59 bytes, text/x-review-board-request
When dealing with things in the purple buffer, we QI twice. Once to get the canonical nsISupports object, and once to get the CC participant. I think that with the way things work now, we can avoid the first one, because we're always passing in a canonical nsISupports pointer, using the Upcast mechanism. This is a low-risk way to avoid some of the overhead.
I think we can avoid both QIs for NoteXPCOMRoot because of the way it is used, though that will be a smaller win.
Comment on attachment 8891664 [details] Bug 1385459 - Don't use QI to canonicalize nsISupports pointers in the purple buffer. https://reviewboard.mozilla.org/r/162738/#review168464
Attachment #8891664 - Flags: review?(bugs) → review+
Pushed by firstname.lastname@example.org: https://hg.mozilla.org/integration/autoland/rev/80647273e2f0 Don't use QI to canonicalize nsISupports pointers in the purple buffer. r=smaug
You need to log in before you can comment on or make changes to this bug.