Closed
Bug 58778
Opened 25 years ago
Closed 25 years ago
Boehm GC's getTypeName() implementation is flaky on Linux
Categories
(Core :: XPCOM, defect, P3)
Tracking
()
RESOLVED
FIXED
People
(Reporter: waterson, Assigned: waterson)
Details
Attachments
(2 files)
|
2.38 KB,
patch
|
Details | Diff | Splinter Review | |
|
2.66 KB,
patch
|
Details | Diff | Splinter Review |
The problem is that calling typeinfo(foo) ends up calling through a random
pointer. Attaching a patch that makes it a bit more robust...
|
Assignee | |
Comment 1•25 years ago
|
||
|
|
Assignee | |
Comment 2•25 years ago
|
||
This:
1. checks against a pointer that is a "known problem"
2. tries to determine if the place you're about to branch to looks at all
like a function
|
|
||
Comment 3•25 years ago
|
||
Good heuristic! We'll need one for PowerPC code as well. I can supply. (A nit,
"looks like a 4-bit aligned pointer" should read 4-byte aligned.) r=beard
|
|
Assignee | |
Comment 4•25 years ago
|
||
|
|
||
Comment 5•25 years ago
|
||
r=beard
|
|
||
Comment 6•25 years ago
|
||
+// The following are pointers that bamboozle are otherwise feeble
"are" s.b. "our", no? "extensible" is misspelled a bit later on.
a=brendan@mozilla.org otherwise.
/be
|
|
Assignee | |
Comment 7•25 years ago
|
||
checked in, with spelling errors fixed.
Status: NEW → RESOLVED
Closed: 25 years ago
Resolution: --- → FIXED
You need to log in
before you can comment on or make changes to this bug.
Description
•