From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; 0.8) Gecko/20010215 BuildID: 2001021508 Look in function nsJARURI::Equals the URI passed to the argument is QueryInterface'd to a nsIJARURI and assigned to a plain old nsIJARURI pointer, otherJar - not an nsCOMPtr. The reference is then never released. Reproducible: Didn't try Steps to Reproduce: I have read the code - I have not tried to measure how much memory this code is leaking. The questionable code is at... http://lxr.mozilla.org/seamonkey/source/netwerk/protocol/jar/src/nsJARURI.cpp#255

Marking NEW.

mass move, v2. qa to me.

Assign the results of the Queryinterface to a nsCOMPtr<nsIJARURI>.

to darin for review of patch

Comment on attachment 60313 [details] [diff] [review] proposed fix >Index: src/nsJARURI.cpp >- nsJARURI* otherJAR; >- rv = other->QueryInterface(NS_GET_IID(nsIJARURI), (void**)&otherJAR); >+ nsCOMPtr<nsIJARURI> otherJAR = nsnull; >+ nsJARURI* tempPtr = nsnull; >+ rv = other->QueryInterface(NS_GET_IID(nsIJARURI), (void**)&tempPtr); > if (NS_FAILED(rv)) > return NS_OK; // not equal >+ else >+ otherJAR = tempPtr; actually, this doesn't fix the leak. tempPtr still holds a reference, and assignment into a COMPtr adds one more reference. the correct solution would be something like this: nsCOMPtr<nsIJARURI> otherJAR( do_QueryInterface(other, &rv) ); if (NS_FAILED(rv)) return rv;

vinay: your patch needs work

vinay: wrong patch :P

Opps - attaching the new patch file. Sorry for the confusion.

Comment on attachment 60469 [details] [diff] [review] correct patch sr=darin

Comment on attachment 60469 [details] [diff] [review] correct patch sr=darin

Comment on attachment 60469 [details] [diff] [review] correct patch Looks good. r=gordon.

fix checked in.