time to create the gdb-7.0.1 release (?)

time to create the gdb-7.0.1 release (?)

[Joel Brobecker]

Hello,

According to our suggested schedule, yesterday (Dec 17) was the time
we were hoping to produce gdb-7.0.1. Looking at the ChangeLog, here
is the list of changes:

 - Fix a build compilation on IRIX 5.3:
   2009-11-23  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
   * dwarf2read.c [HAVE_MMAP] (MAP_FAILED): Define if missing.
   
 - Add support for Thumb2 debugging:
   2009-11-13  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
   Backport from mainline.
     2009-07-28  Daniel Jacobowitz  <dan@codesourcery.com>
     
     * arm-tdep.c (arm_push_dummy_call): Set the low bit of LR for
     a Thumb entry point.
     (thumb_get_next_pc): Handle Thumb-2 and ARM v6 instructions.
     Refuse to single step into IT blocks.
     
     2009-10-28  Daniel Jacobowitz  <dan@codesourcery.com>
     Reported by Antti Hatala <ahatala@nvidia.com>
     * arm-tdep.c (thumb_get_next_pc): Limit check to IT instructions
     correctly.
   
 - Fix a crash on solaris when reading symbols:
   2009-10-22  Paul Pluzhnikov  <ppluzhnikov@google.com>
   PR gdb/10819
   * dwarf2-frame.c (find_cie): Don't call bsearch on empty cie_table.
   * objfiles.c (find_pc_section): Likewise.
   (update_section_map): Don't allocate empty table.
   
 - I think the following change is mostly cosmetic:
   2009-10-19  Don Lee  <don.lee@sunplusct.com>
   * score-tdep.c: Delete some simulator dependent codes.
   * score-tdep.h: Delete some simulator dependent macro definitions.
   
 - Fix a build failure on newer versions of FreeBSD:
   2008-10-16  Steven G. Kargl  <kargl@gcc.gnu.org>  (tiny patch)
   * amd64fbsd-nat.c (amd64fbsd_supply_pcb): Conditionally compile in
   support for pcb->pcb_{fs,ds,es,gs} on FreeBSD older than 8.0.
   
 - GDB SEGV due to stack overflow:
   2009-10-08  Paul Pluzhnikov  <ppluzhnikov@google.com>
   PR gdb/10457
   * elfread.c (elf_symtab_read): Don't use alloca in a loop.

There is also a change in the documentation:

   2009-10-04  Pedro Alves  <pedro@codesourcery.com>
   * gdb.texinfo (Remote Protocol): Don't mention vCont;T.

Anything else that should go in that (corrective) release?

-- 
Joel

Re: time to create the gdb-7.0.1 release (?)

[Ralf Corsepius]

On 12/18/2009 07:17 AM, Joel Brobecker wrote:
> Hello,
>
> According to our suggested schedule, yesterday (Dec 17) was the time
> we were hoping to produce gdb-7.0.1. Looking at the ChangeLog, here
> is the list of changes:

> Anything else that should go in that (corrective) release?

Wasn't there a "shipping wrong COPYING" (GPLv3 / GPLv2) issue?

Ralf

Re: time to create the gdb-7.0.1 release (?)

[Joel Brobecker]

> Wasn't there a "shipping wrong COPYING" (GPLv3 / GPLv2) issue?

There was, indeed. I started working on it (I just sent an RFC on
gdb-patches).

-- 
Joel

Re: time to create the gdb-7.0.1 release (?)

[Hui Zhu]

Hi Joel,

Sorry that I have to be the trouble guy.

http://sourceware.org/ml/gdb-patches/2009-12/msg00010.html

This bug fix is still waiting review.

