Re: Dejagnu as part of repository?

[jtc@redback.com (J.T. Conklin)]

>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
Andrew> Hello, Should the GDB repository contain dejagnu?

The only reservation I have is that dejagnu includes tcl, tk, and
expect.  This can lead to the situation where a user installs a
different version of these programs/libraries atop the versions 
that are already on their machines if they're not careful.  I've
blown away texinfo 4.0 several times already because GDB snapshots
contain a hacked version of texinfo 3.2.

So as you suggest, I think that dejagnu must be a different CVS
'module' than GDB so that those who want it must check it out
explicitly.

        --jtc


-- 
J.T. Conklin
RedBack Networks
From hjl@lucon.org Thu Oct 28 08:34:00 1999
From: hjl@lucon.org (H.J. Lu)
To: bryce@albatross.co.nz (Bryce McKinlay)
Cc: gdb@sourceware.cygnus.com (GDB)
Subject: Re: A patch for shared library support
Date: Thu, 28 Oct 1999 08:34:00 -0000
Message-id: <19991028153421.D072A1B494@ocean.lucon.org>
References: <3818203B.28D094A5@albatross.co.nz>
X-SW-Source: 1999-q4/msg00184.html
Content-length: 258

> 
> Great! Thanks - this bug has been annoying me too, and the patch seems to have
> fixed it. I've noticed that Redhat 6.1's gdb suffers from this as well.
> 
> regards
> 

I'd like to see this bug be fixed in gdb 4.19.

Thanks.

-- 
H.J. Lu (hjl@gnu.org)
From eliz@gnu.org Thu Oct 28 10:57:00 1999
From: Eli Zaretskii <eliz@gnu.org>
To: jimb@cygnus.com
Cc: gdb@sourceware.cygnus.com
Subject: Re: Hardware watchpoints
Date: Thu, 28 Oct 1999 10:57:00 -0000
Message-id: <199910281757.NAA25501@mescaline.gnu.org>
References: <np3dv6k64w.fsf@zwingli.cygnus.com> <19991019235249.917DC1B494@ocean.lucon.org> <199910201401.KAA28719@mescaline.gnu.org> <npvh82htxn.fsf@zwingli.cygnus.com> <199910221200.IAA24556@mescaline.gnu.org> <npn1tbnr5f.fsf@zwingli.cygnus.com> <199910231048.GAA31392@mescaline.gnu.org> <npaep8o97e.fsf@zwingli.cygnus.com> <npu2ndmzyh.fsf@zwingli.cygnus.com> <199910272007.QAA04553@mescaline.gnu.org> <npk8o8m1ch.fsf@zwingli.cygnus.com>
X-SW-Source: 1999-q4/msg00185.html
Content-length: 893

> Instead of adding yet another special flag, we could change
> value_primitive_field, or whoever extracts bitfields, to create a new
> value object

I can try doing that for bitfields, but I don't know what other parts
of GDB might behave in a similar fashion, especially in languages
other than C.

As for the special flag in struct value, I was thinking about walking
the value chain after it is created by evaluate_expression, and
marking all arrays and structs (but not pointers to structs) as being
non-watchable, except if they are at the head of the value chain.  If
this is a good algorithm, then it is easy to implement, and seems to
be safer, since it doesn't interfere with the lazy flag already used
by GDB.  However, it is possible that this seems safer to me because I
still don't feel I understand all the uses of the lazy flag in the
context of value creation and evaluation.
From jingham@leda.cygnus.com Thu Oct 28 11:45:00 1999
From: James Ingham <jingham@leda.cygnus.com>
To: gdb@sourceware.cygnus.com
Subject: Re: Dejagnu as part of repository?
Date: Thu, 28 Oct 1999 11:45:00 -0000
Message-id: <mfemefibd4.fsf@leda.cygnus.com>
References: <381824BF.BCE63370@cygnus.com> <5mogdjbkr2.fsf@jtc.redbacknetworks.com>
X-SW-Source: 1999-q4/msg00186.html
Content-length: 1785

jtc@redback.com (J.T. Conklin) writes:

> >>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
> Andrew> Hello, Should the GDB repository contain dejagnu?
> 
> The only reservation I have is that dejagnu includes tcl, tk, and
> expect.  This can lead to the situation where a user installs a
> different version of these programs/libraries atop the versions 
> that are already on their machines if they're not careful.  I've
> blown away texinfo 4.0 several times already because GDB snapshots
> contain a hacked version of texinfo 3.2.

