Closed Bug 1480927 Opened 2 years ago Closed 1 year ago
Clean up the hazard annotations
Not all annotations are still needed. Go through and rationalize them.
...but first, add some annotations that various other nsISupports subclasses will not GC. Hopefully I'll be able to make use of the builtinclass annotations soon and remove all of these.
Attachment #8997620 - Flags: review?(jcoppeard)
I'm still working on the patch that removes a bunch of annotations. I thought I should have gotten everything, but the shell build is coming back with hazards, for a somewhat amusing reason. There's a PerformanceGroup class with a pure virtual Delete() method, and it's implemented in one place, nsPerformanceGroup. Which is not part of the shell compile -- so the hazard analysis decides it doesn't know how to resolve that virtual method call and assumes it can do anything. So oddly enough, I will probably end up with more hazards reported in the shell than in the full browser!
FWIW, the annotations for the *AutoRealm destructors we can definitely remove AFAIK (because we no longer wrap pending exceptions there). I was considering posting a patch for this but if you're going through them anyway... :)
Attachment #8997620 - Flags: review?(jcoppeard) → review+
Pushed by firstname.lastname@example.org: https://hg.mozilla.org/integration/mozilla-inbound/rev/f6684513f10a Whitelist some more nsISupports virtual functions, r=jonco
Status: ASSIGNED → RESOLVED
Closed: 1 year ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.