Thanks,
Hui

On Fri, Dec 18, 2009 at 14:17, Joel Brobecker <brobecker@adacore.com> wrote:
> Hello,
>
> According to our suggested schedule, yesterday (Dec 17) was the time
> we were hoping to produce gdb-7.0.1. Looking at the ChangeLog, here
> is the list of changes:
>
>  - Fix a build compilation on IRIX 5.3:
>   2009-11-23  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
>   * dwarf2read.c [HAVE_MMAP] (MAP_FAILED): Define if missing.
>
>  - Add support for Thumb2 debugging:
>   2009-11-13  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
>   Backport from mainline.
>     2009-07-28  Daniel Jacobowitz  <dan@codesourcery.com>
>
>     * arm-tdep.c (arm_push_dummy_call): Set the low bit of LR for
>     a Thumb entry point.
>     (thumb_get_next_pc): Handle Thumb-2 and ARM v6 instructions.
>     Refuse to single step into IT blocks.
>
>     2009-10-28  Daniel Jacobowitz  <dan@codesourcery.com>
>     Reported by Antti Hatala <ahatala@nvidia.com>
>     * arm-tdep.c (thumb_get_next_pc): Limit check to IT instructions
>     correctly.
>
>  - Fix a crash on solaris when reading symbols:
>   2009-10-22  Paul Pluzhnikov  <ppluzhnikov@google.com>
>   PR gdb/10819
>   * dwarf2-frame.c (find_cie): Don't call bsearch on empty cie_table.
>   * objfiles.c (find_pc_section): Likewise.
>   (update_section_map): Don't allocate empty table.
>
>  - I think the following change is mostly cosmetic:
>   2009-10-19  Don Lee  <don.lee@sunplusct.com>
>   * score-tdep.c: Delete some simulator dependent codes.
>   * score-tdep.h: Delete some simulator dependent macro definitions.
>
>  - Fix a build failure on newer versions of FreeBSD:
>   2008-10-16  Steven G. Kargl  <kargl@gcc.gnu.org>  (tiny patch)
>   * amd64fbsd-nat.c (amd64fbsd_supply_pcb): Conditionally compile in
>   support for pcb->pcb_{fs,ds,es,gs} on FreeBSD older than 8.0.
>
>  - GDB SEGV due to stack overflow:
>   2009-10-08  Paul Pluzhnikov  <ppluzhnikov@google.com>
>   PR gdb/10457
>   * elfread.c (elf_symtab_read): Don't use alloca in a loop.
>
> There is also a change in the documentation:
>
>   2009-10-04  Pedro Alves  <pedro@codesourcery.com>
>   * gdb.texinfo (Remote Protocol): Don't mention vCont;T.
>
> Anything else that should go in that (corrective) release?
>
> --
> Joel
>

Re: time to create the gdb-7.0.1 release (?)

[Joel Brobecker]

> http://sourceware.org/ml/gdb-patches/2009-12/msg00010.html

I didn't think that this patch was proposed for 7.0.1?

-- 
Joel

Re: time to create the gdb-7.0.1 release (?)

[Hui Zhu]

OK.  Never mind.

(gdb) c
Continuing.
a = 0 b = 0 c = 1
a = 3
a = 3 b = 3 c = 1
a = 3 b = 3 c = 2
a = 1 b = 3 c = 2
The next instruction is syscall exit_group.  It will make the program
exit.  Do you want to stop the program?([y] or n)
Process record: inferior program stopped.

Program received signal SIGTRAP, Trace/breakpoint trap.
0xb7fe3405 in __kernel_vsyscall ()
(gdb) c
Continuing.
The next instruction is syscall exit_group.  It will make the program
exit.  Do you want to stop the program?([y] or n)
Process record: inferior program stopped.

Program received signal SIGABRT, Aborted.
0xb7fe3405 in __kernel_vsyscall ()
(gdb) c
Continuing.

Program terminated with signal SIGABRT, Aborted.
The program no longer exists.


This bug will keep affect 7.0.1.  :(

Hui


On Fri, Dec 18, 2009 at 16:55, Joel Brobecker <brobecker@adacore.com> wrote:
>> http://sourceware.org/ml/gdb-patches/2009-12/msg00010.html
>
> I didn't think that this patch was proposed for 7.0.1?
>
> --
> Joel
>

Re: time to create the gdb-7.0.1 release (?)

