Closed
Bug 585750
Opened 15 years ago
Closed 15 years ago
JM: Fix PIC Resetting
Categories
(Core :: JavaScript Engine, defect)
Tracking
()
RESOLVED
FIXED
People
(Reporter: sstangl, Unassigned)
Details
Attachments
(1 file)
4.58 KB,
patch
|
dvander
:
review+
|
Details | Diff | Splinter Review |
PIC resetting is not well-tested, leading to problems when gczeal(2) is inserted by the fuzzer. One such testcase (bug 585540) revealed a problem with resetting that only segfaulted on x86_64 but was wrong on x86.
There were two problems:
- GetPropCompiler::reset() was attempting to patch the inline path's type guard jump, even if it didn't exist;
- GetElemCompiler had no reset method, instead hijacking GetPropCompiler's, which used invalid values.
The attached patch respects pic->hasTypeGuard() and implements GetElemCompiler::reset().
Attachment #464185 -
Flags: review?(dvander)
![]() |
||
Updated•15 years ago
|
Attachment #464185 -
Flags: review?(dvander) → review+
Reporter | ||
Comment 1•15 years ago
|
||
Status: NEW → RESOLVED
Closed: 15 years ago
Resolution: --- → FIXED
Comment 2•15 years ago
|
||
Good work. I was just working on the same thing, discovered via running trace-tests with gczeal on. This was most likely causing unpredictable Mochitest failures as well.
You need to log in
before you can comment on or make changes to this bug.
Description
•