GDB 7.2 release date

GDB 7.2 release date

[Joel Brobecker]

Dear all,

It is time to decide what day we are targeting for creating the GDB 7.2
release. Please let me know what issues are still remaining on the 7.2
branch, and when you expect them to be fixed. I am only concerned about
critical bugs, regressions, etc.

Ideally, my best day would be tomorrow (Thu), after which I'll be on and
off the office again for the next 10 days. But that might be too tight,
and I'll just try to find some time to make the release during the "on"
time.

-- 
Joel

Re: GDB 7.2 release date

[Pedro Alves]

On Wednesday 28 July 2010 18:50:11, Joel Brobecker wrote:
> It is time to decide what day we are targeting for creating the GDB 7.2
> release. Please let me know what issues are still remaining on the 7.2
> branch, and when you expect them to be fixed. I am only concerned about
> critical bugs, regressions, etc.

A few new commands and remote protocol packets / features made it into
the tree without NEWS entries.  I think we should fix that.

-- 
Pedro Alves

Re: GDB 7.2 release date

[Joel Brobecker]

> A few new commands and remote protocol packets / features made it into
> the tree without NEWS entries.  I think we should fix that.

Agreed. Do you have the list of features that are still missing
documentation, or is it something that needs to be researched?

-- 
Joel

Re: GDB 7.2 release date

[Pedro Alves]

On Wednesday 28 July 2010 23:02:01, Joel Brobecker wrote:
> > A few new commands and remote protocol packets / features made it into
> > the tree without NEWS entries.  I think we should fix that.
> 
> Agreed. Do you have the list of features that are still missing
> documentation, or is it something that needs to be researched?

The latter, I'm afraid.  Out of memory, for example, the observer
mode, and its commands and packets aren't mentioned in NEWS.  I'll
start by mentioning this on the wiki.

-- 
Pedro Alves

Re: GDB 7.2 release date

[Michael Snyder]

What's the current thinking about a release date?

Re: GDB 7.2 release date

[Joel Brobecker]

> What's the current thinking about a release date?

The problem is that we have identified issues with several of us think
are blocking for the release. Until someone has time to step in (holiday
season) and fix them, the release will continue to be delayed.

-- 
Joel

Re: GDB 7.2 release date

[Jack Howarth]

On Mon, Aug 02, 2010 at 10:27:00AM -0700, Joel Brobecker wrote:
> > What's the current thinking about a release date?
> 
> The problem is that we have identified issues with several of us think
> are blocking for the release. Until someone has time to step in (holiday
> season) and fix them, the release will continue to be delayed.
> 
> -- 
> Joel

Joel,
   I posted the testresults for the current 7.2 release snapshot
on http://sourceware.org/ml/gdb-testers/2010-q3/msg00000.html for
x86_64-apple-darwin10. How do those results compare with x86_64-*-linux?
        Jack

Re: GDB 7.2 release date

[Joel Brobecker]

>    I posted the testresults for the current 7.2 release snapshot
> on http://sourceware.org/ml/gdb-testers/2010-q3/msg00000.html for
> x86_64-apple-darwin10. How do those results compare with x86_64-*-linux?

I could run the testsuite on my system and send you the results if you
would like to compare. Typically, though, linux systems enjoy much
better results that darwin.

-- 
Joel

Re: GDB 7.2 release date

[Jack Howarth]

On Mon, Aug 02, 2010 at 10:39:58AM -0700, Joel Brobecker wrote:
> >    I posted the testresults for the current 7.2 release snapshot
> > on http://sourceware.org/ml/gdb-testers/2010-q3/msg00000.html for
> > x86_64-apple-darwin10. How do those results compare with x86_64-*-linux?
> 
> I could run the testsuite on my system and send you the results if you
> would like to compare. Typically, though, linux systems enjoy much
> better results that darwin.

Joel,
   Better yet could you just post the results to gdb-testers? I was 
surprised to see that there isn't the same sort of regress testers
running to post regularly to that list like on FSF gcc.
            Jack
> 
> -- 
> Joel

Re: GDB 7.2 release date

[Michael Snyder]

Joel Brobecker wrote:
>> What's the current thinking about a release date?
> 
> The problem is that we have identified issues with several of us think
> are blocking for the release. Until someone has time to step in (holiday
> season) and fix them, the release will continue to be delayed.

Thanks -- I just needed to know for planning purposes.

Re: GDB 7.2 release date

[Joel Brobecker]

(apologies to everyone for dropping the ball a bit on the release)

> The latter, I'm afraid.  Out of memory, for example, the observer
> mode, and its commands and packets aren't mentioned in NEWS.  I'll
> start by mentioning this on the wiki.

I admit I scanned the emails for the past week a little fast. But I do
not remember seeing any progress on this topic. Is there someone who
is supposed to look at that, or is this going to be on my plate?

