Closed
Bug 683227
Opened 13 years ago
Closed 13 years ago
TI: Assertion failure: newShape != initialShape, at methodjit/PolyIC.cpp:384
Categories
(Core :: JavaScript Engine, defect)
Tracking
()
RESOLVED
FIXED
People
(Reporter: decoder, Unassigned)
References
Details
(Keywords: assertion, testcase, Whiteboard: fixed-in-jaegermonkey)
Attachments
(2 files)
(deleted),
application/x-tar
|
Details | |
(deleted),
patch
|
dvander
:
review+
|
Details | Diff | Splinter Review |
The attached testcase asserts on TI revision 6c8e0affe03e (run with -j -m -n -a), tested on 64 bit.
Updated•13 years ago
|
Attachment #556928 -
Attachment mime type: text/plain → application/x-tar
Comment 1•13 years ago
|
||
If a GC is triggered during analysis or compilation, we do not clear analysis information or destroy JITScripts. GC things pointed to by the JITScripts are pinned, but with TI enabled the jitcode can bake in shape numbers when constructing objects inline and these numbers will be invalidated if the GC is regenerating shapes. The fix just releases all jitcode on every GC; there is no need to keep the jitcode around even if we are compiling, as the compiler does not update the JSScript until it is finished (we do need to abandon compilation if a GC was triggered by it, which this patch does).
Attachment #558071 -
Flags: review?(dvander)
Updated•13 years ago
|
Attachment #558071 -
Flags: review?(dvander) → review+
Comment 2•13 years ago
|
||
Whiteboard: fixed-in-jaegermonkey
Comment 3•13 years ago
|
||
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
You need to log in
before you can comment on or make changes to this bug.
Description
•