Closed Bug 1437374 Opened 8 years ago Closed 8 years ago

Scrolling on a drop-down menu shows blank areas with retain-dl

Categories

(Core :: Web Painting, defect, P2)

defect

Tracking

()

RESOLVED FIXED
mozilla60
Tracking Status
firefox-esr52 --- unaffected
firefox58 --- unaffected
firefox59 --- unaffected
firefox60 --- fixed

People

(Reporter: mayankleoboy1, Assigned: mattwoodrow)

References

()

Details

(Keywords: regression)

Attachments

(2 files)

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:60.0) Gecko/20100101 Firefox/60.0 Build ID: 20180211100408 Steps to reproduce: 1. Create fresh nightly profile 2. Go to https://www.spectra.co/broadband-plans-in-pune 3. choose "Select your area" as "Wakad" from the dropdown 4. In the "select your society" dropdown, drag the vertical scrollbar up and down. Actual results: parts of the list disappears for a few seconds or flashes Expected results: not so Mozregresion gave me: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=a20dc29e502d8b74a1fca9195154d895d7de9ec0&tochange=3403b5f669d1395097f26c69c514e52d302a3ec0 2018-02-11T19:13:57: DEBUG : Using url: https://hg.mozilla.org/integration/mozilla-inbound/json-pushes?changeset=3403b5f669d1395097f26c69c514e52d302a3ec0&full=1 2018-02-11T19:13:59: DEBUG : Found commit message: Bug 1436189 - Part 2: Make sure we descend to the modified frame so that we descend through the containing block and mark out of flow frames for display. r=miko 2018-02-11T19:13:59: INFO : The bisection is done. 2018-02-11T19:13:59: INFO : Stopped If I disable retain-dl, then this does not reproduce.
Blocks: 1436189
Flags: needinfo?(matt.woodrow)
I can repo this, setting to NEW
Status: UNCONFIRMED → NEW
Ever confirmed: true
Assignee: nobody → matt.woodrow
Blocks: 1429932
No longer blocks: 1436189
Flags: needinfo?(matt.woodrow)
Priority: -- → P2
If FrameHasDisplayport == true, then the current frame is the outer scroll frame, and the overflow rect will be clipped to the scrollport. The branch below this code handles clipping the dirty rect to the displayport area, so we don't want to clip to the overflow. Reftest does an invalidation outside of the scrollport, but within the displayport to check that we correctly handle invalidations in that area. I needed reftest-snapshot-all since the code that figures out which part of the canvas to snapshot runs on the client side, and doesn't take the async scroll position into account.
Attachment #8950123 - Flags: review?(mikokm)
Comment on attachment 8950123 [details] [diff] [review] Don't intersect with the overflow rect if we're taking the displayport into account Review of attachment 8950123 [details] [diff] [review]: ----------------------------------------------------------------- LGTM.
Attachment #8950123 - Flags: review?(mikokm) → review+
Pushed by mwoodrow@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/7953b08f4886 Don't intersect the rebuild rect with the frame's overflow area if we're about to take the displayport into account. r=miko
Pushed by mwoodrow@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/b8eeddf07480 Don't intersect the rebuild rect with the frame's overflow area if we're about to take the displayport into account. r=miko
Flags: needinfo?(matt.woodrow)
Pushed by mwoodrow@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/8a15c2c921d8 Don't intersect the rebuild rect with the frame's overflow area if we're about to take the displayport into account. r=miko
Keywords: regression
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla60
Symptoms described at bug 1438621 seems to be fixed after this has landed.
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: