[branching soon!] GDB 13 release -- 2022-12-04 update

[branching soon!] GDB 13 release -- 2022-12-04 update

[Joel Brobecker]

Hello,

*GREAT* progress with the list of issues that were addressed last week,
as you'll see in the summary below. Thanks to everyone who helped with
that strong push!

We're down to one issue left, which have been under discussion.
Hopefully we converge soon. Otherwise, I propose we still go ahead
and branch, and we'll look at backporting the fix to the branch
once it is available.

Based on the above, I propose we aim for branching next weekend,
Sat-Sun Dec 10-11.

Thanks again, everyone.

Fixed Since the Previous Update:
--------------------------------

  * [SimonM[ PR gdb/28275
    commit_resumed_state assertion failure when killing running inferior and running again
    https://sourceware.org/bugzilla/show_bug.cgi?id=28275

  * [TomT] PR symtab/29105
    new DWARF reader still slow
    https://sourceware.org/bugzilla/show_bug.cgi?id=29105

  * [TomDV,AndrewB] PR python/29712
    UNRESOLVED: gdb.python/py-disasm.exp: global_disassembler=ReadMemoryGdbErrorDisassembler: disassemble test
    https://sourceware.org/bugzilla/show_bug.cgi?id=29712

  * [TomT] PR gdb/29787
    Bug 29787 - Using "set debug infrun on" and "set logging enabled" breaks GDB prompt and makes GDB crash
    https://sourceware.org/bugzilla/show_bug.cgi?id=29787

  * [TomT] cli/29800
    Trying to "document" an undefined command causes crash
    https://sourceware.org/bugzilla/show_bug.cgi?id=29800

Added Since the Last Update:
----------------------------

  < none :) >

Other Ongoing Items:
--------------------

  * [Torbjorn] tdep/29738
    Arm M-profile dwarf2 unwinder performance suffers from exponential growth
    https://sourceware.org/bugzilla/show_bug.cgi?id=29738

    patch v2, 2022-11-18, reviewed 2022-11-21:
    https://sourceware.org/pipermail/gdb-patches/2022-November/193960.html

-- 
Joel

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Tom Tromey]

>>>>> "Joel" == Joel Brobecker via Gdb-patches <gdb-patches@sourceware.org> writes:

Joel> We're down to one issue left, which have been under discussion.
Joel> Hopefully we converge soon. Otherwise, I propose we still go ahead
Joel> and branch, and we'll look at backporting the fix to the branch
Joel> once it is available.

I hate to say it but I wonder if we should add the Windows target-async
interrupt problem to the list.  I've been working on it; but if we can't
make it work in time, maybe one option is to disable Windows
target-async for GDB 13.  I think this would probably be ok, because
AFAIK the only thing really relying on target-async is DAP, which isn't
going into GDB 13 anyway.

Tom

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Joel Brobecker]

Hi Tom,

> Joel> We're down to one issue left, which have been under discussion.
> Joel> Hopefully we converge soon. Otherwise, I propose we still go ahead
> Joel> and branch, and we'll look at backporting the fix to the branch
> Joel> once it is available.
> 
> I hate to say it but I wonder if we should add the Windows target-async
> interrupt problem to the list.  I've been working on it; but if we can't
> make it work in time, maybe one option is to disable Windows
> target-async for GDB 13.  I think this would probably be ok, because
> AFAIK the only thing really relying on target-async is DAP, which isn't
> going into GDB 13 anyway.

No problem, Tom.

What do you think of the following approach? Disable Windows target-async
for now, until you have a fix. That way, when I branch, it's all set,
and I can follow the standard procedure for generating the first
pre-release. Then, when we have the fix for the interrupt issue, we can
re-enable it, with the option of backporting it to the GDB 13 branch
if OK to do so.

-- 
Joel

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Tom Tromey]

>>>>> "Joel" == Joel Brobecker via Gdb-patches <gdb-patches@sourceware.org> writes:

Joel> What do you think of the following approach? Disable Windows target-async
Joel> for now, until you have a fix. That way, when I branch, it's all set,
Joel> and I can follow the standard procedure for generating the first
Joel> pre-release. Then, when we have the fix for the interrupt issue, we can
Joel> re-enable it, with the option of backporting it to the GDB 13 branch
Joel> if OK to do so.

I found a fix today, so let's try that unless the fix looks like trouble.
If we don't want that fix, it's easy to disable async, that's a simple
patch, so I can easily do it later this week.

Tom

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Joel Brobecker]

> Joel> What do you think of the following approach? Disable Windows target-async
> Joel> for now, until you have a fix. That way, when I branch, it's all set,
> Joel> and I can follow the standard procedure for generating the first
> Joel> pre-release. Then, when we have the fix for the interrupt issue, we can
> Joel> re-enable it, with the option of backporting it to the GDB 13 branch
> Joel> if OK to do so.
> 
> I found a fix today, so let's try that unless the fix looks like trouble.
> If we don't want that fix, it's easy to disable async, that's a simple
> patch, so I can easily do it later this week.

That's even better. Thanks Tom!

-- 
Joel

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Luis Machado]

Hi Joel,

Thanks for the update.

On 12/4/22 04:33, Joel Brobecker via Gdb-patches wrote:
> Hello,
> 
> *GREAT* progress with the list of issues that were addressed last week,
> as you'll see in the summary below. Thanks to everyone who helped with
> that strong push!
> 
> We're down to one issue left, which have been under discussion.
> Hopefully we converge soon. Otherwise, I propose we still go ahead
> and branch, and we'll look at backporting the fix to the branch
> once it is available.
> 
> Based on the above, I propose we aim for branching next weekend,
> Sat-Sun Dec 10-11.
> 
> Thanks again, everyone.
> 
> Fixed Since the Previous Update:
> --------------------------------
> 
>    * [SimonM[ PR gdb/28275
>      commit_resumed_state assertion failure when killing running inferior and running again
>      https://sourceware.org/bugzilla/show_bug.cgi?id=28275
> 
>    * [TomT] PR symtab/29105
>      new DWARF reader still slow
>      https://sourceware.org/bugzilla/show_bug.cgi?id=29105
> 
>    * [TomDV,AndrewB] PR python/29712
>      UNRESOLVED: gdb.python/py-disasm.exp: global_disassembler=ReadMemoryGdbErrorDisassembler: disassemble test
>      https://sourceware.org/bugzilla/show_bug.cgi?id=29712
> 
>    * [TomT] PR gdb/29787
>      Bug 29787 - Using "set debug infrun on" and "set logging enabled" breaks GDB prompt and makes GDB crash
>      https://sourceware.org/bugzilla/show_bug.cgi?id=29787
> 
>    * [TomT] cli/29800
>      Trying to "document" an undefined command causes crash
>      https://sourceware.org/bugzilla/show_bug.cgi?id=29800
> 
> Added Since the Last Update:
> ----------------------------
> 
>    < none :) >
> 
> Other Ongoing Items:
> --------------------
> 
>    * [Torbjorn] tdep/29738
>      Arm M-profile dwarf2 unwinder performance suffers from exponential growth
>      https://sourceware.org/bugzilla/show_bug.cgi?id=29738
> 
>      patch v2, 2022-11-18, reviewed 2022-11-21:
>      https://sourceware.org/pipermail/gdb-patches/2022-November/193960.html
> 

It would be nice if someone could take a look at Torbjörn's patches for the performance issue.

They touch generic code, so we need to make sure it works correctly and that the strategy is sound.

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Eli Zaretskii]

> Date: Fri, 9 Dec 2022 11:14:17 +0000
> Cc: Torbjorn SVENSSON <torbjorn.svensson@foss.st.com>
> From: Luis Machado via Gdb-patches <gdb-patches@sourceware.org>
> 
> Hi Joel,
> 
> Thanks for the update.
> 
> On 12/4/22 04:33, Joel Brobecker via Gdb-patches wrote:
> > Hello,
> > 
> > *GREAT* progress with the list of issues that were addressed last week,
> > as you'll see in the summary below. Thanks to everyone who helped with
> > that strong push!
> > 
> > We're down to one issue left, which have been under discussion.
> > Hopefully we converge soon. Otherwise, I propose we still go ahead
> > and branch, and we'll look at backporting the fix to the branch
> > once it is available.
> > 
> > Based on the above, I propose we aim for branching next weekend,
> > Sat-Sun Dec 10-11.
> > 
> > Thanks again, everyone.
> > 
> > Fixed Since the Previous Update:
> > --------------------------------
> > 
> >    * [SimonM[ PR gdb/28275
> >      commit_resumed_state assertion failure when killing running inferior and running again
> >      https://sourceware.org/bugzilla/show_bug.cgi?id=28275
> > 
> >    * [TomT] PR symtab/29105
> >      new DWARF reader still slow
> >      https://sourceware.org/bugzilla/show_bug.cgi?id=29105
> > 
> >    * [TomDV,AndrewB] PR python/29712
> >      UNRESOLVED: gdb.python/py-disasm.exp: global_disassembler=ReadMemoryGdbErrorDisassembler: disassemble test
> >      https://sourceware.org/bugzilla/show_bug.cgi?id=29712
> > 
> >    * [TomT] PR gdb/29787
> >      Bug 29787 - Using "set debug infrun on" and "set logging enabled" breaks GDB prompt and makes GDB crash
> >      https://sourceware.org/bugzilla/show_bug.cgi?id=29787
> > 
> >    * [TomT] cli/29800
> >      Trying to "document" an undefined command causes crash
> >      https://sourceware.org/bugzilla/show_bug.cgi?id=29800
> > 
> > Added Since the Last Update:
> > ----------------------------
> > 
> >    < none :) >
> > 
> > Other Ongoing Items:
> > --------------------
> > 
> >    * [Torbjorn] tdep/29738
> >      Arm M-profile dwarf2 unwinder performance suffers from exponential growth
> >      https://sourceware.org/bugzilla/show_bug.cgi?id=29738
> > 
> >      patch v2, 2022-11-18, reviewed 2022-11-21:
> >      https://sourceware.org/pipermail/gdb-patches/2022-November/193960.html
> > 

What is the decision regarding target-async support on MS-Windows?
Will it be in GDB 13 or will it be deferred to GDB 14?

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Joel Brobecker]

> What is the decision regarding target-async support on MS-Windows?
> Will it be in GDB 13 or will it be deferred to GDB 14?

Pending any new issues found with this feature, I think there is a very
good chance that it'll go in GDB 13, thanks to Tom having sent some
patches to fix the issue he's found (interrupting programs).

-- 
Joel

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Eli Zaretskii]

> Date: Fri, 9 Dec 2022 17:23:26 +0400
> From: Joel Brobecker <brobecker@adacore.com>
> Cc: Luis Machado <luis.machado@arm.com>, brobecker@adacore.com,
> 	gdb-patches@sourceware.org, torbjorn.svensson@foss.st.com
> 
> > What is the decision regarding target-async support on MS-Windows?
> > Will it be in GDB 13 or will it be deferred to GDB 14?
> 
> Pending any new issues found with this feature, I think there is a very
> good chance that it'll go in GDB 13, thanks to Tom having sent some
> patches to fix the issue he's found (interrupting programs).

But it wasn't committed yet, right?  And you say you will branch
tomorrow?

Re: [branching soon!] GDB 13 release -- 2022-12-04 update

[Joel Brobecker]

> > Date: Fri, 9 Dec 2022 17:23:26 +0400
> > From: Joel Brobecker <brobecker@adacore.com>
> > Cc: Luis Machado <luis.machado@arm.com>, brobecker@adacore.com,
> > 	gdb-patches@sourceware.org, torbjorn.svensson@foss.st.com
> > 
> > > What is the decision regarding target-async support on MS-Windows?
> > > Will it be in GDB 13 or will it be deferred to GDB 14?
> > 
> > Pending any new issues found with this feature, I think there is a very
> > good chance that it'll go in GDB 13, thanks to Tom having sent some
> > patches to fix the issue he's found (interrupting programs).
> 
> But it wasn't committed yet, right?  And you say you will branch
> tomorrow?

It hasn't no. I'm planning to wait a few extra days if necessary
in order to give this fix a chance to get in on time. This will allow us
to avoid a pre-release with the feature disabled, only to perhaps
re-enable soon after.

-- 
Joel