Martin Robinson
Martin Robinson
I'm not sure if we have any plans to release numbered versions any time soon and we likely won't make a tag for every nightly, so I'm not sure this...
Servo isn't ready to be packaged as it does not expose a stable C-compatible API nor is it prepared to be a Cargo crate due to #26275.
The layout 2013 failures are because we do not run layout 2013 WPT tests by default so results are only updated once a week during WPT imports.
> Does this mean I don't need to bother updating layout_2013 test results on new PRs? Correct.
Closing in favor of #32812.
I suspect that different modules will want to handle IPC failures in different ways. For instance in the case of WebGPU, it will likely need to shut down the task...
> Or are you suggesting that we should actually replace all such patterns with actual error handling not just logging, so why bother with making it easier. Yes, I think...
> should I do it part of this PR? If it's possible to write a macro to halt execution of the current function, trigger the shutdown of the WebGPU task...
> > > should I do it part of this PR? > > > > > > If it's possible to write a macro to halt execution of the current...
> They are cleaned when Drop of associated dom objects is called (dropping GPUBuffer will send `DropBuffer` to wgpu thread), although if the thread is dead (that's probably the reason...