Re: Regressions problem (200 failures)

[Jim Kingdon <kingdon@redhat.com>]

> I think GCC shouldn't put out any line notes for the prologue in the
> first place.  That's what's causing the problem, indirectly.  Does GDB
> require a line note in the prologue, or can we wait until the first
> bit of real code?

GDB expects the first line number to be for the real code (unless
something has changed, or I'm remembering it wrong or something - I
didn't actually play around with the test cases).
From aj@suse.de Sat Apr 01 00:00:00 2000
From: Andreas Jaeger <aj@suse.de>
To: shebs@shebs.cnchost.com
Cc: gdb@sourceware.cygnus.com, linux390@de.ibm.com
Subject: Re: GDB for S390
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <u8og9ussad.fsf@gromit.rhein-neckar.de>
References: <389CB0ED.755AD37C@shebs.cnchost.com>
X-SW-Source: 2000-q1/msg00114.html
Content-length: 769

>>>>> Stan Shebs writes:

 > Noticed this while surfing in strange corners of the web:
 > http://www.software.ibm.com/developerworks/opensource/linux390/download_src.html 

 > Not only is there a port of Linux to IBM's 390 mainframes,
 > but the GCC and GDB patches are available!  GDB patches
 > are against 4.18 and at first glance seem reasonable; we
 > should contact those guys and get them to donate...

There're also glibc and binutils patches ;-).  

I hope that these will soon be integrated (so far the patches have only
be added to the Linux 2.2 kernel)  I'm CC'ing the linux390 
developers - perhaps one of them can clarify the situation?  Or just donate
the port;-).

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.rhein-neckar.de
From kevinb@cygnus.com Sat Apr 01 00:00:00 2000
From: Kevin Buettner <kevinb@cygnus.com>
To: gdb@sourceware.cygnus.com
Cc: Eric Bachalo <ebachalo@cygnus.com>, Franz Sirl <Franz.Sirl-kernel@lauterbach.com>, khendricks@admin.ivey.uwo.ca
Subject: Linux/PPC support (was Re: Preparing for the GDB 5.0 ... release)
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <1000209051829.ZM2093@ocotillo.lan>
References: <00020900262503.05214@enzo.bigblue.local> <Franz.Sirl-kernel@lauterbach.com>
X-SW-Source: 2000-q1/msg00185.html
Content-length: 1659

I am at least partially to blame for nothing happening with regard to
getting the linux/ppc patches merged and committed to cvs.  My excuse
is that I've been assigned some challenging work that has sucked up
all of my work time as well as much of my personal time.

Earlier today, I spoke to my manager about this situation and I've
gotten his agreement (I think) to spend a week getting linux/ppc
support merged and committed. :-)  Better still, the week in question
is next week (2/14 thru 2/18, plus whatever I put in on the weekends).

The reason for doing it next week as opposed to some other random week
is that I'm in the midst of a two week stint of AIX work and it's
better for me to do it now when I'm focused on the PPC and related
architectures and have current (working) build trees for AIX.  Plus,
once committed, it'll give those of you who use linux/ppc on a daily
basis a chance to put it through its paces before the next gdb
release.

When I did the gdb port for linux/ppc, I made a copy of rs6000-tdep.c
and converted it to use the SysV ABI (along with some other linux
specific details).  At the time, this was the right decision since I
didn't have access to an AIX machine to build and test on.  Had I
attempted to work directly in rs6000-tdep.c, I would almost certainly
have broken the AIX port.  But now the situation is a bit different
and I think the right thing to do (maintenance-wise) is to merge
linux/ppc support into rs6000-tdep.c and (partially/minimally)
multi-arch the thing in the process.  There will still be a separate
ppclinux-nat.c file though.

Okay?

Kevin

-- 
Kevin Buettner
kev@primenet.com, kevinb@redhat.com
From kettenis@wins.uva.nl Sat Apr 01 00:00:00 2000
From: Mark Kettenis <kettenis@wins.uva.nl>
To: ac131313@cygnus.com
Cc: aj@suse.de, gdb@sourceware.cygnus.com
Subject: Re: core dump from GNU/Linux <sys/procfs.h>; Was: Build failure on Linux/i686
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <200002110736.e1B7afa00398@delius.kettenis.local>
References: <u8d7q4j4df.fsf@gromit.rhein-neckar.de> <200002101919.UAA23595@landau.wins.uva.nl> <38A353C3.DDD9A240@cygnus.com>
X-SW-Source: 2000-q1/msg00219.html
Content-length: 1144

   Date: Fri, 11 Feb 2000 11:11:47 +1100
   From: Andrew Cagney <ac131313@cygnus.com>

   Mark Kettenis wrote:

   > This is the solution I proposed:
   > 
   >    http://sourceware.cygnus.com/ml/gdb-patches/2000-q1/msg00103.html
   > 
   > Note that you'll need to patch <sys/procfs.h> too, otherwise GDB will
   > segfault when you try to debug a multithreaded app!

   Um, can you expand on this a little?  (You may have already).

The first version of glibc that includes the threads debugging library
(libthread.so) will be 2.1.3, which has not been officially released
yet.  Since both Andreas and I are glibc developers we were among the
first to test a GDB that makes use of that functionality apart from
the people who actually implemented it.  It looks as if Ulrich Drepper
(the glibc maintainer) and Michael have let things go slighty out of
sync.  So without patches to both GDB and the glibc prereleases it
won't work.

