Consider removing the Debugger's instrumentation mechanism
Categories
(Core :: JavaScript Engine, task, P3)
Tracking
()
Tracking | Status | |
---|---|---|
firefox91 | --- | fixed |
People
(Reporter: jandem, Assigned: iain)
References
(Blocks 2 open bugs)
Details
(Whiteboard: [smooshmonkey-reserve])
Attachments
(6 files)
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review | |
48 bytes,
text/x-phabricator-request
|
Details | Review |
This was added in bug 1554524 for Web Replay but with Web Replay code gone from the tree it might make sense to remove this too. This way SmooshMonkey's bytecode emitter doesn't need to implement this.
Comment 1•5 years ago
|
||
It is unfortunate, but I agree we should remove. We don't have any use cases of this API on roadmaps. The design that was implemented was an interesting balance of concerns, but we don't want to maintain compatibility of instrumentation points inside new frontend.
Updated•5 years ago
|
Updated•5 years ago
|
Updated•4 years ago
|
Assignee | ||
Comment 2•3 years ago
|
||
I bumped into an issue with this code while working on bug 1672626 (JSOp::Resume
in the baseline interpreter does not initialize BaselineFrame::interpreterICEntry_
because the first op after resuming will always be a jump target that will set it, unless the realm instrumentation code has inserted InstrumentationActive/JumpIfFalse/...
, in which case we use a garbage IC entry in JumpIfFalse
; it hasn't been a problem until now because of the fast-path in JumpIfFalse
that skips the IC if the input is already a bool, but I'm removing that fast-path.)
It's as good a time as any to clean this up.
Assignee | ||
Comment 3•3 years ago
|
||
This is no longer needed because we are removing the realm instrumentation code (which was only ever used for WebRR).
Assignee | ||
Comment 4•3 years ago
|
||
This patch stack is based on the patches for bug 1554524 in more-or-less reverse order, with a few modifications so that the intermediate patches build. This is part 8.
Depends on D116304
Assignee | ||
Comment 5•3 years ago
|
||
This removes the changes from bug 1554524 part 6 (and equivalent changes that were made later for Stencil).
Depends on D116305
Assignee | ||
Comment 6•3 years ago
|
||
This removes the changes from bug 1554524 part 5.
Depends on D116306
Assignee | ||
Comment 7•3 years ago
|
||
This removes bug 1554524 Part 7 and the relevant parts of bug 1618198 part 26.
Depends on D116307
Assignee | ||
Comment 8•3 years ago
|
||
This patch removes the changes from bug 1554524 part 4. (Parts 1 and 3 of bug 1554524 were bytecode emitter cleanups that don't need to be reverted; part 2 was abandoned.)
Depends on D116308
Comment 10•3 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/56f9a8184a88
https://hg.mozilla.org/mozilla-central/rev/27cea02bdee4
https://hg.mozilla.org/mozilla-central/rev/8b94a17c1934
https://hg.mozilla.org/mozilla-central/rev/202e208f88b5
https://hg.mozilla.org/mozilla-central/rev/135a0e6af641
https://hg.mozilla.org/mozilla-central/rev/e590eb34efc5
Description
•