--NEW-LEAKS-----------------------------------leaks------leaks% nsCaseConversionImp2 12 - 2524 MOZ_CO_DATE=2006:12:01:10:44 2536 MOZ_CO_DATE=2006:12:01:12:16 Check-ins: http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=HEAD&branchtype=match&dir=&file=&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=2006%3A12%3A01%3A10%3A44%3A00&maxdate=2006%3A12%3A01%3A12%3A16%3A00&cvsroot=%2Fcvsroot Not a huge deal, but probably something that should be fixed.
I think this is WONTFIX as the basic idea was to leave a static instance hanging around that would persist through XPCOM re-initialization.
agree, but why are we leaking 12 objects? or is that 12 bytes?
That's 12 bytes ( = 1 instance). -> WONTFIX
We should make it not screw up the leak stats. Can we just remove refcnt logging from this class?
If I'm reading this right then we'd have to manually implement addref/release. Is there no other way to disable refcount logging?
Created attachment 247467 [details] [diff] [review] Patch Here we go.
11 years ago
Why don't we just make the refcount always 1? This object never gets destroyed anyway.
Why do we have a refcount at all?
Wait, so you want them to be no-ops? Right now no one else creates one of these guys, but do we really want to neuter all instances?
Why do we need more than one of them? Always return the singleton instance.
Created attachment 247899 [details] [diff] [review] Patch Okay, now just returns 1.
Comment on attachment 247899 [details] [diff] [review] Patch This is ok, but I think you still need a singleton-constructor.
Created attachment 250367 [details] [diff] [review] Patch Now with singleton constructor. I also removed the unused and potentially problematic NS_NewCaseConversion function.
Fixed on trunk.
And nsCaseConversionImp2 has dropped from the rlk logs