Sorry, I mentioned this on irc to emilio but didn't document it here. We can probably take this patch, since it's strictly better than what we have, but it still has bugs where we don't invalidate. The current invalidation model is that explicit invalidations are marked on the frame, and reflow position/size changes are detected by DLBI geometry comparison. This patches handles the former, but not the latter. This could happen if an ancestor-size changes size, which causes the ancestor of the text to be reflowed into a new position. This changes the resulting position of the text within the mask, and we need to invalidate, but nothing on either the text frame or background frame changed, and we wouldn't detect it. I'm not sure if there's a great solution to fix this, but I'll think about this more. Is there an urgent rush to fix this, the bug has existed for a while?
Bug 1525372 Comment 11 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Sorry, I mentioned this on irc to emilio but didn't document it here. We can probably take this patch, since it's strictly better than what we have, but it still has bugs where we don't invalidate. The current invalidation model is that explicit invalidations are marked on the frame, and reflow position/size changes are detected by DLBI geometry comparison. This patches handles the former, but not the latter. This could happen if an ancestor's sibling changes size, which causes the ancestor of the text to be reflowed into a new position. This changes the resulting position of the text within the mask, and we need to invalidate, but nothing on either the text frame or background frame changed, and we wouldn't detect it. I'm not sure if there's a great solution to fix this, but I'll think about this more. Is there an urgent rush to fix this, the bug has existed for a while?