-- 
Joel

Re: GDB 7.2 release date

[Pedro Alves]

On Monday 16 August 2010 16:15:13, Joel Brobecker wrote:
> (apologies to everyone for dropping the ball a bit on the release)
> 
> > The latter, I'm afraid.  Out of memory, for example, the observer
> > mode, and its commands and packets aren't mentioned in NEWS.  I'll
> > start by mentioning this on the wiki.
> 
> I admit I scanned the emails for the past week a little fast. But I do
> not remember seeing any progress on this topic. Is there someone who
> is supposed to look at that, or is this going to be on my plate?

I was kind of hoping that patches would have appeared, avoiding the need for
an audit, but that was wishful thinking.  :-)  I'll try to do this, but
probably not today.

-- 
Pedro Alves

Re: GDB 7.2 release date

[Mathew Yeates]

Will someone please look at my gfortran problem where the nsym != NULL
assertion fails?

On Mon, Aug 16, 2010 at 8:15 AM, Joel Brobecker <brobecker@adacore.com> wrote:
> (apologies to everyone for dropping the ball a bit on the release)
>
>> The latter, I'm afraid.  Out of memory, for example, the observer
>> mode, and its commands and packets aren't mentioned in NEWS.  I'll
>> start by mentioning this on the wiki.
>
> I admit I scanned the emails for the past week a little fast. But I do
> not remember seeing any progress on this topic. Is there someone who
> is supposed to look at that, or is this going to be on my plate?
>
> --
> Joel
>

Re: GDB 7.2 release date

[Stan Shebs]

Pedro Alves wrote:
> On Monday 16 August 2010 16:15:13, Joel Brobecker wrote:
>   
>> (apologies to everyone for dropping the ball a bit on the release)
>>
>>     
>>> The latter, I'm afraid.  Out of memory, for example, the observer
>>> mode, and its commands and packets aren't mentioned in NEWS.  I'll
>>> start by mentioning this on the wiki.
>>>       
>> I admit I scanned the emails for the past week a little fast. But I do
>> not remember seeing any progress on this topic. Is there someone who
>> is supposed to look at that, or is this going to be on my plate?
>>     
>
> I was kind of hoping that patches would have appeared, [...]
Is that shorthand for "I was hoping that Stan would do his job"?  :-)

I can certainly take care of the observer mode part at least.

Stan

Re: GDB 7.2 release date

[Joel Brobecker]

> I can certainly take care of the observer mode part at least.

Thanks! Anything else we think might have slipped in without
the necessary NEWS update? Apart from pure ChangeLog review, can we
verify this by just comparing the documentation changes vs the NEWS
files changes? (this assumes that all new packets were documented,
is that correct).

On my end, I will look at the -file-list-exec-source-files tomorrow
(BTW: I'm in Europe for the next 10 days), if no one else has a chance
to look at it by then.

-- 
Joel

Re: GDB 7.2 release date

[Joel Brobecker]

> Will someone please look at my gfortran problem where the nsym != NULL
> assertion fails?

ISTR that you had a few pointers provided by Tom, but apparently nothing
concrete enough to fix your issue... That's a bummer, but we do not
assign bugs. People are usually willing to help, but the amount of help
they are willing to provide depends on pretty much every conceivable
factor, including the phase of the moon.

Going back to the 7.2 release: I am currently not planning on holding
the release until this bug is fixed, especially if we don't have some
kind of estimate as to when this might be. We would need more information
regarding this bug before we consider it release-critical (Eg: is this
a regression, under what conditions does it appear, is this only with
Fortran, etc).

-- 
Joel

Re: GDB 7.2 release date

[Pedro Alves]

On Monday 16 August 2010 20:17:43, Joel Brobecker wrote:
> > I can certainly take care of the observer mode part at least.
> 
> Thanks! Anything else we think might have slipped in without
> the necessary NEWS update? Apart from pure ChangeLog review, can we
> verify this by just comparing the documentation changes vs the NEWS
> files changes? (this assumes that all new packets were documented,
> is that correct).

I've remembered "$_thread" as well.  IMO that's NEWS worthy.

And this one too (Tracepoint source strings feature and the
QTDPsrc packet & co):

 <http://sourceware.org/ml/gdb-patches/2010-03/msg01043.html>

:-)

-- 
Pedro Alves

Re: GDB 7.2 release date

[Tom Tromey]

>>>>> "Mathew" == Mathew Yeates <mat.yeates@gmail.com> writes:

Mathew> Will someone please look at my gfortran problem where the nsym
Mathew> != NULL assertion fails?

Could you try the patch in PR 11313 and say whether it works for you?

This isn't a real fix for the problem, but I think it would be an ok
workaround.

