Christopher Vollick

This is being proposed in https://github.com/iNPUTmice/Conversations/pull/4214

Hello! We agree that it looked like the "transport" instead of "fingerprint" thing was a copy-and-paste error, and it turned out to be hiding another error. So, we've changed those...

I thought about that, but they clearly did this on purpose. The comment docs say "when the body is too big we'll ignore it", and they have code that reads...

Ok, don't merge this yet. Doing further testing I appear to be leaking the paintables. I need to go figure out how GLib / Vala / GTK handles deconstruction, I...

Ok, I've used a weakref for the delay now. I use a static to avoid capturing `this`, and then make an explicit weakref so I can test it in the...

Ok, I found more sources of leaks that are unrelated to my changes. These were just already leaking widgets sometimes, but that's more obvious with the animated gifs because the...

Right, that's fair... but it's definitely more involved. As soon as we have UI that controls the animation, now the image view needs to handle click, and then needs to...

> In case of click wouldn't the system just use the "viewer" app and that would animate it? Oh true! Ok, if what we want is a setting that's either...

Right, yeah. That's why the PR I'm working on right now (but got sidetracked by the memory leak hunt) allows collapsing of images, both animated and otherwise. Just cleaning up...

I've added some changes to the upstream weak timeout branch, and have updated this to include those changes.