviewed-By: Rafael Gonzaga <rafael.nunu@hotmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>
PR-URL: nodejs/node#47520 Reviewed-By: Robert Nagy <ronagy@icloud.com> Reviewed-By: Rafael Gonzaga <rafael.nunu@hotmail.com> Reviewed-By: Gerhard Stöbich <deb2001-github@yahoo.de>
| Back | FazBrowse Home | New Git URL |
I discovered a bug in some of the changes I landed with TracingChannel. Technically the bug existed before, but we've been discouraging use of channel.subscribe(...) and channel.unsubscribe(...). With the TracingChannel changes I included a fix for the GC issue that allows channel.subscribe(...) and channel.unsubscribe(...) to work as they were originally intended, and some simplifications were made to make the top-level functions delegate to those. Unfortunately, there was an unnoticed issue where if the subscriber count ever dropped to zero the weakref would be deleted from the channels map but channel objects could still be held and therefore could still have subscribers added which would attempt an incRef on a weakref that's not there anymore. To handle that safely the delete actually has to happen only when there are no more strong refs, meaning when GC happens, so I've changed it to use FinalizationRegistry instead.