Closed Bug 1267885 Opened 9 years ago Closed 8 years ago

Text appears to jump / flicker a few seconds after scrolling ends (non-hidpi monitor only)

Categories

(Core :: Graphics: Text, defect)

48 Branch
defect
Not set
normal

Tracking

()

RESOLVED DUPLICATE of bug 1012752
Tracking Status
firefox46 --- unaffected
firefox47 --- unaffected
firefox48 --- wontfix

People

(Reporter: jryans, Unassigned)

References

()

Details

(Keywords: qawanted, regression, Whiteboard: [gfx-noted])

Attachments

(1 file)

Attached image jumping-text.gif (deleted) —
Nightly 2016-04-26 STR: 1. Go to a page like https://bugzilla.mozilla.org/page.cgi?id=splinter.html&bug=920956&attachment=8745277 2. Click on browser_responsiveui_touch.js so that a diff is on screen 3. Scroll around the page using for example page down / page up keys ER: Text should scroll naturally, like it used to until a while ago. (I am not sure how long I've noticed issues, perhaps around a month.) AR: Once the page stops scrolling, some of the text seems to jump vertically by 1 - 2 px. I have attached a screen capture of the issue.
I am on OS X 10.11.4.
OS: Unspecified → Mac OS X
I see correct behavior if I view the page on the Retina / HiDPI screen built into my MacBook Pro Retina 15. The issue only occurs on my external non-Retina / 1x display.
Summary: Text appears to jump / flicker a few seconds after scrolling ends → Text appears to jump / flicker a few seconds after scrolling ends (non-hidpi monitor only)
:mstange, does this issue sound the same as bug 1220371? My issue happens with regular web content though, not DevTools.
Flags: needinfo?(mstange)
This is "intentional" but we're planning to fix it in bug 1012752 because it hurts performance and looks really bad (though bug 1012752 was primarily filed with performance in mind). Bug 1220371 is something else - that one's about the blending of the text anti-aliasing, and this one's about the vertical position of the text.
Flags: needinfo?(mstange)
Thanks! I'll mark bug 1012752 as a dependency then.
Depends on: 1012752
Whiteboard: [gfx-noted]
Marking as a regression and setting 48+ as affected as per dup bug 1272408.
Version: unspecified → 48 Branch
Now that bug 1012752 has landed, can somebody verify this is gone?
Keywords: qawanted
OS: Mac OS X → Unspecified
Ryan, do you still see this in nightly?
Flags: needinfo?(jryans)
(In reply to Milan Sreckovic [:milan] from comment #10) > Ryan, do you still see this in nightly? At the moment, it does seem resolved, which is great! However, looking through bug 1012752, it seems :mstange tried to back it out for causing regressions, but then it was re-landed by a sheriff because the backout caused conflicts. Will bug 1012752 be backed out again?
Flags: needinfo?(jryans) → needinfo?(mstange)
Yes. Maybe later today.
Flags: needinfo?(mstange)
Bug 1012752 was re-landed On August 5th and seems to have stuck. Ryan, can you confirm this is fixed now?
Flags: needinfo?(jryans)
(In reply to Kartikaya Gupta (email:kats@mozilla.com) from comment #13) > Bug 1012752 was re-landed On August 5th and seems to have stuck. Ryan, can > you confirm this is fixed now? Testing from Nightly 20160818, yes, it looks like this issue is fixed, thanks! Not sure how you'd like the bug flags updated, so I'll let someone else update them. :)
Flags: needinfo?(jryans)
Thanks! I'm just going to dupe it over.
Status: NEW → RESOLVED
Closed: 8 years ago
No longer depends on: 1012752
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: