Guy Harris wrote:
On Fri, Jul 16, 2004 at 11:53:58PM +0200, Joerg Mayer wrote:
Looking thourgh specflies it found the following patch in RedHats Fedora:
RedHat/Fedora ethereal.spec:
* Mon Jun 14 2004 Phil Knirsch <pknirsch@xxxxxxxxxx> 0.10.4-2
- Include ia64 segfault patch (#113726).
diff -rU4 -N ethereal-0.10.0a/epan/proto.c ethereal-0.10.0a-ia64-segfault.patch/epan/proto.c
--- ethereal-0.10.0a/epan/proto.c 2003-12-06 13:46:55.000000000 -0500
+++ ethereal-0.10.0a-ia64-segfault.patch/epan/proto.c 2004-03-04 16:31:31.000000000 -0500
@@ -1975,9 +1975,9 @@
}
FIELD_INFO_NEW(fi);
- fi->hfinfo = hfinfo;
+ fi->hfinfo = gpa_hfinfo.hfi[hfindex];
fi->start = start;
fi->start+=(tvb)?TVB_RAW_OFFSET(tvb):0;
fi->length = *length;
fi->tree_type = -1;
As I failed to understand what it does and as the RedHat bugzilla doesn't
allow anonymous access to this bug: Can anyone explain to ma what this
fix does (aka why it is necessary)?
I can't, given the way the macro works, unless, somehow, the
"g_assert()" fails, or the C preprocessor fails on IA-64.
I tried logging in to Bugzilla, and they won't let me look at the bug,
either; Phil, what is the reason for that change?
The fix is only neede for our older toolchain for AS2.1 on IA64 which is
prone to some really strange bugs. So this "fix" is nothing but a
workaround for yet another toolchain bug in AS2.1 (which strangely
enough almost always hit ethereal ;-).
So i see no reason at all to include it in the upstream version, the
original should have worked perfectly (which it does on all releases and
archs except IA64...)
Hope that explains it a little.
Read ya, Phil
--
Philipp Knirsch | Tel.: +49-711-96437-470
Development | Fax.: +49-711-96437-111
Red Hat GmbH | Email: Phil Knirsch <phil@xxxxxxxxx>
Hauptstaetterstr. 58 | Web: http://www.redhat.de/
D-70178 Stuttgart
Motd: You're only jealous cos the little penguins are talking to me.
