This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] Adjust `pc-fp.exp' for ppc64/s390x (PR 12659)
- From: Tom Tromey <tromey at redhat dot com>
- To: Pedro Alves <palves at redhat dot com>
- Cc: Sergio Durigan Junior <sergiodj at redhat dot com>, GDB Patches <gdb-patches at sourceware dot org>, Jan Kratochvil <jan dot kratochvil at redhat dot com>
- Date: Wed, 01 Aug 2012 13:52:18 -0600
- Subject: Re: [PATCH] Adjust `pc-fp.exp' for ppc64/s390x (PR 12659)
- References: <m3mx2fmxmb.fsf@redhat.com> <5018ECBE.4020007@redhat.com>
>>>>> "Pedro" == Pedro Alves <palves@redhat.com> writes:
>> # Regression test for
>> # http://sourceware.org/bugzilla/show_bug.cgi?id=12659
>> gdb_test "info register pc fp" \
>> - "pc: ${valueof_pc}\[\r\n\]+fp: ${valueof_fp}\[\r\n\]+"
>> + "pc(:)?.*${valueof_pc}(.*${hex} <.*>)?\[\r\n\]+fp:
>> ${valueof_fp}\[\r\n\]+"
Pedro> Relaxing the output like that means that inadvertent changes to x86's
Pedro> or ppc/s390x output might go unnoticed. It's best to have
In this particular case, the check is really just to verify that the
named register, and nothing else, appears at the start of the line.
Before 12659 was fixed, "info register pc fp" printed:
sp fp: blah blah
fp: blah blah
The "fp" on the first line was the bogus bit.
I think the test would remain correct, with regards to what it was
intended to check, if it even went as far as "pc: .*\[\r\n\]+fp: .*";
checking the values is additional here.
Tom