This is the mail archive of the
mailing list for the GDB project.
Re: [RFC/testsuite] Update "info float" output for i386
- From: Fernando Nasser <fnasser at redhat dot com>
- To: Mark Kettenis <kettenis at chello dot nl>
- Cc: Daniel Jacobowitz <drow at mvista dot com>,gdb-patches at sources dot redhat dot com
- Date: Wed, 15 Jan 2003 09:48:41 -0500
- Subject: Re: [RFC/testsuite] Update "info float" output for i386
- Organization: Red Hat , Inc. - Toronto
- References: <20030114041526.GA1885@nevyn.them.org> <email@example.com>
Yes, I missed Mark's patch. Sorry.
If Daniel does not object, I prefer Mark's patch because it preserves the nature
of default.exp (tests with no inferior) and in doing so it can test for the
specific message that we are expecting in that condition.
I was also surprised how few tests we have for floating point handling. So
naming the new file float.exp is a good idea -- it may encourage more tests for
We could create a catch all file for info, but I guess it is tested in all
places where the object being "infoe'ed" is tested :-)
Mark Kettenis wrote:> Daniel Jacobowitz <firstname.lastname@example.org> writes:
I'm still in testsuite maintenance mode, for a little while longer. I'm on
the home stretch for GCC 2.95.3 + stabs; as soon as I finish, I'll do my
monthly apt-get dist-upgrade, the default compiler on my Debian machine will
change to GCC 3.2 + DWARF-2, and I'll be right back where I started at
dozens of failures. But I'll have a tremendous sense of accomplishment!
For what it's worth, I will continue to use my i386-unknown-freebsd4.7
system, which has GCC 2.95.4 with stabs as my primary development
machine for the foreseeable future. This GCC 2.95.4 is the default
compiler on FreeBSD 4.7-RELEASE which, to be honest, I've patched to
fix a debug info generation problem caused by the FreeBSD-specific
Some time ago Mark Kettenis updated i387-tdep.c to use frame_register_read
for "info float". This changed the result of "info float" with no running
program; I think for the better, and so did he, if I recall rightly. Can
anyone think of a reason not to update the testsuite correspondingly, as
with the attached patch?
At the time I checked in that patch, I did submit a patch to the testsuite:
Unfortunately it seems to have fallen between the cracks, and I
totally forgot about it; it has been happily sitting in my tree ever
since :-(. I sort-of prefer my patch over yours. Should I check it
in without Fernando's approval?
Red Hat - Toronto E-Mail: email@example.com
2323 Yonge Street, Suite #300
Toronto, Ontario M4P 2C9