   If there is a header file found in a standard GNU/Linux distribution
   that can cause GDB to dump core we're going to need some sort of evasive
   action.  Sigh.

That's what I'm trying to prevent :-)

Mark
From ac131313@cygnus.com Sat Apr 01 00:00:00 2000
From: Andrew Cagney <ac131313@cygnus.com>
To: GDB Discussion <gdb@sourceware.cygnus.com>
Cc: GDB Patches <gdb-patches@sourceware.cygnus.com>
Subject: New file gdb/CONTRIBUTE guidelins for the contributor
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <38A3780F.2FE7A510@cygnus.com>
X-SW-Source: 2000-q1/msg00209.html
Content-length: 4972

Hello,

I'd like to put the attatched file forward as an addition to the other
readme files in the GDB directory.
Originally I was going to include this in the MAINTAINERS file but
decided that a separate file was probably of more benefit and more
likely to be noticed.

In terms of content, this really is a direct lift of GCC's
how-to-contribute page.

The one problem I can see at present is that the notes don't clearly
differentiate between GDB (src/gdb) and BINUTILS (src/bfd, src/include).

	Andrew


			Contributing to GDB

		[much of this is lifted from the GCC page]

We strongly encourage contributions of code, bugfixes, new
optimizations, new features, documentation updates, tests, web page
improvements, etc. for GDB.

There are certain legal requirements and style issues which all
contributions must meet.


o	Coding Standards

	All contributions must conform to the GNU Coding Standard.
	http://www.gnu.ai.mit.edu/prep/standards_toc.html
	Submissions which do not conform to the standards will be
	returned with a request to reformat the changes.

	For GDB, that standard is more tightly defined. GDB's
	coding standard is determined by the output of
	gnu-indent.

	This situation came about because, by the start of '99,
	GDB's coding style was so bad an inconsistent that it was
	decided to restart things from scratch.


o	Copyright Assignment

	There are certain legal requirements 

	Before we can accept code contributions from you, we need a
	copyright assignment form filled out.

	If you've developed some addition or patch to GDB that you
	would like to contribute, you should fill out a copyright
	assignment form and send it in to the FSF. We are unable to
	use code from you until this is on-file at the FSF, so get
	that paperwork in!  This form covers one batch of changes.
	Ref: http://gcc.gnu.org/fsf-forms/assignment-instructions.html

	If you think you're going to be doing continuing work on GDB, it
	would be easier to use a different form, which arranges to
	assign the copyright for all your future changes to GDB. It is
	called assign.future. Please note that if you switch
	employers, the new employer will need to fill out the
	disclaim.future form; there is no need to fill out the
	assign.future form again.
	Ref: http://gcc.gnu.org/fsf-forms/assign.future
	Ref: http://gcc.gnu.org/fsf-forms/disclaim.future

	There are several other forms you can fill out for different
	circumstances (e.g. to contribute an entirely new program, to
	contribute significant changes to a manual, etc.)
	Ref: http://gcc.gnu.org/fsf-forms/copyrights.html

	Small changes can be accepted without a copyright assignment
	form on file.

	This is pretty confusing! If you are unsure of what is
	necessary, just ask the GCC mailing list and we'll figure out
	what is best for you.

	Note: Many of these forms have a place for "name of
	program". Insert the name of one program in that place -- in
	this case, "GDB".


o	Submitting Patches

	Every patch must have several pieces of information before we
	can properly evaluate it.

	A description of the bug and how your patch fixes this
	bug. A reference to a testsuite failure is very helpful. For
	new features a description of the feature and your
	implementation.

	A ChangeLog entry as plaintext (separate from the patch); see
	the various ChangeLog files for format and content. Note that,
	unlike some other projects, we do require ChangeLogs also for
	documentation (i.e., .texi files).

	The patch itself. If you are accessing the CVS repository at:
	Cygnus, use "cvs update; cvs diff -c3p"; else, use "diff -c3p
	OLD NEW" or "diff -up OLD NEW". If your version of diff does
	not support these options, then get the latest version of GNU
	diff.

	We accept patches as plain text (preferred for the compilers
	themselves), MIME attachments (preferred for the web pages),
	or as uuencoded gzipped text.

	When you have all these pieces, bundle them up in a mail
	message and send it to gdb-patches@sourceware.cygnus.com. All
	patches and related discussion should be sent to the
	gcc-patches mailinglist. For further information on the GDB
	CVS repository, see the Anonymous read-only CVS access and
	Read-write CVS access page.

--

Supplemental information for GDB:

o	Please try to run the relevant testsuite before and after
	committing a patch

	If the contributor doesn't do it then the maintainer will.  A
	contributor might include before/after test results in their
	contribution.


o	For bug fixes, please try to include a way of
	demonstrating that the patch actually fixes something.

	The best way of doing this is to ensure that the
	testsuite contains one or more test cases that
	fail without the fix but pass with the fix.

	People are encouraged to submit patches that extend
	the testsuite.


o	Please read your patch before submitting it.

	A patch containing several unrelated changes or
	arbitrary reformats will be returned with a request
	to re-formatting / split it.