1755266 - Inspector fails on framesets

Status: RESOLVED FIXED

(DevTools :: Inspector, defect, P2)

Description

[David]

Attachment: img1.png

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:97.0) Gecko/20100101 Firefox/97.0

Steps to reproduce:

1. Go to https://aprende-web.net/html/ejemplos/ejemplo34.html
2. rigth click and select "Inspect"

Actual results:

The node tree is closed

Expected results:

The node tree should be open

Comment 1

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

The Bugbug bot thinks this bug should belong to the 'DevTools::Inspector' component, and is moving the bug to that component. Please revert this change in case you think the bot is wrong.

Comment 2

[Julian Descottes [:jdescottes]]

Thanks for filing, seems to work fine on ESR, would be great to have a regression window.

Comment 4

[Alice0775 White]

regression window:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=853874dee29b2a068a74b2a550151afaffd939ce&tochange=fe65ddb40037dcbd40fa07de66c7cbc619b55e12

Comment 5

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

:alice0775, since this bug is a regression, could you fill (if possible) the regressed_by field?
For more information, please visit auto_nag documentation.

Comment 6

[Alice0775 White]

Suspect Bug 1605330, Bug 1742332, 1741808, Bug 1742541, tentatively, regressed by Bug 1741808

Comment 7

[Julian Descottes [:jdescottes]]

Thanks for the investigation. This would suggest that the regression is linked to the prefs devtools.webconsole.input.context or devtools.every-frame-target.enabled. When testing previously it didn't seem to be the case, but maybe one of the values gets somehow cached, so I'll test again with clean profiles.

Comment 8

[Julian Descottes [:jdescottes]]

Indeed, the regression is linked to devtools.every-frame-target.enabled, it seems that if you already started devtools once, changing the value in about:config afterwards has no impact. I guess this is because we cache the value at JSM scope in some server files eg https://searchfox.org/mozilla-central/rev/68a5327697ec43aa55b458c504c4b313c9c80528/devtools/server/connectors/js-window-actor/DevToolsFrameChild.jsm#29

So I can confirm this was regressed by Bug 1741808

Edit: and it doesn't seem impacted by devtools.webconsole.input.context at all.

Comment 9

[Markus Keller]

(In reply to Julian Descottes [:jdescottes] from comment #8)
Thank you for digging out that option!
Setting devtools.every-frame-target.enabled to false in about:config and then restarting Firefox made the Inspect command work again in Firefox 97.0.1.

Comment 10

[Laurent Barbareau]

(In reply to Markus Keller from comment #9)

Setting devtools.every-frame-target.enabled to false in about:config and then restarting Firefox made the Inspect command work again in Firefox 97.0.1.
Thx !

Comment 11

[Nicolas Chevobbe [:nchevobbe]]

The problem comes from devtools/shared/layout/utils.js#886,892
We explicitely check if the node is an <iframe>, so we miss the case where we have a <frame>.
I'll fix that

Comment 12

[Nicolas Chevobbe [:nchevobbe]]

Attachment: Bug 1755266 - [devtools] Fix isFrameWithChildTarget for frame element. r=jdescottes.

We do create dedicated targets for frame documents, but
in isFrameWithChildTarget, we were only checking if the
passed element was an iframe, making some area of the code
not behaving correctly (e.g. using the node picker, or
the Inspect Element context menu entry).
This patch fixes this and add a test case to make sure we
don't regress.

Comment 13

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/d26db7cd35a3
[devtools] Fix isFrameWithChildTarget for frame element. r=jdescottes.

Comment 14

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/d26db7cd35a3