It seems necessary to fix WebRenderLayerManager::FlushRendering() work like ClientLayerManager::FlushRendering().

Here you have changed something WebRender specific and it has to do with mozilla::ipc::IPCResult. I had sometimes crashes with Ctrl+W tab closing when webrender was enabled. bp-0c38dcef-1d98-4336-b9dc-5fb310170510 (bug 1354198). Maybe this is the fix? :O

(In reply to Darkspirit from comment #2) > Here you have changed something WebRender specific and it has to do with > mozilla::ipc::IPCResult. I had sometimes crashes with Ctrl+W tab closing > when webrender was enabled. bp-0c38dcef-1d98-4336-b9dc-5fb310170510 (bug > 1354198). Maybe this is the fix? :O I looked into Bug 1354198, it seems like a different problem.

Rebased.

Comment on attachment 8872491 [details] [diff] [review] patch - Fix WebRenderLayerManager::FlushRendering() Review of attachment 8872491 [details] [diff] [review]: ----------------------------------------------------------------- ::: gfx/layers/wr/WebRenderBridgeParent.cpp @@ +937,5 @@ > } > } > > void > +WebRenderBridgeParent::FlushRendering(bool aSync) nit: (very lame nit, sorry): my eyes always trip on the convention that forces us to put 'a' before 'Sync' for arguments which make it look like the exact opposite word (async). "aIsSync" or "aForceSync" are less elegant but they don't cause me to misinterpret what I read. Our layers code is filled with aSync parameters so it's not a problem of this patch, but it'd be nice if we took on the habit of naming this parameter some way that is harder to misread (or maybe I'll just get used to it eventually).

Pushed by sikeda@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/63cd7aaf05d8 Fix WebRenderLayerManager::FlushRendering() r=nical