Tom

Re: GDB 7.2 release date

[Mathew Yeates]

Are you referring to the gfortran patch? I tried that and it didn't work.

On Mon, Aug 16, 2010 at 1:06 PM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> "Mathew" == Mathew Yeates <mat.yeates@gmail.com> writes:
>
> Mathew> Will someone please look at my gfortran problem where the nsym
> Mathew> != NULL assertion fails?
>
> Could you try the patch in PR 11313 and say whether it works for you?
>
> This isn't a real fix for the problem, but I think it would be an ok
> workaround.
>
> Tom
>

Re: GDB 7.2 release date

[Tom Tromey]

>>>>> "Mathew" == Mathew Yeates <mat.yeates@gmail.com> writes:

Mathew> Are you referring to the gfortran patch? I tried that and it
Mathew> didn't work.

No, the patch here:

http://sourceware.org/bugzilla/show_bug.cgi?id=%31%31%33%31%33#c5

Tom

Re: GDB 7.2 release date

[Mathew Yeates]

I'm confused. At the end of the post Randy says he can't print the
variable "str".
-----
(gdb) p str1
$1 = (REF TO -> ( character*16 )) @0x0: <error reading variable>
---------------

But, I'll try it if you think it will work

-Mathew

On Mon, Aug 16, 2010 at 1:22 PM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> "Mathew" == Mathew Yeates <mat.yeates@gmail.com> writes:
>
> Mathew> Are you referring to the gfortran patch? I tried that and it
> Mathew> didn't work.
>
> No, the patch here:
>
> http://sourceware.org/bugzilla/show_bug.cgi?id=%31%31%33%31%33#c5
>
> Tom
>

Re: GDB 7.2 release date

[Tom Tromey]

>>>>> "Mathew" == Mathew Yeates <mat.yeates@gmail.com> writes:

Mathew> I'm confused. At the end of the post Randy says he can't print the
Mathew> variable "str".
Mathew> -----
Mathew> (gdb) p str1
Mathew> $1 = (REF TO -> ( character*16 )) @0x0: <error reading variable>
Mathew> ---------------

Mathew> But, I'll try it if you think it will work

I think it will stop gdb from crashing.  I think this would be an
improvement over the current situation.

The 'str' thing may be a separate bug.  I haven't looked at that.

That PR concerns ifort, not gfortran, which IIRC is what you are using.
So it might be the case that this fix is all that you need.  Or, it may
be that this patch doesn't help you for some reason as yet unknown to
us.  That is what I would like to know.

Tom

Re: GDB 7.2 release date

[Mathew Yeates]

I don't crash anymore after the patch.

On Mon, Aug 16, 2010 at 1:50 PM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> "Mathew" == Mathew Yeates <mat.yeates@gmail.com> writes:
>
> Mathew> I'm confused. At the end of the post Randy says he can't print the
> Mathew> variable "str".
> Mathew> -----
> Mathew> (gdb) p str1
> Mathew> $1 = (REF TO -> ( character*16 )) @0x0: <error reading variable>
> Mathew> ---------------
>
> Mathew> But, I'll try it if you think it will work
>
> I think it will stop gdb from crashing.  I think this would be an
> improvement over the current situation.
>
> The 'str' thing may be a separate bug.  I haven't looked at that.
>
> That PR concerns ifort, not gfortran, which IIRC is what you are using.
> So it might be the case that this fix is all that you need.  Or, it may
> be that this patch doesn't help you for some reason as yet unknown to
> us.  That is what I would like to know.
>
> Tom
>

Re: GDB 7.2 release date

[Joel Brobecker]

> On my end, I will look at the -file-list-exec-source-files tomorrow
> (BTW: I'm in Europe for the next 10 days), if no one else has a chance
> to look at it by then.

Tomorrow turned out to be the day after tomorrow, but I'm happy to say
that I think I've got it. Currently testing the obvious fix right now.

Any news on the NEWS file ;-) ? I've let the release date slip a lot
longer than usual, and it's mostly my fault, but I am getting anxious
to get it out (if we wait that much longer, we might as well skip it
and start working on 7.3!!! [ahem, half joking])

-- 
Joel

Re: GDB 7.2 release date

[Stan Shebs]

Joel Brobecker wrote:
>> On my end, I will look at the -file-list-exec-source-files tomorrow
>> (BTW: I'm in Europe for the next 10 days), if no one else has a chance
>> to look at it by then.
>>     
>
> Tomorrow turned out to be the day after tomorrow, but I'm happy to say
> that I think I've got it. Currently testing the obvious fix right now.
>
> Any news on the NEWS file ;-) ? I've let the release date slip a lot
> longer than usual, and it's mostly my fault, but I am getting anxious
> to get it out (if we wait that much longer, we might as well skip it
> and start working on 7.3!!! [ahem, half joking])
>   

I'm grinding through doc/ChangeLog now, should have a patch by the end 
of the day.

Stan

Re: GDB 7.2 release date

[Joel Brobecker]

Following Stan's NEWS patch (thank you!), we should be pretty close to
being ready for the 7.2 release. I propose to create it either Friday
afternoon (UTC+2 - should be morning in the US), or sometime Monday if
we find more blocking issues.

BTW: I think that we're still missing NEWS entries for at least one
feature that Pedro identified... I will try to help with those.

-- 
Joel

Re: GDB 7.2 release date

[Pedro Alves]

On Thursday 19 August 2010 14:09:52, Joel Brobecker wrote:
> Following Stan's NEWS patch (thank you!), we should be pretty close to
> being ready for the 7.2 release. I propose to create it either Friday
> afternoon (UTC+2 - should be morning in the US), or sometime Monday if
> we find more blocking issues.
> 
> BTW: I think that we're still missing NEWS entries for at least one
> feature that Pedro identified... I will try to help with those.

I just diffed 7.1 vs 7.2, and grepped for things like

  grep "add_" ... | grep  "cmd"

I've documented in the wiki the new commands that are missing an
entry in news or in the manual.  Here's a copy of what I placed there:

    * Missing in NEWS:
          o [Stan]: Tracepoint source strings, observer mode, corresponding commands and packets. $_thread.
                  http://sourceware.org/ml/gdb-patches/2010-08/msg00325.html 

          o [Joel]: $_thread.

                  http://sourceware.org/ml/gdb-patches/2010-08/msg00353.html 

          o set|show record memory-query. 
    * Missing in NEWS and manual:
          o record pic
          o set|show record pic
          o set record pic hide-nofunction
          o set record pic hide-nosource
          o set record pic hide-sameaddr
          o set record pic type

            [Hui]: http://sourceware.org/ml/gdb-patches/2010-07/msg00033.html

It appeared to me that for most commands and features there are already
patches pending, except for "set|show record memory-query", if I'm not mistaken.
Hui, what's the status of that patch?

I've also diffed remote.c in both trees, and confirmed that all new
packets are either already mentioned in NEWS, or the patches pointed at
above already take care of mentioning them.

Joel, I think you missed that Stan had already documented $_thread
in his patch.  I've no idea if both descriptions could be merged so
we end up with a better description, or whether one of them ends
up redundant.

-- 
Pedro Alves

Re: GDB 7.2 release date

[Joel Brobecker]

>     * Missing in NEWS and manual:
>           o record pic
>           o set|show record pic
>           o set record pic hide-nofunction
>           o set record pic hide-nosource
>           o set record pic hide-sameaddr
>           o set record pic type
> 
>             [Hui]: http://sourceware.org/ml/gdb-patches/2010-07/msg00033.html

Eli, it looks like this patch might have fallen through the cracks?
Is Hui waiting for your review of the last iteration of his patch?

> It appeared to me that for most commands and features there are
> already patches pending, except for "set|show record memory-query",
> if I'm not mistaken.

I'm going to write the documentation for that patch.

-- 
Joel

Re: GDB 7.2 release date

[Joel Brobecker]

> > It appeared to me that for most commands and features there are
> > already patches pending, except for "set|show record memory-query",
> > if I'm not mistaken.
> 
> I'm going to write the documentation for that patch.

Once again, Stan beat me to it. Now who's the one who owes a drink???  ;-)

-- 
Joel

Re: GDB 7.2 release date

[Eli Zaretskii]

> Date: Fri, 20 Aug 2010 10:55:55 +0200
> From: Joel Brobecker <brobecker@adacore.com>
> Cc: gdb@sourceware.org
> 
> >     * Missing in NEWS and manual:
> >           o record pic
> >           o set|show record pic
> >           o set record pic hide-nofunction
> >           o set record pic hide-nosource
> >           o set record pic hide-sameaddr
> >           o set record pic type
> > 
> >             [Hui]: http://sourceware.org/ml/gdb-patches/2010-07/msg00033.html
> 
> Eli, it looks like this patch might have fallen through the cracks?
> Is Hui waiting for your review of the last iteration of his patch?

Yes, sorry.  The last patch was sitting in my INBOX.  This sometimes
happens with patches that need a lot of work; I need to be pinged if I
don't respond in a week or so.

I reviewed it now; it is still not ready for prime time.

Re: GDB 7.2 release date

[Joel Brobecker]

> I reviewed it now; it is still not ready for prime time.

Thanks for the quick review! Hui seems to be away or busy. I don't want
a documentation patch to continue delaying the release, so I will wait
until Tuesday morning to see if Hui might be back. Otherwise, I'll take
the documentation patch over and do my best at filling in the gaps
that you pointed out.

-- 
Joel