Closed
Bug 1242273
Opened 8 years ago
Closed 8 years ago
[e10s] [APZ] Flash doesn't appear after I close newly opened New Tab page
Categories
(Core :: Graphics: Layers, defect, P2)
Tracking
()
RESOLVED
DUPLICATE
of bug 1229429
Tracking | Status | |
---|---|---|
e10s | + | --- |
firefox46 | --- | unaffected |
firefox47 | --- | unaffected |
firefox48 | --- | wontfix |
firefox49 | --- | wontfix |
firefox50 | --- | fix-optional |
firefox51 | --- | fix-optional |
firefox52 | --- | fix-optional |
People
(Reporter: arni2033, Unassigned)
References
()
Details
(Whiteboard: [gfx-noted][tpi:+][waiting on dependency])
Attachments
(1 file)
154.14 KB,
video/webm
|
Details |
>>> My Info: Win7_64, Nightly 46, 32bit, ID 20160121030208 STR: 1. Open new firefox window 2. Open many tabs in that window to cause overflow in Tabs toolbar. Open 5 more tabs 3. Open the following "data:" url in the last tab in that window > data:text/html,<iframe src="https://www.youtube.com/v/60og9gwKh1o" height="344" width="425"></iframe><style>body{height:10000px;}iframe{position:fixed;} 4. Make sure that flash appeared on the page. Otherwise reload the page. 5. Open new tab: Click on the free place on the page, Press Ctrl+T or click "New tab" button 6. Close the newly opened (last) tab: Press Ctrl+W or click "Close tab" button Result: Flash doesn't appear on the page Expectations: Flash should be on the page Regressed 2 times (sic!). Timescale: [0] Past ----------- [1] 1st regression ----------- [2] 2nd regression ----------- [3] Now [1] bug 1148855 (I believe): > https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=4d6d69f0f499dfc5c67df291858feda7bd2b9eac&tochange=d0fc7202b4cbe1ac8de823ed9e1e6dad706fc4d0 [2] bug 1157745, because enabling APZ in the last good build allows me to reproduce this bug: > https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=738d8ee106fe760c44f54b5f250d0b36a84818c8&tochange=af67b5f73381f154f41ded5926d823709a0ebd08 During [0]---[1] this bug never happened. During [1]---[2] this happened on lwthemes+HWA_on and Default theme. lwthemes+HWA_off were OK During [2]---[3] this happens on lwthemes and Default theme. lwtheme+HWA_off+APZ_off is OK Notes: 1) So now to avoid this issue I need to set lwtheme (e.g. Dev.Edition), disable APZ and HWA 2) I'm leaving keyword "regressionwindow-wanted" just in case I've mistaken in [1]
Comment 1•8 years ago
|
||
I can *even* reproduce the problem on the Bad build 4d6d69f0f499 w/ new profile. https://hg.mozilla.org/integration/mozilla-inbound/rev/4d6d69f0f499 Mozilla/5.0 (Windows NT 6.1; WOW64; rv:40.0) Gecko/20100101 Firefox/40.0 ID:20150402114335 And I can repro even if layers.async-pan-zoom.enabled = false on latest Nightly46.0a1. https://hg.mozilla.org/mozilla-central/rev/d6d81655dd9e146c300a64c0fcaeb04ca3300a19 Mozilla/5.0 (Windows NT 6.1; WOW64; rv:46.0) Gecko/20100101 Firefox/46.0 ID:20160124030209
(In reply to Alice0775 White from comment #1) > I can *even* reproduce the problem on the Bad build 4d6d69f0f499 w/ new profile. I think you meant "on the «Good» build 4d6d69f0f499 w/ new profile", because I described it as good. Well, of course it's possible to run into bug 1229429 if you open-close tab very quickly Is that what you did? In comment 0, I meant only "normal", not super-fast tab operations to exclude other bugs. So please confirm that it happens on build 4d6d69f0f499 with opening/closing tab "at normal speed"
Flags: needinfo?(alice0775)
> And I can repro even if layers.async-pan-zoom.enabled = false on latest Nightly46.0a1.
Oh, and also confirm that you tested this with lwtheme enabled, HWA disabled and APZ disabled at_the_same_time, as said in the description of time period [2]---[3].
