Closed Bug 1307711 Opened 9 years ago Closed 8 years ago

New RDM breaks a current page when enabling RDM repeatedly

Categories

(DevTools :: Responsive Design Mode, defect, P3)

All
Windows
defect

Tracking

(firefox49 unaffected, firefox50 unaffected, firefox51 unaffected, firefox52 wontfix)

RESOLVED DUPLICATE of bug 1407830
Tracking Status
firefox49 --- unaffected
firefox50 --- unaffected
firefox51 --- unaffected
firefox52 --- wontfix

People

(Reporter: magicp.jp, Unassigned)

Details

(Whiteboard: [multiviewport][reserve-rdm])

Attachments

(3 files)

Attached video new-rdm.mp4
User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Firefox/52.0 Build ID: 20161004030204 Steps to reproduce: 1. Start Nightly 2. Go to any sites (e.g. "Getting Started" https://www.mozilla.org/en-US/firefox/central/) 3. Open DevTools 4. On/Off RDM button several times Actual results: Current page turns blank and Exit Firefox is slower. The following errors are in Console. this.webNavigation is null browser.xml:876 TypeError: win.gBrowser is undefined[Learn More] ReaderParent.jsm:85:1 TypeError: cannot use the given object as a weak map key RemoteAddonsParent.jsm:1084:7 Got a removal of an actor we didn't know about: server1.conn0.child1/domnode43 inspector.js:779 Got an addition of an actor we didn't know about: server1.conn0.child1/domnode51 inspector.js:794 Got a removal of an actor we didn't know about: server1.conn0.child1/domnode51 inspector.js:779 Got an addition of an actor we didn't know about: server1.conn0.child1/domnode52 inspector.js:794 TypeError: this.globals.get(...) is undefined ExtensionContent.jsm:883:5 this.webNavigation is null browser.xml:876 TypeError: win.gBrowser is undefined[Learn More] ReaderParent.jsm:85:1 TypeError: cannot use the given object as a weak map key RemoteAddonsParent.jsm:1084:7 Got a removal of an actor we didn't know about: server1.conn0.child1/domnode52 inspector.js:779 Got an addition of an actor we didn't know about: server1.conn0.child1/domnode53 inspector.js:794 Got a removal of an actor we didn't know about: server1.conn0.child1/domnode53 inspector.js:779 Got an addition of an actor we didn't know about: server1.conn0.child1/domnode54 inspector.js:794 TypeError: this.globals.get(...) is undefined ExtensionContent.jsm:883:5 this.webNavigation is null browser.xml:876 TypeError: win.gBrowser is undefined[Learn More] ReaderParent.jsm:85:1 TypeError: cannot use the given object as a weak map key RemoteAddonsParent.jsm:1084:7 Got a removal of an actor we didn't know about: server1.conn0.child1/domnode54 inspector.js:779 Got an addition of an actor we didn't know about: server1.conn0.child1/domnode55 inspector.js:794 Got a removal of an actor we didn't know about: server1.conn0.child1/domnode55 inspector.js:779 Got an addition of an actor we didn't know about: server1.conn0.child1/domnode56 inspector.js:794 TypeError: this.globals.get(...) is undefined ExtensionContent.jsm:883:5 this.webNavigation is null browser.xml:876 TypeError: cannot use the given object as a weak map key RemoteAddonsParent.jsm:1084:7 TypeError: this.globals.get(...) is undefined ExtensionContent.jsm:883:5 Expected results: New RDM works correctly.
Has Regression Range: --- → yes
Has STR: --- → yes
Component: Untriaged → Developer Tools: Responsive Design Mode
OS: Unspecified → Windows 10
Hardware: Unspecified → x86_64
Tracking to be sure we look into this before RDM goes to all channels. However, in a quick test locally, I couldn't actually reproduce. I tried clicking the RDM button as rapidly as possible, and it appeared to toggle correctly each team. :magicp, can you think of any additional steps or specific customizations you might have enabled? Do you see the same behavior in a new profile?
Blocks: enable-rdm
Flags: needinfo?(magicp.jp)
Priority: -- → P3
Whiteboard: [multiviewport][reserve-rdm]
(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #1) > However, in a quick test locally, I couldn't actually reproduce. I tried > clicking the RDM button as rapidly as possible, and it appeared to toggle > correctly each team. I meant to say "it appeared to toggle correctly each _time_".
(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #1) > :magicp, can you think of any additional steps or specific customizations > you might have enabled? Do you see the same behavior in a new profile? I can reproduce this in a new profile. It is occurred after toggling for two or twenty times.
Flags: needinfo?(magicp.jp)
Confirmed on Windows 7 and Windows 8.1
OS: Windows 10 → Windows
