GDB 9.1 release -- 2019-11-01 Update

GDB 9.1 release -- 2019-11-01 Update

[Joel Brobecker]

Hi everyone,

Let's do a quick update on where we are respective to the next release.

Objective #1 is to branch. Before we do so, we need to identify
the list of changes that are large and/or tricky enough that we'd
want those change to be done in master, prior to branching.

Below is the list of patches that I know about. If your name is
next to it, it means you're the patch's champion ;-).

If there are others you think need to be pushed before we branch,
please let me know. Another way to catch my attention is to use PRs
with the target milestone being set to "9.1".

Still Pending:
--------------

  - [Tom DV] Only force INTERP_CONSOLE ui_out for breakpoint commands in MI mode
    https://sourceware.org/ml/gdb-patches/2019-10/msg00099.html
    Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/28

    Pedro gave it the OK, but mentioned that he'd like Andrew to review.

  - [Tom DV] Prefer var def over decl
    https://sourceware.org/ml/gdb-patches/2019-09/msg00161.html
    Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/29

    Tom T said the patch is reasonable, but was wondering whether
    we shouldn't even be in that situation in the first place, if
    we were to not create symbols for declarations...

  - [Christian B] The threaded symbol loading that tromey Christian have been working on
    Missing an URL to the patch, so can't check progress

Completed since the last update (2019-10-12):
---------------------------------------------

  - [Christian B] Load system gdbinit files from a directory
    https://sourceware.org/ml/gdb-patches/2019-10/msg00125.html
    https://www.sourceware.org/ml/gdb-patches/2019-10/msg01035.html

  - [Tom DV] Fix 'Unexpected register class' assert in amd64_push_arguments
    https://sourceware.org/ml/gdb-patches/2019-10/msg00293.html
    https://www.sourceware.org/ml/gdb-cvs/2019-10/msg00130.html

  - [Tom DV] Fix inferior call arg passing for amd64
    https://sourceware.org/ml/gdb-patches/2019-10/msg00307.html
    https://www.sourceware.org/ml/gdb-cvs/2019-10/msg00131.html

Thank you!
-- 
Joel

Re: GDB 9.1 release -- 2019-11-01 Update

[Philippe Waroquiers]

On Sun, 2019-11-03 at 17:16 -0800, Joel Brobecker wrote:
> Completed since the last update (2019-10-12):
> ---------------------------------------------
Note that I pushed 
https://sourceware.org/ml/gdb-patches/2019-10/msg01123.html
Convenience functions $_gdb_setting/$_gdb_setting_str.


If there is some more review bandwidth before 9.1, here are the list
of remaining pending patches:


RFC Have an option to tell GDB to detect and possibly handle mismatched exec-files
https://sourceware.org/ml/gdb-patches/2019-09/msg00580.html

Allow the user to define default leading args for commands and aliases
https://sourceware.org/ml/gdb-patches/2019-09/msg00583.html

Implement 'print -raw-values' and 'set print raw-values on|off'
https://sourceware.org/ml/gdb-patches/2019-09/msg00582.html

More flexible user-defined commands prefixing and naming.
https://sourceware.org/ml/gdb-patches/2019-09/msg00588.html

Thanks

Philippe

Re: GDB 9.1 release -- 2019-11-01 Update

[Jonah Graham]

On Sun., Nov. 3, 2019, 20:16 Joel Brobecker, <brobecker@adacore.com> wrote:

> Hi everyone,
>
> Let's do a quick update on where we are respective to the next release.
>
>
Hi Joel,

Can an update be done to the schedule (current/next) at the top of
https://www.gnu.org/software/gdb/schedule/ please?

Also, as an irregular follower of the list and got commits I have noticed
that we are jumping to 9.x, but is there a 9.0 release, or is the next
release 9.1?

Thanks,
Jonah


https://www.gnu.org/software/gdb/schedule/
>

Re: GDB 9.1 release -- 2019-11-01 Update

[Christian Biesinger via gdb-patches]

On Sun, Nov 3, 2019 at 7:16 PM Joel Brobecker <brobecker@adacore.com> wrote:
> Still Pending:
> --------------
[...]
>   - [Christian B] The threaded symbol loading that tromey Christian have been working on
>     Missing an URL to the patch, so can't check progress

Sorry about that, this is the patchset ending in:
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/176

(Ideally my two followup patches as well; not necessarily the RFC one
at the top)

Christian

Re: GDB 9.1 release -- 2019-11-01 Update

[Tom Tromey]

>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:

Joel> Objective #1 is to branch. Before we do so, we need to identify
Joel> the list of changes that are large and/or tricky enough that we'd
Joel> want those change to be done in master, prior to branching.

Joel> Below is the list of patches that I know about. If your name is
Joel> next to it, it means you're the patch's champion ;-).

I have a lot of patches pending in gerrit.  Most of them aren't really
important for any particular release, though this one is regression fix:

https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/371

Whether it goes in before or after the branch is made isn't too
important to me.


Plus, there's the series to fix the windows-nat crash.

Tom

Re: GDB 9.1 release -- 2019-11-01 Update

[Simon Marchi]

On 2019-11-05 3:54 p.m., Tom Tromey wrote:
>>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:
> 
> Joel> Objective #1 is to branch. Before we do so, we need to identify
> Joel> the list of changes that are large and/or tricky enough that we'd
> Joel> want those change to be done in master, prior to branching.
> 
> Joel> Below is the list of patches that I know about. If your name is
> Joel> next to it, it means you're the patch's champion ;-).
> 
> I have a lot of patches pending in gerrit.  Most of them aren't really
> important for any particular release, though this one is regression fix:
> 
> https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/371
> 
> Whether it goes in before or after the branch is made isn't too
> important to me.
> 
> 
> Plus, there's the series to fix the windows-nat crash.
> 
> Tom
> 

To help you track the patches that are on Gerrit that people would like to get
in 9.1, we have tagged them with the hashtag "for-9.1":

  https://gnutoolchain-gerrit.osci.io/r/q/hashtag:for-9.1

We have only tagged the tip of the series, so in reality it includes the
tagged patch and those it depends on.

Simon

Re: GDB 9.1 release -- 2019-11-01 Update

[Andrew Burgess]

* Joel Brobecker <brobecker@adacore.com> [2019-11-03 17:16:26 -0800]:

> If there are others you think need to be pushed before we branch,
> please let me know. Another way to catch my attention is to use PRs
> with the target milestone being set to "9.1".

I'd was hoping to get these merged into 9.1 as they provide MI access
to some new CLI commands:

  https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/264
  https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/265
  https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/266
  https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/267
  https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/268
  https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/269

Most of these have already had an earlier versions reviewed, so
hopefully there isn't much more that needs doing.

Also this fixes a regressions from earlier versions:

  https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/511

but this can easily be merged into the release branch later.

thanks,
Andrew

Re: GDB 9.1 release -- 2019-11-01 Update

[Joel Brobecker]

Hi Jonah,

> Can an update be done to the schedule (current/next) at the top of
> https://www.gnu.org/software/gdb/schedule/ please?

Yes, good point. I somehow thought that I had already done that before,
but clearly not. I took this opportunity to include the schedule for
GDB 10 as well, even though it's a bit early to think about it. It's
only tentative, and it's based on making major releases of GDB twice
yearly, minus a month to help avoid slippage over time.

> Also, as an irregular follower of the list and got commits I have noticed
> that we are jumping to 9.x, but is there a 9.0 release, or is the next
> release 9.1?

The release numbering is changing starting with this release cycle.
The 9.0[.x] numbering will be used for the pre-releases, and the first
official release will be 9.1, followed by the corrective release 9.2.

Each new release cycle will then increase the major version, and follow
the same logic as above.

The following wiki page summarizes the process (skip to the diagram
at the end):

    https://sourceware.org/gdb/wiki/Internals%20Versions

-- 
Joel

Re: GDB 9.1 release -- 2019-11-01 Update

[Joel Brobecker]

Hi Philippe,

> If there is some more review bandwidth before 9.1, here are the list
> of remaining pending patches:
> 
> 
> RFC Have an option to tell GDB to detect and possibly handle mismatched exec-files
> https://sourceware.org/ml/gdb-patches/2019-09/msg00580.html
> 
> Allow the user to define default leading args for commands and aliases
> https://sourceware.org/ml/gdb-patches/2019-09/msg00583.html
> 
> Implement 'print -raw-values' and 'set print raw-values on|off'
> https://sourceware.org/ml/gdb-patches/2019-09/msg00582.html
> 
> More flexible user-defined commands prefixing and naming.
> https://sourceware.org/ml/gdb-patches/2019-09/msg00588.html

I've added those to the list I'll track as, unlike some other patches
that others highlighted, these don't seem overly risky for landing
just before branching, or being backported.

-- 
Joel

Re: GDB 9.1 release -- 2019-11-01 Update

[Joel Brobecker]

> >   - [Christian B] The threaded symbol loading that tromey Christian have been working on
> >     Missing an URL to the patch, so can't check progress
> 
> Sorry about that, this is the patchset ending in:
> https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/176
> 
> (Ideally my two followup patches as well; not necessarily the RFC one
> at the top)

Thanks for the link, Christian. I looked at it, and as I mentioned
in today's update
(https://www.sourceware.org/ml/gdb-patches/2019-11/msg00293.html)
I am proposing we do _not_ include this in 9.1 -- seems a bit risky
to be making a change of this kind just prior to release...