Comment 4•8 years ago
|
||
I think this is normal speed. My STR: 1. Start firefox w/ new profile(w/ e10s) 2. Open many tabs in that window to cause overflow in Tabs toolbar. e.g Open 10 about:home tabs 3. Open URL 4. Make sure that flash appeared on the page. Otherwise reload the page. 5. Open new tab: Click "New tab" button 6. Wait about 5-10 sec 7. Close the newly opened (last) tab: Click "Close tab" button 8. Repeat from Step 5 if any My regression range: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=741a15659b09&tochange=5969eb0fe8b5 So, I think Bug 1095754 causes this.
Flags: needinfo?(alice0775)
Updated•8 years ago
|
tracking-e10s:
--- → ?
Updated•8 years ago
|
tracking-e10s:
? → ---
Updated•8 years ago
|
tracking-e10s:
--- → ?
Updated•8 years ago
|
Blocks: e10s-plugins
Flags: needinfo?(twalker)
Comment 5•8 years ago
|
||
The following STR’s (very similar to Alice’s) are 100% reproducible on Win7 and Ubuntu. For me, it is imperative that APZ is enabled and that there are enough tabs open to cause overflow of the tab toolbar even after closing the tab in step 8. For me, this bug is not reproducible if APZ is disabled. It is also not reproducible if the tab bar does not remain in tab overflow after step 8. Win 7 - Nightly 47.0a1, Build ID 20160127030236 - FAIL Ubuntu - Nightly 47.0a1, Build ID 20160128030208 - FAIL Mac 10.9.5 - Nightly 47.0a1, Build ID 20160128030208 - PASS STR: 0) Ensure APZ is on (layers.async-pan-zoom.enabled = true, set to true if it is not and restart the browser) 1) Start firefox w/ new profile(w/ e10s) 2) Open many tabs in that window to cause overflow in Tabs toolbar. The number of tabs required varies depending on your browser window width. 3) Open one more new tab. 4) In that tab Open the test url with flash content. 5) Make sure the the Flash content appears on the page. Otherwise reload the page. 6) Open new tab by clicking "New tab” button. (if you open new tab any other way, make sure you set focus to the new tab) 7) Wait a a few seconds, then 8) Close the newly opened (last) tab: Click "Close tab” button Tested results: In the Tab that regains focus, the flash content is not there. However, wait a few seconds (on my machine it is about 5 seconds), then the flash content appears. Expected results: A soon as the tab is closed and focus given back to the tab with the flash content, the flash content should be present.
Flags: needinfo?(twalker)
Updated•8 years ago
|
Blocks: apz-desktop
Comment 6•8 years ago
|
||
I can reproduce this on any windowed flash test case I have. Flipping tabs solves the problem, or the problem corrects itself after about two or three seconds. Feel like polish, I wouldn't block on this.
Updated•8 years ago
|
Comment 7•8 years ago
|
||
Attachment 8713926 [details] [diff] from bug 1243413 may fix this.
Component: Untriaged → Graphics: Layers
Depends on: 1243413
Comment 8•8 years ago
|
||
appears to be fixed.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → WORKSFORME
>>> My Info: Win7_64, Nightly 46, 32bit, ID 20160208030244 There're 3 issues in this bug (it seems), which have different regression ranges. Comment 4, which implies tab with Flash content to be first, looks fixed: Flash content appears on the page once animation in tabs toolbar is finished. Comment 0, which implies tab with Flash content to be the last (before Step 5) is not fixed
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
Reporter | ||
Comment 10•8 years ago
|
||
(In reply to arni2033 from comment #9) > Comment 4, which implies tab with Flash content to be first, looks fixed: > Flash content appears on the page once animation in tabs toolbar is finished. Or is it fixed? IMO Flash content should appear simultaneously with the rest of content. As if I just switched from one visible tab to another visible tab. But of course I don't know exactly what is "expected" here (e10s+Flash behavior is really strange)
Comment 11•8 years ago
|
||
I can still reproduce the problem on Aurora46.0a2 and Nightly47.0a1 with clean profile. https://hg.mozilla.org/releases/mozilla-aurora/rev/f19a31a82ae36283b82d36a0cb2f4264a8356787 Mozilla/5.0 (Windows NT 6.1; WOW64; rv:46.0) Gecko/20100101 Firefox/46.0 ID:20160208004010 https://hg.mozilla.org/mozilla-central/rev/ac338559876df7b2e81388f2aac28d2e95ceb5ff Mozilla/5.0 (Windows NT 6.1; WOW64; rv:47.0) Gecko/20100101 Firefox/47.0 ID:20160208030244
Updated•8 years ago
|
status-firefox47:
--- → affected
Comment 12•8 years ago
|
||
The str I'm following: 1) open new tab pages until tab overflow kicks in 2) open data:text/html,<iframe%20src="https://www.youtube.com/v/60og9gwKh1o"%20height="344"%20width="425"></iframe><style>body{height:10000px;}iframe{position:fixed;} 3) open one more tab 4) wait about three seconds 4) close new tab page It helps if you don't use or move your mouse around after step 4 (use keyboard shortcuts). result: youtube video window takes a few seconds to reappear. flipping between tabs or any remove content interaction in the video page will bring it back sooner.