Hardware: x86_64 → All
This time I tried again on a Windows 10 VM (previous test was on macOS 10.11), but I still can't reproduce. I tried various different variations, but I did not manage to get it stuck. Please let me know if you think of any additional details, I'd really like to figure this out!
mozregression str: 1. Start Nightly > about:home 2. Repeat Ctrl+Shift+M
(In reply to magicp from comment #7) > mozregression str: > 1. Start Nightly > about:home > 2. Repeat Ctrl+Shift+M I've tried these steps as well, but I am still unable to reproduce so far.
(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #8) > I've tried these steps as well, but I am still unable to reproduce so far. Can you see errors of comment #1 in console?
Flags: needinfo?(jryans)
(In reply to magicp from comment #9) > (In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #8) > > > I've tried these steps as well, but I am still unable to reproduce so far. > > Can you see errors of comment #1 in console? I can see most, but not all. I don't see the "Got [x] of an actor we didn't know about" errors. The other errors (that I do see here) are noise that happens with RDM at the moment. There are various bugs to clean them up, but they aren't a problem. So far, I am guessing this a race condition somewhere, but I'll need some way to trigger it locally first.
Flags: needinfo?(jryans)
I have confirmed regression range again using latest mozregression-gui v0.9.4. This regression range is same with the following errors... Is there a possibility of this bug? this.webNavigation is null browser.xml:880 TypeError: win.gBrowser is undefined ReaderParent.jsm:85:1 TypeError: cannot use the given object as a weak map key RemoteAddonsParent.jsm:1085:7 Regression range: https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=f6c5daab390c0772cd67dc624359e50ba539c6b4&tochange=e4e13b10667c36eeb0d24ad09f37851504977208 Bug 1305376 - Use nsTArray for cursor image list storage. r=xidorn
Flags: needinfo?(jryans)
> this.webNavigation is null browser.xml:880 > TypeError: win.gBrowser is undefined ReaderParent.jsm:85:1 > TypeError: cannot use the given object as a weak map key RemoteAddonsParent.jsm:1085:7 FYI, these errors occur on Mac OS X, Linux and Windows, but whiteout is only on Windows in my desktop and VM.
(In reply to magicp from comment #11) > I have confirmed regression range again using latest mozregression-gui > v0.9.4. This regression range is same with the following errors... Is there > a possibility of this bug? > > this.webNavigation is null browser.xml:880 > TypeError: win.gBrowser is undefined ReaderParent.jsm:85:1 > TypeError: cannot use the given object as a weak map key > RemoteAddonsParent.jsm:1085:7 These errors are noise that happens with all RDM usage currently. Bug 1273721 tracks fixing them. So, those shouldn't be related to the issue here. It's possible other errors could be related, if they appear. > Regression range: > https://hg.mozilla.org/integration/autoland/ > pushloghtml?fromchange=f6c5daab390c0772cd67dc624359e50ba539c6b4&tochange=e4e1 > 3b10667c36eeb0d24ad09f37851504977208 > > Bug 1305376 - Use nsTArray for cursor image list storage. r=xidorn Bug 1305376 is not the culprit here. It just so happens that the changeset just prior to that bug is for bug 1296498 where we first enabled the new RDM redesign in Nightly only. My current theory is that this bug is might be same as bug 1309631, but I don't know for sure yet. For me, the other bug only reproduces in a release build (not a local build), and I am guessing the same is true here since the symptoms are similar. I'll continue investigating.
Flags: needinfo?(jryans)
"TypeError: this.globals.get" is repeated when failing rdm.
Mass wontfix for bugs affecting firefox 52.
I think this should now be fixed by bug 1407830. magicp, could you update to Nightly 20171017100127 or later and re-test?
Flags: needinfo?(magicp.jp)
This bug fix (bug 1407830) was verified in the latest Nightly build 20171017141229. Thanks a lot! - in Nightly Build 20171017100113 reproduce - in Nightly Build 20171017141229 not reproduce
Flags: needinfo?(magicp.jp)
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Status: RESOLVED → VERIFIED
Wonderful, thank you for filing and for your patience, I know it took a while to get this fixed. I'll track this as a duplicate of the bug with the fix.
Status: VERIFIED → RESOLVED
Closed: 8 years ago8 years ago
Resolution: FIXED → DUPLICATE
Product: Firefox → DevTools
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: