Issue metadata
Sign in to add a comment
|
32ms startup performance regression in OmniboxTextView::CalculatePreferredSize() |
||||||||||||||||||
Issue descriptionData from the UMA Sampling Profiler shows that OmniboxTextView::CalculatePreferredSize() regressed browser process UI thread startup on 64-bit Chrome on Windows by 32ms. This occurred in the 69.0.3480.0 Canary release. The responsible CL appears to be https://chromium.googlesource.com/chromium/src.git/+/45bf0faac15aab71a2e0ea6eb090ff2fdad01017 Execution profile difference of Canary 69.0.3479.0 vs. 69.0.3480.0: https://uma.googleplex.com/p/chrome/callstacks?sid=18626d20573158a0d3b9987bca7afdb0
,
Jul 9
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/6aac6ccf30deb69d888247a853b9087610606bd4 commit 6aac6ccf30deb69d888247a853b9087610606bd4 Author: Leonard Grey <lgrey@chromium.org> Date: Mon Jul 09 14:50:17 2018 Views: Use cached size for Omnibox match separator Missed that the separator was an OmniboxTextView. Bug: 860789, 712268 Change-Id: I670fa25754f23a5aaaa9201f548d9b735f92ecf3 Reviewed-on: https://chromium-review.googlesource.com/1128187 Reviewed-by: Justin Donnelly <jdonnelly@chromium.org> Commit-Queue: Leonard Grey <lgrey@chromium.org> Cr-Commit-Position: refs/heads/master@{#573304} [modify] https://crrev.com/6aac6ccf30deb69d888247a853b9087610606bd4/chrome/browser/ui/views/omnibox/omnibox_match_cell_view.cc
,
Jul 13
wittman@ are we still seeing this as a regression or did it cancel out in layouts? I probably don't want to pull the change either way, but we might be able to split the difference somehow on creating the popup.
,
Jul 13
The regression is still there. The change above landed in 69.0.3487.0 and the diff between 69.0.3486.0 and 69.0.3487.0 shows no improvement: https://uma.googleplex.com/p/chrome/callstacks?sid=b726db3d43124a46971bbb314ca42ba1 |
|||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||
Comment 1 by lgrey@chromium.org
, Jul 6