Inspector doesn't show any rules
Categories
(DevTools :: about:debugging, defect, P2)
Tracking
(firefox-esr68 unaffected, firefox-esr78 unaffected, firefox80 unaffected, firefox81+ fixed, firefox82 unaffected)
Tracking | Status | |
---|---|---|
firefox-esr68 | --- | unaffected |
firefox-esr78 | --- | unaffected |
firefox80 | --- | unaffected |
firefox81 | + | fixed |
firefox82 | --- | unaffected |
People
(Reporter: karlcow, Assigned: jdescottes)
References
()
Details
(Keywords: regression)
Attachments
(6 files, 1 obsolete file)
Setup:
- Desktop: Firefox Nightly 81.0a1 (2020-08-18) (64-bit) macOS 10.15.6 (19G2021)
- Mobile: Firefox Nightly 81.0a1-20200813092915 Android 10. Pixel 2
- Mobile: Go to https://www.usedcarsni.com/cars (or any websites)
- Desktop: Go to about:debugging
- Open the remote debugger context
- Click on inspector
- select a node
Expected:
Have the CSS rules displayed in the center column
Actual:
All CSS rules are hidden.
Comment 1•4 years ago
|
||
Note that I'm seeing a lot of
Error: Protocol error (NS_ERROR_UNEXPECTED): Component returned failure code: 0x8000ffff (NS_ERROR_UNEXPECTED) [nsIPrefBranch.getCharPref] from: server0.conn1.child40/pagestyle26 (resource://devtools/server/actors/compatibility.js:104:0)
in the browser console.
Assignee | ||
Comment 2•4 years ago
|
||
Updated•4 years ago
|
Assignee | ||
Comment 3•4 years ago
|
||
I spotted one backward compatibility issue, but that's probably not what you are experiencing here. Mine also makes the markup view go blank. I will look into the compatibility actor report from comment 1.
Assignee | ||
Updated•4 years ago
|
Assignee | ||
Comment 4•4 years ago
|
||
Ok pref("devtools.inspector.compatibility.target-browsers", "");
should be moved to https://searchfox.org/mozilla-central/source/modules/libpref/init/all.js otherwise it will be undefined on Fenix
Assignee | ||
Comment 5•4 years ago
|
||
Depends on D87603
Assignee | ||
Comment 6•4 years ago
|
||
Depends on D87611
Assignee | ||
Comment 7•4 years ago
|
||
Depends on D87611
Updated•4 years ago
|
Pushed by jdescottes@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/be6efbcfec5d Add a setter for isTopLevelDocument on Node Front for backward compatibility r=daisuke https://hg.mozilla.org/integration/autoland/rev/c9645c284601 Move server-side devtools pref for Compatibility panel to modules/libpref/init/all.js r=daisuke https://hg.mozilla.org/integration/autoland/rev/103c9f61a101 Set a default value for target-browsers from the client r=daisuke
Comment 9•4 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/be6efbcfec5d
https://hg.mozilla.org/mozilla-central/rev/c9645c284601
https://hg.mozilla.org/mozilla-central/rev/103c9f61a101
Reporter | ||
Comment 10•4 years ago
|
||
Julian,
This is still busted for me as of today.
Is it the same for you?
macOS 10.15.6 (19G2021)
desktop Firefox Nightly 82.0a1 (2020-08-25) (64-bit)
Android 10
mobile Firefox Nightly 200825 06:07 (Build #2015759987)
81.0a1-20200820093209
just to stress that it is quite important for the daily debugging work on diagnosing issues for Firefox (for the webcompat team).
Thanks for any progress, and tell me how I can help you with better information.
Comment 11•4 years ago
|
||
Comment 12•4 years ago
|
||
[Tracking Requested - why for this release]: Remote debugging inspector is broken. It would nice to not ship this to release?
Updated•4 years ago
|
Updated•4 years ago
|
Updated•4 years ago
|
Assignee | ||
Comment 13•4 years ago
|
||
While testing this morning I spotted another regression caused by Bug 1660266, which I am fixing via Bug 1662058 on the server (will file another Bug to add a workaround on the client).
To workaround this, you can set a default value for the missing pref.
Run Services.prefs.setBoolPref("devtools.overflow.debugging.enabled", false)
in the console of the "Multiprocess toolbox" for Fenix.
After setting a default value, the inspector works for me.
Firefox desktop: latest central
Firefox mobile: Nightly 200829 06:05
Note that the Fenix build mentioned in the previous comments (81.0a1-20200820093209) does not include the fix that initially landed here, it is in the next one: https://ftp.mozilla.org/pub/firefox/nightly/2020/08/2020-08-20-21-21-07-mozilla-central/ . This means it didn't include the server side fix, which can explain why it didn't work for you? I am a bit surprised that the client side workaround in https://hg.mozilla.org/mozilla-central/rev/103c9f61a101 didn't properly create the preference.
Can you please test again?
Comment 14•4 years ago
|
||
OK, yeah. That did work for me:
- in Firefox Nightly, open about:debugging and connect to the latest Fenix Nightly build.
- Select the device on the left, then go to Processes at the bottom and click Inspect from the Multiprocess Toolbox
- Run
Services.prefs.setBoolPref("devtools.overflow.debugging.enabled", false)
in the console of the Multiprocess Toolbox (chrome://geckoview/content/geckoview.xhtml should be the selected frame/context/whatever, if not select it from the icon next to the meatballs icon) - Disconnect device from top of about:debugging (Disconnect button)
- Hit Connect again from about:debugging
At this point, the Inspector works as expected.
Reporter | ||
Comment 15•4 years ago
|
||
Mike Steps worked for me too with
Mobile Nightly 200831 06:01
and Desktop Nightly 82.0a1 (2020-08-30) (64-bit)
Assignee | ||
Comment 16•4 years ago
•
|
||
Thanks for testing!
The issue with devtools.overflow.debugging.enabled
is tracked by Bug 1662058, so I think we can close this.
I am also working on a more generic client side workaround with Bug 1662067.
Comment 17•4 years ago
•
|
||
Bug 1662058 was a new regression in 82, but this bug also affects 81? Is there something we can do to fix this on Beta still? Or are we saying that this bug was tracking two different problems - one fixed by the original patches and one subsequent regression fixed by bug 1662058?
Updated•4 years ago
|
Comment 18•4 years ago
|
||
Thank you very much Julian for helping to get this working again. :)
Assignee | ||
Comment 19•4 years ago
|
||
(In reply to Ryan VanderMeulen [:RyanVM] from comment #17)
Bug 1662058 was a new regression in 82, but this bug also affects 81? Is there something we can do to fix this on Beta still? Or are we saying that this bug was tracking two different problems - one fixed by the original patches and one subsequent regression fixed by bug 1662058?
Normally the initial fix that landed here in 81 (https://hg.mozilla.org/mozilla-central/rev/c9645c284601) is sufficient to fix the issue.
There is some delay before Fenix incorporates changes so when Mike & Karl tested again, Fenix 81 didn't have the fix yet. But the "final" Fenix 81 contains the fix and doesn't require any new patch.
I am trying to update the flags in the bug accordingly.
Description
•