Poor Omnibox Color Scheme |
|||||||||||||
Issue descriptionGoogle Chrome 55.0.2883.75 (Official Build) (64-bit) OS: Linux What steps will reproduce the problem? (1) Have a profile configured such as zero-suggest triggers. (Google default search engine, signed in and syncing history without a custom passphrase, regular non-incognito window, maybe other conditions that I forget) (2) Go to http://www.imdb.com/title/tt2096673/ (3) After the page loads, focus in the omnibox. (I did it with control-l). What is the expected result? (4) All lines in the omnibox are visible. What happens instead? (4) The top line has practically no contrast and is not readable. See attached screenshot.
,
Dec 7 2016
I do not have a theme installed (or at least I'm not aware of ever installing a theme).
,
Dec 7 2016
pkasting, can you triage? I know I've seen you commenting on theming/coloring bugs recently.
,
Dec 7 2016
How does normal, non-zerosuggest interaction with the dropdown look on that same machine? The grey background thing is weird. Is that the selected result background? If you select something different, is the text of the first row all the same grey or something? Pretty sure this is Linux-specific.
,
Dec 7 2016
the theme you are using is the GTK theme. I'm sure it's specific to that setting (and probably your particular system theme).
,
Dec 7 2016
> Is that the selected result background? If you select something different, > is the text of the first row all the same grey or something? It is the selected result background. If I arrow to another selection, everything looks correct, including the now non-selected original item (no colored background). Furthermore, if I arrow back to the top item so it selected again, it looks right too. By the way, the selection color when doing all this is a kind of darkish-reddish-orange. If I use the mouse to hover over an item in the dropdown, I see the item gets selected using a lightish-reddish-orange. I can hover and thus highlight the bottom items in the dropdown. The top item remains with the weird gray background. If I use the mouse to hover over the top item, I don't see the light reddish-orange highlight; it keeps the weird gray background. > How does normal, non-zerosuggest interaction with the dropdown > look on that same machine? Looks fine, no issue. For example, top item as I type is always highlighted in the dark reddish-orange.
,
Dec 7 2016
Hovering a selected row leaves it painted as the selected row. It won't repaint it. It sounds to me as if there is some sort of strange bug where, only in the zerosuggest case, and only until you interact with the dropdown, the "selected row" background color is computed as grey instead of orange? What's the text color for selected rows -- is it that grey? Could we be wrongly using the text color as the background color this first time? (I did test on Windows Dev, and I can't reproduce problems -- the row highlight color is correct for zerosuggest there.)
,
Dec 13 2016
> It sounds to me as if there is some sort of strange bug where, only in the > zerosuggest case, and only until you interact with the dropdown, the "selected > row" background color is computed as grey instead of orange? That sounds like an accurate description of the problem. > What's the text color for selected rows -- is it that grey? When I highlight a row with the mouse (to get the light reddish-orange background color), the text color remains the same as unhighlighted rows: black of queries and green for URLs. When I use the arrow keys to explicitly change the selection, the text in the selected row (which has the dark reddish-orange highlight) is white-ish. Maybe I'll call it gray. In particular query text is a light gray (close to white) and URL text is a more typical gray. > Could we be wrongly using the text color as the background color this first time? I can believe this. If I stare closely at the screen, I can believe that the initial (bad) background color (above the current page URL suggestion) is the same as the text color on a line with a URL suggestion (the middle gray that appears when a URL suggestion is selected).
,
Dec 14 2016
thanks for the report and discussion. Fix is here: https://codereview.chromium.org/2578443002/
,
Dec 14 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/cfd026f99e1868040b970891ba75d9394c5b8efa commit cfd026f99e1868040b970891ba75d9394c5b8efa Author: estade <estade@chromium.org> Date: Wed Dec 14 03:48:23 2016 Refresh appearance of omnibox results when native theme changes. Previous to this patch, the render texts were created just once. In some cases that was before the view was actually added to a widget hierarchy and therefore GetNativeTheme returned the wrong thing. This only affects GTK theme mode because on all other platforms, there's only ever one NativeTheme. (Incognito does use a different NativeTheme as well but this only affected suggestions which we don't provide in incognito.) BUG= 672226 Review-Url: https://codereview.chromium.org/2578443002 Cr-Commit-Position: refs/heads/master@{#438418} [modify] https://crrev.com/cfd026f99e1868040b970891ba75d9394c5b8efa/chrome/browser/ui/views/omnibox/omnibox_result_view.cc [modify] https://crrev.com/cfd026f99e1868040b970891ba75d9394c5b8efa/chrome/browser/ui/views/omnibox/omnibox_result_view.h
,
Dec 14 2016
,
Dec 14 2016
so I think this bug has probably been around for about 21 months (when the code was added, or whenever it was turned on by default for the first time). That means it certainly affects m56, which branched recently. It also means merging this may not be a high priority, but the patch is really simple and merging it would get the fix out to users 8 weeks faster.
,
Dec 15 2016
Your change meets the bar and is auto-approved for M56 (branch: 2924)
,
Dec 19 2016
This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible! If all merges have been completed, please remove any remaining Merge-Approved labels from this issue. Thanks for your time! To disable nags, add the Disable-Nags label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Dec 19 2016
,
Dec 19 2016
,
Dec 23 2016
This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible! If all merges have been completed, please remove any remaining Merge-Approved labels from this issue. Thanks for your time! To disable nags, add the Disable-Nags label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Dec 23 2016
,
Jan 4 2017
Unable to reproduce the issue on Ubuntu 14.04 with chrome version #55.0.2883.75 Seen the same behavior in chrome version #56.0.2924.51 Attaching the screen-cast for your reference, could you please look into it and let us know any steps missed while reproducing the issue.
,
Jan 4 2017
I'm not sure if you followed all the steps in part (1) of the repro instructions. mpearson@ might be the best one to verify this, if we feel a strong need for verification.
,
Jan 6 2017
mpearson@ could you please confirm whether this issue is fixed in ##56.0.2924.51 ???
,
Jan 10 2017
Verified. Looks fine to me now on top of tree. Sorry, it took me a while to check.
,
Jan 11 2017
thanks for report and verification |
|||||||||||||
►
Sign in to add a comment |
|||||||||||||
Comment 1 by mpear...@chromium.org
, Dec 7 2016