[Joel Brobecker]

> This bug will keep affect 7.0.1.  :(

Given that 7.1 is not far behind, I don't think that's going to be
a big issue.

-- 
Joel

Re: time to create the gdb-7.0.1 release (?)

[Jack Howarth]

On Fri, Dec 18, 2009 at 10:17:36AM +0400, Joel Brobecker wrote:
> Hello,
> 
> According to our suggested schedule, yesterday (Dec 17) was the time
> we were hoping to produce gdb-7.0.1. Looking at the ChangeLog, here
> is the list of changes:
> 
>  - Fix a build compilation on IRIX 5.3:
>    2009-11-23  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
>    * dwarf2read.c [HAVE_MMAP] (MAP_FAILED): Define if missing.
>    
>  - Add support for Thumb2 debugging:
>    2009-11-13  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
>    Backport from mainline.
>      2009-07-28  Daniel Jacobowitz  <dan@codesourcery.com>
>      
>      * arm-tdep.c (arm_push_dummy_call): Set the low bit of LR for
>      a Thumb entry point.
>      (thumb_get_next_pc): Handle Thumb-2 and ARM v6 instructions.
>      Refuse to single step into IT blocks.
>      
>      2009-10-28  Daniel Jacobowitz  <dan@codesourcery.com>
>      Reported by Antti Hatala <ahatala@nvidia.com>
>      * arm-tdep.c (thumb_get_next_pc): Limit check to IT instructions
>      correctly.
>    
>  - Fix a crash on solaris when reading symbols:
>    2009-10-22  Paul Pluzhnikov  <ppluzhnikov@google.com>
>    PR gdb/10819
>    * dwarf2-frame.c (find_cie): Don't call bsearch on empty cie_table.
>    * objfiles.c (find_pc_section): Likewise.
>    (update_section_map): Don't allocate empty table.
>    
>  - I think the following change is mostly cosmetic:
>    2009-10-19  Don Lee  <don.lee@sunplusct.com>
>    * score-tdep.c: Delete some simulator dependent codes.
>    * score-tdep.h: Delete some simulator dependent macro definitions.
>    
>  - Fix a build failure on newer versions of FreeBSD:
>    2008-10-16  Steven G. Kargl  <kargl@gcc.gnu.org>  (tiny patch)
>    * amd64fbsd-nat.c (amd64fbsd_supply_pcb): Conditionally compile in
>    support for pcb->pcb_{fs,ds,es,gs} on FreeBSD older than 8.0.
>    
>  - GDB SEGV due to stack overflow:
>    2009-10-08  Paul Pluzhnikov  <ppluzhnikov@google.com>
>    PR gdb/10457
>    * elfread.c (elf_symtab_read): Don't use alloca in a loop.
> 
> There is also a change in the documentation:
> 
>    2009-10-04  Pedro Alves  <pedro@codesourcery.com>
>    * gdb.texinfo (Remote Protocol): Don't mention vCont;T.
> 
> Anything else that should go in that (corrective) release?
> 
> -- 
> Joel

Joel,
   I don't see any mention of fixes for MacOS X. Have all the recent
changes for darwin in gdb trunk been backported into gdb 7.0 branch?
                      Jack

Re: time to create the gdb-7.0.1 release (?)

[Joel Brobecker]

> I don't see any mention of fixes for MacOS X. Have all the recent
> changes for darwin in gdb trunk been backported into gdb 7.0 branch?

I only showed the changes that went into the branch. So the answer
is mostly likely negative. I don't necessarily object to doing so
but I would like the relevant patches to be identified and approved
quickly, so that the release can be made soon.  Remember that 7.1
is only 4-6 weeks away.

-- 
Joel

Re: time to create the gdb-7.0.1 release (?)

[Michael Snyder]

Joel Brobecker wrote:

> Anything else that should go in that (corrective) release?

Well, we have process record save/restore, bookmarks, floating point
support (387), and hardware watchpoint support.  Those will all go in,
right?

