Use default alias analysis GetObject semantics for some opcodes
Categories
(Core :: JavaScript Engine: JIT, task)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox75 | --- | fixed |
People
(Reporter: anba, Assigned: anba)
Details
Attachments
(4 files)
|
47 bytes,
text/x-phabricator-request
|
Details | Review | |
|
47 bytes,
text/x-phabricator-request
|
Details | Review | |
|
Bug 1618917 - Part 3: Fall through to the default case when getAliasSet() isn't overriden. r=jandem!
47 bytes,
text/x-phabricator-request
|
Details | Review | |
|
47 bytes,
text/x-phabricator-request
|
Details | Review |
Some potential clean-ups noticed while working on bug 1614704.
| Assignee | ||
Comment 1•6 years ago
|
||
| Assignee | ||
Comment 2•6 years ago
|
||
Depends on D64776
| Assignee | ||
Comment 3•6 years ago
|
||
These three opcodes don't override getAliasSet(), so it's better to let the
default case handle them, because then we can assert the alias information is
correct in debug builds.
Depends on D64777
| Assignee | ||
Comment 4•6 years ago
|
||
Similar to part 3, it seems better to get the debug check for MCompare for
free. Also this make MCompare more similar to other opcodes like MToString,
MBitNot, and MBinaryBitwiseInstruction, which all three either have a "none"
alias-set or the default "store-any" alias-set.
Depends on D64778
Comment 6•6 years ago
|
||
| bugherder | ||
https://hg.mozilla.org/mozilla-central/rev/40a351c7e4b1
https://hg.mozilla.org/mozilla-central/rev/3d690e518780
https://hg.mozilla.org/mozilla-central/rev/f83f504514b6
https://hg.mozilla.org/mozilla-central/rev/ce635c0f44ec
Description
•