-- 
Joel

Re: GDB 9.1 release -- 2019-11-01 Update

[Joel Brobecker]

Hi Andrew,

On Sun, Nov 10, 2019 at 10:40:10PM +0000, Andrew Burgess wrote:
> * Joel Brobecker <brobecker@adacore.com> [2019-11-03 17:16:26 -0800]:
> 
> > If there are others you think need to be pushed before we branch,
> > please let me know. Another way to catch my attention is to use PRs
> > with the target milestone being set to "9.1".
> 
> I'd was hoping to get these merged into 9.1 as they provide MI access
> to some new CLI commands:
> 
>   https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/264
>   https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/265
>   https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/266
>   https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/267
>   https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/268
>   https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/269
> 
> Most of these have already had an earlier versions reviewed, so
> hopefully there isn't much more that needs doing.

Those indeed look OK to me to include late in the game. So I've
added them to the list of patches to track (but in the "if done
in time, category, although I have no doubt it'll be done in time;
we can wait an extra week or two if it helps).

> Also this fixes a regressions from earlier versions:
> 
>   https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/511
> 
> but this can easily be merged into the release branch later.

Thanks. Added to the list as well, but in the "should have"
category.

Thanks again!
-- 
Joel

Re: GDB 9.1 release -- 2019-11-01 Update

[Christian Biesinger via gdb-patches]

On Mon, Nov 11, 2019 at 9:48 AM Joel Brobecker <brobecker@adacore.com> wrote:
>
> > >   - [Christian B] The threaded symbol loading that tromey Christian have been working on
> > >     Missing an URL to the patch, so can't check progress
> >
> > Sorry about that, this is the patchset ending in:
> > https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/176
> >
> > (Ideally my two followup patches as well; not necessarily the RFC one
> > at the top)
>
> Thanks for the link, Christian. I looked at it, and as I mentioned
> in today's update
> (https://www.sourceware.org/ml/gdb-patches/2019-11/msg00293.html)
> I am proposing we do _not_ include this in 9.1 -- seems a bit risky
> to be making a change of this kind just prior to release...

I was hoping that it would land sooner :(

However, I would like to make a proposal? What if we push it with a
setting to turn it off by default, so that people who want can turn it
on sooner? It is a massive performance benefit for big programs.

Christian

Re: GDB 9.1 release -- 2019-11-01 Update

[Joel Brobecker]

> > > >   - [Christian B] The threaded symbol loading that tromey Christian have been working on
> > > >     Missing an URL to the patch, so can't check progress
> > >
> > > Sorry about that, this is the patchset ending in:
> > > https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/176
> > >
> > > (Ideally my two followup patches as well; not necessarily the RFC one
> > > at the top)
> >
> > Thanks for the link, Christian. I looked at it, and as I mentioned
> > in today's update
> > (https://www.sourceware.org/ml/gdb-patches/2019-11/msg00293.html)
> > I am proposing we do _not_ include this in 9.1 -- seems a bit risky
> > to be making a change of this kind just prior to release...
> 
> I was hoping that it would land sooner :(
> 
> However, I would like to make a proposal? What if we push it with a
> setting to turn it off by default, so that people who want can turn it
> on sooner? It is a massive performance benefit for big programs.

That would indeed be a very good compromise!

-- 
Joel

Re: GDB 9.1 release -- 2019-11-01 Update

[Christian Biesinger via gdb-patches]

On Thu, Nov 14, 2019 at 8:17 PM Joel Brobecker <brobecker@adacore.com> wrote:
>
> > > > >   - [Christian B] The threaded symbol loading that tromey Christian have been working on
> > > > >     Missing an URL to the patch, so can't check progress
> > > >
> > > > Sorry about that, this is the patchset ending in:
> > > > https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/176
> > > >
> > > > (Ideally my two followup patches as well; not necessarily the RFC one
> > > > at the top)
> > >
> > > Thanks for the link, Christian. I looked at it, and as I mentioned
> > > in today's update
> > > (https://www.sourceware.org/ml/gdb-patches/2019-11/msg00293.html)
> > > I am proposing we do _not_ include this in 9.1 -- seems a bit risky
> > > to be making a change of this kind just prior to release...
> >
> > I was hoping that it would land sooner :(
> >
> > However, I would like to make a proposal? What if we push it with a
> > setting to turn it off by default, so that people who want can turn it
> > on sooner? It is a massive performance benefit for big programs.
>
> That would indeed be a very good compromise!

OK, I pushed a new patch to gerrit that implements that:
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/687

Christian