Consider switching the layout debugger to use e10s
Categories
(Core :: Layout, task)
Tracking
()
Tracking | Status | |
---|---|---|
firefox70 | --- | fixed |
People
(Reporter: Gijs, Assigned: heycam)
References
(Regressed 1 open bug)
Details
Attachments
(6 files)
47 bytes,
text/x-phabricator-request
|
Details | Review | |
47 bytes,
text/x-phabricator-request
|
Details | Review | |
47 bytes,
text/x-phabricator-request
|
Details | Review | |
47 bytes,
text/x-phabricator-request
|
Details | Review | |
47 bytes,
text/x-phabricator-request
|
Details | Review | |
47 bytes,
text/x-phabricator-request
|
Details | Review |
Right now the layout debugger loads resources in a non-e10s <browser>
.
It'd be safer, but also probably less convenient in some ways to have it load things in a remote browser instead. bug 1565965 comment #3 suggests we would ideally switch to using e10s. OTOH, based on https://groups.google.com/d/msg/mozilla.dev.platform/cJMzxi7_PmI/NfM6D3OPBQAJ , it seems some folks prefer having something single-processed to help with debugging.
I don't know to what degree switching to a single-child-process (so e10s-enabled) layout debugger would hamper debugging - I suppose it's somewhat less annoying than "real" Firefox but probably more annoying than the current non-e10s state.
I don't really have a horse in this race, but it seemed worth getting a bug on file seeing as I couldn't find any.
I think having e10s is more convenient when debugging in rr, because then all the stuff you care about is in its own process!
Assignee | ||
Comment 2•6 years ago
|
||
Adding remote="true"
to the <browser>
element is enough to make it use a content process, but all of the nsLayoutDebuggingTools
methods that need to poke at the docshell, e.g. for frame tree dumping, need another mechanism to ask the content process to do the work.
Assignee | ||
Comment 3•6 years ago
|
||
Assignee | ||
Comment 4•6 years ago
|
||
Assignee | ||
Comment 5•6 years ago
|
||
Assignee | ||
Comment 6•6 years ago
|
||
Many of these features are non-functional these days, but I'll keep them
hooked up in case we decide to fix them.
Assignee | ||
Comment 7•6 years ago
|
||
Assignee | ||
Comment 8•6 years ago
|
||
Assignee | ||
Comment 9•6 years ago
|
||
Assignee | ||
Updated•6 years ago
|
Assignee | ||
Comment 10•6 years ago
|
||
One thing I only just noticed after uploading the patches is that I can no longer view file: URLs with these changes, unless I turn off sandboxing. I think that the file: URL is being loaded in a regular Web content process, and not a file: content process. Also when I load say about:config, that also gets loaded in the content process.
Gijs, do you know what I might need to do beyond using remote="true"
on the <browser>
to cause URLs to be loaded in the expected process?
Assignee | ||
Comment 11•6 years ago
|
||
I think maybe I need to do some things similar to _loadURI
in browser.js.
Assignee | ||
Comment 12•6 years ago
|
||
I've worked it out.
Assignee | ||
Updated•6 years ago
|
Comment 13•6 years ago
|
||
Comment 14•6 years ago
|
||
Comment 15•6 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/32f959e898c6
https://hg.mozilla.org/mozilla-central/rev/184025bba151
https://hg.mozilla.org/mozilla-central/rev/f84c119572a1
https://hg.mozilla.org/mozilla-central/rev/ffbc0f9c06c4
https://hg.mozilla.org/mozilla-central/rev/5d1cf5706da5
https://hg.mozilla.org/mozilla-central/rev/5b7f4799a1a8
https://hg.mozilla.org/mozilla-central/rev/db8f3ee41bdf
Description
•