Note: it seems that this did indeed fix things for attachment 9158023 [details] (asuth's persnosco testcase, "slow-grid.html bad things happen on scroll"). I can scroll that in current Nightly with no noticeable rendering delays (whereas if I turn off the pref, I see extreme many-second delays and sometimes hangs). However, the reporter's testcase that was quoted in comment 0 (which I attached as attachment 9221148 [details]) seems to still be pretty slow -- its render time has been cut in half (so it takes ~4 seconds instead of ~8 seconds, roughly). That testcase is approximately the same as the testcase in bug 1682686, so let's track that remaining slowness over there.
Bug 1591366 Comment 22 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Note: it seems that this did indeed fix things for attachment 9158023 [details] (asuth's persnosco testcase, "slow-grid.html bad things happen on scroll"). I can scroll that testcase in current Nightly with no noticeable rendering delays (whereas if I turn off the pref, I see extreme many-second delays and sometimes hangs). However, the reporter's testcase that was quoted in comment 0 (which I attached as attachment 9221148 [details]) seems to still be pretty slow -- its render time has been cut in half (so it takes ~4 seconds instead of ~8 seconds, roughly). That testcase is approximately the same as the testcase in bug 1682686, so let's track that remaining slowness over there.
Note: it seems that this did indeed fix things for attachment 9158023 [details] (asuth's persnosco testcase, "slow-grid.html bad things happen on scroll"). I can scroll that testcase in current Nightly with no noticeable rendering delays (whereas if I turn off the pref, I see extreme many-second delays and sometimes hangs). However, the reporter's testcase that was quoted in comment 0 (which I attached as attachment 9221148 [details]) seems to still be pretty slow -- its render time has been cut in half (so it takes ~4 seconds instead of ~8 seconds, roughly; vs. it has no render delay in Chrome). That testcase is approximately the same as the testcase in bug 1682686, so let's track that remaining slowness over there.