It turns out that the combination of Dejagnu + the gdb testsuites is
not currently known to work with Tcl8.2.  There were some subtle
changes to the regexp parser made in order to support the wider Perl regexp
syntax, which will probably mean that some of the regexp's in dejagnu
or more likely the gdb testsuite will need to be tweaked.  

My guess is that this will not involve any major changes, but still,
to be consistent, for now we actually need to make sure that someone
running the testsuite for gdb is NOT using their already installed
version of Tcl if that version is 8.2.  So this argues for packaging
all this stuff together.

B.T.W. even if users did blow away their Tcl installation, it would
probably not be a big deal.  Our version of Tcl is a slightly modified 
version of 8.0.5, the most recent 8.0 variety, so most users would not 
even notice the change.  And different minor version number variants
of Tcl can coexist in the same location, so if they had an 8.1 or 8.2, 
we would not touch that.

Jim

-- 
++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++
Jim Ingham                                              jingham@cygnus.com
Cygnus Solutions Inc.
From jimb@cygnus.com Thu Oct 28 16:38:00 1999
From: Jim Blandy <jimb@cygnus.com>
To: Andreas Jaeger <aj@suse.de>
Cc: gdb@sourceware.cygnus.com
Subject: Re: Patch for warning in gdb/linux-thread.c
Date: Thu, 28 Oct 1999 16:38:00 -0000
Message-id: <npemefm4xt.fsf@zwingli.cygnus.com>
References: <u8ogdjx4xt.fsf@gromit.rhein-neckar.de>
X-SW-Source: 1999-q4/msg00187.html
Content-length: 243

> The current cvs version of gdb produces a number of harmless warnings
> in linuxthreads.  I've fixed the following warnings with the appended
> patch.

Applied --- thanks!

By the way, patches should go to gdb-patches@sourceware.cygnus.com.
From jimb@cygnus.com Fri Oct 29 14:15:00 1999
From: Jim Blandy <jimb@cygnus.com>
To: Eli Zaretskii <eliz@gnu.org>
Cc: hjl@lucon.org, gdb@sourceware.cygnus.com
Subject: Re: Hardware watchpoints
Date: Fri, 29 Oct 1999 14:15:00 -0000
Message-id: <npyacllvha.fsf@zwingli.cygnus.com>
References: <np3dv6k64w.fsf@zwingli.cygnus.com> <19991019235249.917DC1B494@ocean.lucon.org> <199910201401.KAA28719@mescaline.gnu.org> <npvh82htxn.fsf@zwingli.cygnus.com> <199910221200.IAA24556@mescaline.gnu.org> <npn1tbnr5f.fsf@zwingli.cygnus.com> <199910231048.GAA31392@mescaline.gnu.org> <npaep8o97e.fsf@zwingli.cygnus.com> <npu2ndmzyh.fsf@zwingli.cygnus.com> <199910272007.QAA04553@mescaline.gnu.org>
X-SW-Source: 1999-q4/msg00188.html
Content-length: 314

> As promised, diffs for all the changes that avoid inserting/removing
> watchpoints for lazy values are attached below.

I've applied this patch, with minor changes (after running it by
Michael for approval).  So Linux should be able to watch individual
members of structures now, in a general and non-hacky way.
From mharig@landmark.com Fri Oct 29 16:11:00 1999
From: "Mark Harig" <mharig@landmark.com>
To: gdb@sourceware.cygnus.com
Cc: "Mark Harig" <mharig@landmark.com>
Subject: Debugging threads on AIX 4.2/4.3
Date: Fri, 29 Oct 1999 16:11:00 -0000
Message-id: <85256819.007F5EC7.00@lnserver.landmark.com>
X-SW-Source: 1999-q4/msg00189.html
Content-length: 447

Can anyone tell me whether gdb (4.17.x or 4.18 or a recent snapshot) is able to
recognize and debug threads (Posix or otherwise) on AIX 4.2.x or AIX 4.3.x?  I'm
unable to find any description in the gdb documents (install or info) that
describe how gdb might be built to recognize threads, and haven't been able to
find any mention of it in these mailing lists.

Thank you in advance for any help on this matter.

Mark Harig
mharig@landmark.com