Closed Bug 1234513 Opened 10 years ago Closed 10 years ago

Gif tweets not detaching from timeline properly at one click and comments are hidden

Categories

(Core :: Layout, defect)

29 Branch
defect
Not set
normal

Tracking

()

RESOLVED FIXED
Tracking Status
firefox43 --- wontfix
firefox44 --- wontfix
firefox45 --- wontfix
firefox46 --- fixed
firefox47 --- fixed
firefox48 --- fixed

People

(Reporter: bmaris, Unassigned)

Details

(Keywords: regression)

Attachments

(2 files)

Affected builds: - Firefox 43.0.2 RC build 1 - latest Developer Edition 45.0a2 - latest Nightly 46.0a1 Affected OS's: - Windows 7 64-bit - Mac OS X 10.9.5 - Ubuntu 12.04 32-bit STR: 1. Start Firefox 2. Visit https://twitter.com/ 3. Login to account 4. Scroll down the timeline to a GIF. 5. Click the GIF 6. Repeat step 5. Expected results: The tweet detaches from the timeline at the bottom and at the top after a click on it and reattaches after the second click on the gif. Also the comments are displayed. Actual results: The tweet does not detach at first click , it requires two and the comments are not displayed. Notes: 1. This is not a recent regression, reproduces with Firefox 36.0 build 2 as well. Will try and find a regression though ASAP. 2. Attaching two gifs one with Firefox behavior, one with Chrome.
And here is the Gif showing the behavior on Chrome.
And here is the regression range (only central, inbound not available because of the old regression): M-C Last good: 2014-01-14 First bad: 2014-01-15 Last good revision: 34dddf6f7ec1 (2014-01-14) First bad revision: 81bced59e8b3 (2014-01-15) Pushlog: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=34dddf6f7ec1&tochange=81bced59e8b3 Maybe this is the culprit: 0d816dd5e310 Andreas Pehrson — Bug 877115 - Moz2Dify CanvasLayerD3D10 drawing shared surface into D3D10_MAPPED_TEXTURE2D. r=nical e6da040d1f81 Andreas Pehrson — Bug 877115 - Moz2Dify CanvasLayerD3D9 drawing shared surface into D3DLOCKED_RECT. r=nical Adding needinfo to Nicolas Silva since he was mentor in this bug and maybe he can point out to someone else if this is not the bug that caused this issue.
Flags: needinfo?(nical.bugzilla)
Keywords: regression
This is definitely not graphics related. It also reproduces on linux so the D3D patches here (only affecting windows) can't be causing this. Could be related to the way video controls are implemented/stylable in gecko, perhaps, or something closer to DOM logic. Not that these gifs on twitter are in fact mp4 videos made to look like gifs, but as far as the browser sees them they are videos. Paul, maybe you know something about how video controls could interact with this ?
Flags: needinfo?(nical.bugzilla) → needinfo?(padenot)
No, sorry. Jared, is there something weird with <video> and clicks in Firefox when controls are not shown ?
Flags: needinfo?(padenot) → needinfo?(jaws)
Videos without the controls showing should act like any other replaced element as far as I know. The XBL binding is only attached when the 'controls' attribute is present.
Flags: needinfo?(jaws)
Could also come from bug 865407 perhaps? Rick, does that make any sense to you? Bogdan, are you able to get a smaller regression window? Perhaps using `hg bisect` and building locally?
Flags: needinfo?(rick.eyre)
Flags: needinfo?(bogdan.maris)
(In reply to Jared Wein [:jaws] (please needinfo? me) from comment #6) > Could also come from bug 865407 perhaps? Rick, does that make any sense to > you? > > Bogdan, are you able to get a smaller regression window? Perhaps using `hg > bisect` and building locally? I finally managed to do that, It was a struggle though since hg bisect would not want to bisect properly. I needed to improvise a bit but could not get a range though. All the builds I got were good, until I ran into bug 1235051 which was Win XP only. Apparently it does reproduce on Ubuntu as well. So bottom line is that I got nowhere with this...
Flags: needinfo?(bogdan.maris)
Jet, can you find an assignee?
Flags: needinfo?(bugs)
I couldn't reproduce this one. Twitter GIFs seem to behave the same on Firefox and Chrome. Do we have new repro steps?
Flags: needinfo?(rick.eyre)
Flags: needinfo?(bugs)
Flags: needinfo?(bogdan.maris)
I can't reproduce this either. It's possible that Twitter fixed this on their side. Lawrence, did you have some more information when you marked the affected versions?
Flags: needinfo?(lmandel)
I concur. Looks like Twitter has changed their UI. I cannot reproduce this. I had set the status flags to carry forward based on the information in the bug. I'm resolving as fixed. Bogdan - If you can still reproduce, please provide additional guidance via the STR.
Status: NEW → RESOLVED
Closed: 10 years ago
Flags: needinfo?(lmandel)
Resolution: --- → FIXED
(In reply to Lawrence Mandel [:lmandel] (use needinfo) from comment #11) > I had set the status flags to carry forward based on the information in the > bug. I'm resolving as fixed. Thanks, that's what I was thinking but just wanted to make sure.
I can't reproduce anymore across platforms. The UI has changed on twitter's side.
Flags: needinfo?(bogdan.maris)
Version: unspecified → 29 Branch
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: