two bug fixes on archer-tromey-call-frame-cfa

two bug fixes on archer-tromey-call-frame-cfa

[Tom Tromey]

FYI, I've checked in a couple of fixes to archer-tromey-call-frame-cfa.

Revision 8b4e2c4f9bfd463e24cc8c0927fb197447f7e0b1 fixes a bug reported
in RH bugzilla:

    https://bugzilla.redhat.com/show_bug.cgi?id=516627

I also finally added a test case for DW_OP_call_frame_cfa.
This patch is probably ready for resubmission upstream.


I also noticed a possible memory leak when evaluating a dwarf
expression.  This is fixed in c09f5506acfca898b69b99e6d849d8ed22bff9a8.


Jan, you should at least merge the first fix.  Whether you want the
second one is up to you, I suggest looking at it to see if you agree it
is correct.

Tom

Re: two bug fixes on archer-tromey-call-frame-cfa

[Jan Kratochvil]

On Tue, 11 Aug 2009 19:39:34 +0200, Tom Tromey wrote:
> I also noticed a possible memory leak when evaluating a dwarf
> expression.  This is fixed in c09f5506acfca898b69b99e6d849d8ed22bff9a8.
> 
> 
> Jan, you should at least merge the first fix.  Whether you want the
> second one is up to you, I suggest looking at it to see if you agree it
> is correct.

Included for Fedora.
(similiar fix is in the forked/unsubmitted vla branch, dwarf_expr_prep_ctx)


Thanks,
Jan

Re: two bug fixes on archer-tromey-call-frame-cfa

[Tom Tromey]

>>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:

Jan> (similiar fix is in the forked/unsubmitted vla branch, dwarf_expr_prep_ctx)

I'm going to send one upstream soon.  There's no point in letting these
bug fixes hide on a branch.  If you have a preference for which one, let
me know.  Otherwise I'll just push mine.

Tom

Re: two bug fixes on archer-tromey-call-frame-cfa

[Jan Kratochvil]

On Tue, 11 Aug 2009 20:38:23 +0200, Tom Tromey wrote:
> >>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:
> 
> Jan> (similiar fix is in the forked/unsubmitted vla branch, dwarf_expr_prep_ctx)
> 
> I'm going to send one upstream soon.  There's no point in letting these
> bug fixes hide on a branch.  If you have a preference for which one, let
> me know.  Otherwise I'll just push mine.

Sure push yours, I just made such a stupid notice, your patch is even more
clean (introducing the new make_cleanup_free_dwarf_expr_context stub) and
complete (both the cases are fixed).


Thanks,
Jan