http://bugzilla.novell.com/show_bug.cgi?id=590820
http://bugzilla.novell.com/show_bug.cgi?id=590820#c2
Michael Matz changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P5 - None |P2 - High
Status|NEW |ASSIGNED
--- Comment #2 from Michael Matz 2010-03-24 15:46:41 UTC ---
This is a problem only in the listing code, when the assembler file contained
errors. What happens is this: there's uninitialized space allocated for
operands, and a fixup is recorded (referring to that space). If there are
no errors fixup_segment (via write_object_file) fills that formerly
uninitialized space, and all is well, because listings are generated after
object files are written.
With errors on the other hand write_object_file doesn't get called, the space
remains uninitialized, and the listing code accesses it. With malloc_perturb
the output then isn't as expected anymore.
Now instead of forcefully initializing the space always (which isn't needed
if there are no errors), or only if there are errors (which might not even
work, as the errors might have resulted in internally inconsistent data
structures), I'm leaning towards just disabling malloc checking while
running the testsuite.
--
Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.