1638490 - Rendering issue, some content stuck from previously open tab

Status: NEW

(Core :: Graphics: WebRender, defect)

Description

[ospis12]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:76.0) Gecko/20100101 Firefox/76.0

Steps to reproduce:

1. Open and log in to floatplane.com
2. Wait for page to load
3. Open second tab and go to pclab.pl
4. Wait for page to load
5. Go back to floatplane.com
6. Profit(Bug is showing)
   Almost 100% reproducible, does happen in nightly.
   Website devs couldn't find issue on their side.

Actual results:

Image of issue https://imgur.com/a/cO3UJKh .
Part of previews tab is still being rendered in active tab.

Expected results:

I would expect to see only content of website I'm currently using.

Comment 1

[Andrei Purice]

Hey Ospis12,

I tried reproducing this issue on the latest Nightly 78.0a1 (2020-05-20), beta 77.0b8 and release 76.0.1 versions without any success though.

Can you test the issue while in Safe Mode. You can find helpful info here : https://support.mozilla.org/en-US/kb/troubleshoot-firefox-issues-using-safe-mode .

Also a fresh new profile could help. You can find more about creating a new profile here : https://support.mozilla.org/en-US/kb/troubleshoot-and-diagnose-firefox-problems#w_6-create-a-new-firefox-profile .

Comment 2

[ospis12]

In response to questions asked, I was able to reproduce it with new profile and no addons in nightly 78.0a1 (220-05-20), but not consistently. I had to switch tabs when pclab was loading, and there was no ads loaded.
In Safe Mode I was not able to reproduce bug.
With uBlock Origin I can reproduce every time.

Comment 3

[ospis12]

Did some more testing, turning off hardware acceleration but remaining in normal mode did not change behavior of bug. I was still able to reproduce it.
If you think it will be helpful I will post my hardware/software configuration.

Comment 4

[Andrei Purice]

Setting a component for this issue in order to get the dev team involved.
If you feel it's an incorrect one please feel free to change it to a more appropriate one.

Comment 5

[Tomislav Jovanovic :zombie]

Andrei, were you able to reproduce this? Can you confirm a more specific STR?

Comment 6

[Andrei Purice]

Hey Tomi,

I was able to reproduce this issue using the steps provided by the author on the latest Nightly 79.0a1 (2020-06-25). I still couldn't repro it on the release version 77.0.1 or beta 78.0 . As an extra step when i reached floatplane.com i also chose one of the creators (i.e.: Robert Neal) and i could repro it on his page. I have attached a file.

Comment 7

[Andrei Purice]

Attachment: float.jpg

Comment 8

[Tomislav Jovanovic :zombie]

Thanks Andrei.

Does this reproduce only with an extension (like the mentioned uBO), or even without one? Can you please find a regression window?

Comment 9

[Andrei Purice]

Hello Tomislav,
I can reproduce this bug without any extension installed and I am able to reproduce it on my personal pc (configuration : Nvidia Gtx 1660, Intel i5 ) but not on my work pc (configuration : Ati Radeon 3000 Graphics, Amd fx(tm)-83320 Eight-Core Processor ). Let me know if I can further assisst you on this matter.

Here is the regression range I found.

Last known good build : https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=810e31bf46f160c31e8cfa8d77aead4c902f2276&tochange=6814870342efc284b2edec7c9aeb45240cb0fe09

First bad build : https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=31b56def2a8db104c791ba220e3eea5664dd906a&tochange=6814870342efc284b2edec7c9aeb45240cb0fe09

Found commit message:
Bug 1579235 - Part 6 - Support an opaque/alpha native surface per slice. r=Bert

Comment 10

[Tomislav Jovanovic :zombie]

Thanks again Andrei. Since it's reproducible without any extensions, this looks like a webrender regression, moving there.

Comment 11

[Glenn Watson [:gw]]

I haven't been able to reproduce locally, but it's likely that it depends on the current compositing mode (e.g. DirectComposition) and/or driver version. Could you paste the output of your about:support here?

Comment 12

[ospis12]

Attachment: about:support

Since my first posting, I've updated my graphics drivers. Now I cannot get this bug to reproduce.
Here is my <code> about:support </code>
<code>

</code>

Comment 13

[Andrew Osmond [:aosmond] (he/him)]

ospis12, do you know what driver version you had before? Thanks!

Comment 14

[ospis12]

I'm sorry, but I don't recall version I was using. I tried to test with version ranging from 441.87 to 442.74 and got nothing. So right now I cannot reproduce this bug.

Comment 15

[Jim Mathies [:jimm]]

unknown driver version issue, no str.

Comment 16

[BugBot [:suhaib / :marco/ :calixte]]

The bug has a release status flag that shows some version of Firefox is affected, thus it will be considered confirmed.