Updated•8 years ago
|
Keywords: regressionwindow-wanted
Comment 13•8 years ago
|
||
Should this be blocking an apz meta bug? Since it is a regression that only shows up with apz enabled I want to make sure your team is aware.
Flags: needinfo?(bugmail.mozilla)
Comment 14•8 years ago
|
||
(In reply to Liz Henry (:lizzard) (needinfo? me) from comment #13) > Should this be blocking an apz meta bug? Since it is a regression that only > shows up with apz enabled I want to make sure your team is aware. Yup, it's blocking the apz-desktop meta bug already.
Flags: needinfo?(bugmail.mozilla)
Updated•8 years ago
|
Flags: needinfo?(jmathies)
Comment 15•8 years ago
|
||
Still reproducible. This is basically bug 1229429 with different str.
Depends on: 1229429
Flags: needinfo?(jmathies)
Updated•8 years ago
|
Priority: -- → P1
Comment 16•8 years ago
|
||
This is e10s specific, which won't roll out with 46.
Updated•8 years ago
|
Whiteboard: [gfx-noted]
Updated•8 years ago
|
Flags: needinfo?(howareyou322)
Updated•8 years ago
|
status-firefox48:
--- → affected
status-firefox49:
--- → affected
Updated•8 years ago
|
Version: Trunk → 48 Branch
Given comment 15, should bug 1229429 be tracked as a 48 regression?
Dependency on bug 1229429 based on comment 15.
Keeping the timing in step with bug 1229429.
Updated•8 years ago
|
Priority: P1 → P2
Updated•8 years ago
|
OS: Unspecified → Windows
Hardware: Unspecified → All
Whiteboard: [gfx-noted] → [gfx-noted][tpi:+]
Updated•8 years ago
|
Whiteboard: [gfx-noted][tpi:+] → [gfx-noted][tpi:+][e10st?]
Updated•8 years ago
|
Whiteboard: [gfx-noted][tpi:+][e10st?] → [gfx-noted][tpi:+]
Updated•8 years ago
|
Flags: needinfo?(howareyou322)
Comment 20•8 years ago
|
||
Adding a whiteboard thing to note that this bug is not actionable, it's just waiting on bug 1229429. We might want to consider duping it over, at the risk of annoying arni :)
Whiteboard: [gfx-noted][tpi:+] → [gfx-noted][tpi:+][waiting on dependency]
Updated•8 years ago
|
status-firefox51:
--- → affected
status-firefox52:
--- → affected
Comment 21•8 years ago
|
||
Seems like bug 1229429 is the root problem here. If anyone disagrees, feel free to un-dupe.
Status: REOPENED → RESOLVED
Closed: 8 years ago → 8 years ago
Resolution: --- → DUPLICATE
Comment 22•8 years ago
|
||
(Removing regression keyword since it's on the bug this is duped to. Please re-add if we un-dupe.)
Updated•8 years ago
|
Updated•8 years ago
|
Keywords: regression
You need to log in
before you can comment on or make changes to this bug.
Description
•