https://bugs.kde.org/show_bug.cgi?id=401512
--- Comment #8 from David E. Narvaez <david.narv...@computer.org> --- (In reply to Kevin Kofler from comment #7) > The patch from comment #2 prevents the tuple from being de-refcounted all > the way down to 0 (because it adds a bogus unowned reference), so the tuple > is leaked, and its continued existence also keeps the arguments referenced. > (It is the destruction of the tuple that decreases the reference count of > each argument.) Ah, makes sense. > On the other hand, the patch from comment #5 looks like the correct fix to > me. I agree. Franco, can you post your patch to Phabricator? Against the release/19.12 branch please. -- You are receiving this mail because: You are watching all bug changes.