Re: time to create the gdb-7.0.1 release (?)

[Joel Brobecker]

> >Anything else that should go in that (corrective) release?
> 
> Well, we have process record save/restore, bookmarks, floating point
> support (387), and hardware watchpoint support.  Those will all go in,
> right?

I suspect a confusion between 7.0.1, which is a minor corrective release
over 7.0 (release made off the same branch), and 7.1, which is the next
"major" release, with new features etc.

-- 
Joel

Re: time to create the gdb-7.0.1 release (?)

[Michael Snyder]

Joel Brobecker wrote:
>>> Anything else that should go in that (corrective) release?
>> Well, we have process record save/restore, bookmarks, floating point
>> support (387), and hardware watchpoint support.  Those will all go in,
>> right?
> 
> I suspect a confusion between 7.0.1, which is a minor corrective release
> over 7.0 (release made off the same branch), and 7.1, which is the next
> "major" release, with new features etc.

OK.  ;-)

Re: time to create the gdb-7.0.1 release (?)

[Jan Kratochvil]

Hi Joel,

On Fri, 18 Dec 2009 07:17:36 +0100, Joel Brobecker wrote:
> Anything else that should go in that (corrective) release?

http://sourceware.org/ml/gdb-patches/2009-11/msg00422.html
http://sourceware.org/ml/gdb-cvs/2009-11/msg00169.html
[commit] SEGV trying to complete "catch exception" command
	* completer.c (complete_line_internal): Make sure the command
	completer is not NULL before calling it.

http://sourceware.org/ml/gdb-patches/2009-11/msg00269.html
http://sourceware.org/ml/gdb-cvs/2009-11/msg00118.html
Re: [RFA] dcache invalidate fix
	* dcache.c (dcache_block): Replace member newer with next,prev.
	(dcache_struct): Delete member newest.
	(block_func): New typedef.
	(append_block, remove_block, for_each_block): New functions.
	(invalidate_block, free_block): New functions.
	(dcache_invalidate): Update
	(dcache_invalidate_line, dcache_alloc): Update to use new list
	accessors.
	(dcache_free): Ditto.  Fix memory leak.
Locking up GDB on ppc64 inferiors.
While this unchecked-in is less intrusive
http://sourceware.org/ml/gdb-patches/2009-11/msg00233.html
it forgot to update `dcache->newest'.

Optionally a set of testsuite results stabilization patches:
http://sourceware.org/ml/gdb-patches/2009-11/msg00594.html
http://sourceware.org/ml/gdb-cvs/2009-11/msg00233.html
Subject: [patch] testsuite: bigcore.exp fuzzy PASS message fix
http://sourceware.org/ml/gdb-patches/2009-11/msg00593.html
http://sourceware.org/ml/gdb-cvs/2009-12/msg00013.html
Subject: [patch] testcase: Fix spurious structs.exp FAILs
http://sourceware.org/ml/gdb-patches/2009-11/msg00573.html
http://sourceware.org/ml/gdb-cvs/2009-11/msg00232.html
Subject: [patch] testsuite fuzzy results fixup: foll-fork.exp


Regards,
Jan

Re: time to create the gdb-7.0.1 release (?)

[Jan Kratochvil]

On Fri, 18 Dec 2009 07:17:36 +0100, Joel Brobecker wrote:
> Anything else that should go in that (corrective) release?

http://sourceware.org/ml/gdb-patches/2009-08/msg00448.html
http://sourceware.org/ml/gdb-cvs/2009-08/msg00139.html
Re: [patch] Support constant DW_AT_data_member_location by GCC PR debug/40659
	Support constant DW_AT_data_member_location by GCC PR debug/40659.
	* dwarf2read.c
	(dwarf2_add_field <DW_TAG_member> <DW_AT_data_member_location>):
	Initialize BYTE_OFFSET to 0 by default.  Explicitly check if
	attr_form_is_block.
	(dwarf2_add_field <DW_TAG_inheritance> <DW_AT_data_member_location>)
	(read_common_block <DW_AT_data_member_location>): New variable
	byte_offset.  Fix crash on non-DW_BLOCK ATTR values.
	Support constant DW_AT_data_member_location by GCC PR debug/40659.
	* gdb.dwarf2/dw2-inheritance.exp, gdb.dwarf2/dw2-inheritance.S: New.

As otherwise gdb-7.0 SEGVs on normal binaries from either gcc-4.5 or
gcc-4.4fedora12.


Regards,
Jan

Re: time to create the gdb-7.0.1 release (?)

[Joel Brobecker]

Nice research work! 

I checked all the patches you proposed, except for one (see below).
I hesitated a bit on the dcache invalidate fix, but checked it in
in the end, mostly because of the symptoms (infinite loop). If we have
any problem with it, either 7.1 will soon be out, or I'll make an
emergency 7.0.2.

I did not see much point in regarding the testsuite patches, but checked
them in anyway, in case some people are in fact interested in the
testsuite results for the branch (I personally only follow the results
on the HEAD).

The patch that I did not commit was:

http://sourceware.org/ml/gdb-patches/2009-08/msg00448.html
http://sourceware.org/ml/gdb-cvs/2009-08/msg00139.html
Re: [patch] Support constant DW_AT_data_member_location by GCC PR debug/40659
        Support constant DW_AT_data_member_location by GCC PR debug/40659.
        * dwarf2read.c
        (dwarf2_add_field <DW_TAG_member> <DW_AT_data_member_location>):
        Initialize BYTE_OFFSET to 0 by default.  Explicitly check if
        attr_form_is_block.
        (dwarf2_add_field <DW_TAG_inheritance> <DW_AT_data_member_location>)
        (read_common_block <DW_AT_data_member_location>): New variable
        byte_offset.  Fix crash on non-DW_BLOCK ATTR values.
        Support constant DW_AT_data_member_location by GCC PR debug/40659.
        * gdb.dwarf2/dw2-inheritance.exp, gdb.dwarf2/dw2-inheritance.S: New.

As far as I can tell, this one is already in the branch (was checked
in before the branch got cut). Do you actually see a SEGV with gdb-7.0?
If yes, must be something else... Let me know, because I'd like to make
the release RSN.

-- 
Joel

Re: time to create the gdb-7.0.1 release (?)

[Jan Kratochvil]

On Mon, 21 Dec 2009 07:00:46 +0100, Joel Brobecker wrote:
> I did not see much point in regarding the testsuite patches, but checked
> them in anyway, in case some people are in fact interested in the
> testsuite results for the branch

It was intended to make life easier for distro maintainers

> (I personally only follow the results on the HEAD).

while upstream probably does not wish they would use HEAD.


> The patch that I did not commit was:
> 
> http://sourceware.org/ml/gdb-patches/2009-08/msg00448.html
> http://sourceware.org/ml/gdb-cvs/2009-08/msg00139.html
> Re: [patch] Support constant DW_AT_data_member_location by GCC PR debug/40659
...
> As far as I can tell, this one is already in the branch (was checked
> in before the branch got cut).

OK, sorry, too many branches around.
Also this one was checked in HEAD already before branching:
http://sourceware.org/ml/gdb-cvs/2009-09/msg00009.html
http://sourceware.org/ml/gdb-patches/2009-09/msg00027.html
Re: RFC: DW_OP_call_frame_cfa, again


> Let me know, because I'd like to make the release RSN.

Fine with me, thanks.


Regards,
Jan

Re: time to create the gdb-7.0.1 release (?)

[Vladimir Prus]

Joel Brobecker wrote:

> Hello,
> 
> According to our suggested schedule, yesterday (Dec 17) was the time
> we were hoping to produce gdb-7.0.1. Looking at the ChangeLog, here
> is the list of changes:

It seems to me that
 
	http://permalink.gmane.org/gmane.comp.gdb.patches/54059

fixes a sufficiently nasty bug, and is sufficiently obvious itself, to
warrant including in 7.0.1

- Volodya