GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Joel Brobecker]

Hello everyone,

It's been a long time coming, and I think we're at a point where
we can branch Real Soon Now.

At this point, I have removed the section "if reviewed in time".
We do have a couple of patches identified as needed, but one is
about to be pushed, while the other has comments which don't seem
super hard to address -- and the patch itself looks sufficiently
small that it can be backported post branch creation.

Therefore, unless there are issues that should delay the creation
of the branch (please let us know!), I propose creating that branch on:

        Wednesday Dec 11th, sometime in the evening of Western Europe

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

  - [TomT/PedroA] Handle pending stops from the Windows kernel
    Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/414

    Add pending stop support to gdbserver's Windows port
    Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/432

        The first patch has been approved, and thus should get pushed
        fairly quickly.

        The second has been reviewed, and changes have been requested.
        Without looking too much into details, the comments don't seem
        too hard to address.


Completed since the last update (2019-11-24):
---------------------------------------------

  - [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

  - [Christian B] The threaded symbol loading that tromey Christian have been working on
    Gerrit: https://gnutoolchain-gerrit.osci.io/r/q/topic:%22minsyms-threads%22+(status:open%20OR%20status:merged)

-- 
Joel

Re: GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Philippe Waroquiers]

On Sat, 2019-12-07 at 20:05 -0500, Joel Brobecker wrote:

> Completed since the last update (2019-11-24):
> ---------------------------------------------

>   - [PhilippeW] More flexible user-defined commands prefixing and naming
>     https://sourceware.org/ml/gdb-patches/2019-09/msg00588.html
has been pushed the 30 of November.

while the below should be ok to go in before Wednedsday, as the only thing still
to review are 3 additional gdb_test in gdb.python/py-prettyprint.exp
>   - [PhilippeW] Implement 'print -raw-values' and 'set print raw-values on|off'
>     https://sourceware.org/ml/gdb-patches/2019-09/msg00582.html

Thanks
Philippe

Re: GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Tom Tromey]

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

Joel>   - [TomT/PedroA] Handle pending stops from the Windows kernel
Joel>     Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/414

Joel>     Add pending stop support to gdbserver's Windows port
Joel>     Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/432

Joel>         The first patch has been approved, and thus should get pushed
Joel>         fairly quickly.

I think these should be dropped from the release.  They've required
fairly extensive rewriting.  Also, I haven't completed the rewrites from
the last round of reviews, and due to work stuff I'm not likely to in
the next couple of weeks.

Tom

Re: GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Joel Brobecker]

Hi Tom,

> Joel>   - [TomT/PedroA] Handle pending stops from the Windows kernel
> Joel>     Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/414
> 
> Joel>     Add pending stop support to gdbserver's Windows port
> Joel>     Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/432
> 
> Joel>         The first patch has been approved, and thus should get pushed
> Joel>         fairly quickly.
> 
> I think these should be dropped from the release.  They've required
> fairly extensive rewriting.  Also, I haven't completed the rewrites from
> the last round of reviews, and due to work stuff I'm not likely to in
> the next couple of weeks.

Thanks for the update on those. It seems wise to drop them indeed.

-- 
Joel

Re: GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Eli Zaretskii]

> Date: Thu, 12 Dec 2019 23:36:01 +0100
> From: Joel Brobecker <brobecker@adacore.com>
> Cc: gdb-patches@sourceware.org
> 
> > I think these should be dropped from the release.  They've required
> > fairly extensive rewriting.  Also, I haven't completed the rewrites from
> > the last round of reviews, and due to work stuff I'm not likely to in
> > the next couple of weeks.
> 
> Thanks for the update on those. It seems wise to drop them indeed.

Does this mean the pretest published yesterday doesn't represent the
upcoming release well enough, as these changesets are still part of
it?  If so, can we please have a better pretest soon?  I planned on
building the pretest on MinGW soon, as I've seen many changes that
might "need work" in the MinGW port.

TIA

Re: GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Joel Brobecker]

> > > I think these should be dropped from the release.  They've required
> > > fairly extensive rewriting.  Also, I haven't completed the rewrites from
> > > the last round of reviews, and due to work stuff I'm not likely to in
> > > the next couple of weeks.
> > 
> > Thanks for the update on those. It seems wise to drop them indeed.
> 
> Does this mean the pretest published yesterday doesn't represent the
> upcoming release well enough, as these changesets are still part of
> it?  If so, can we please have a better pretest soon?  I planned on
> building the pretest on MinGW soon, as I've seen many changes that
> might "need work" in the MinGW port.

What we mean is that we we will not be including these patches in
the GDB 9.1 release. So you can pre-test this pre-release. At
this point, we no longer have any issue that's been identified
as being release blocking (famous last words, as this is the point
when problems start show up).

One thing I'd like to suggest is if you could do a first pre-check
when I announce that the branch creation is a few days away.
If you find something sufficiently bad, this would be grounds
for me to either hold the branch creation, or else hold the
creation of the pre-release.  I don't who else could help with that,
because I don't anyone but you building GDB with the configuration
that you use -- I know we build a MinGW version of GDB on Windows,
but we use MinGW64 and do not see the issues that you have been
reporting in the past -- so aside from you, I don't know who else
can do it.

-- 
Joel

Re: GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Eli Zaretskii]

> Date: Sat, 14 Dec 2019 00:47:44 +0100
> From: Joel Brobecker <brobecker@adacore.com>
> Cc: tromey@adacore.com, gdb-patches@sourceware.org
> 
> > Does this mean the pretest published yesterday doesn't represent the
> > upcoming release well enough, as these changesets are still part of
> > it?  If so, can we please have a better pretest soon?  I planned on
> > building the pretest on MinGW soon, as I've seen many changes that
> > might "need work" in the MinGW port.
> 
> What we mean is that we we will not be including these patches in
> the GDB 9.1 release.

But they are included in the pretest tarball?

> One thing I'd like to suggest is if you could do a first pre-check
> when I announce that the branch creation is a few days away.
> If you find something sufficiently bad, this would be grounds
> for me to either hold the branch creation, or else hold the
> creation of the pre-release.  I don't who else could help with that,
> because I don't anyone but you building GDB with the configuration
> that you use -- I know we build a MinGW version of GDB on Windows,
> but we use MinGW64 and do not see the issues that you have been
> reporting in the past -- so aside from you, I don't know who else
> can do it.

Thanks, I'm okay with testing the first pretest, and fixing any
problems in the release branch.  I was asking whether the current
pretest is a good approximation for the release, as I don't enjoy
solving problems I don't need to solve, and Gnulib has been
historically an important source of problems for MinGW.

If there are significant changes (not bugfixes) expected before the
release, I'd prefer a pretest after those changes.  I'm not sure I
understand what was the decision regarding the specific issue
mentioned above, if there was a decision.  Am I missing something.

Re: GDB 9.1 Release: Creating the branch on *WED* Dec 11th!

[Joel Brobecker]

> > > Does this mean the pretest published yesterday doesn't represent the
> > > upcoming release well enough, as these changesets are still part of
> > > it?  If so, can we please have a better pretest soon?  I planned on
> > > building the pretest on MinGW soon, as I've seen many changes that
> > > might "need work" in the MinGW port.
> > 
> > What we mean is that we we will not be including these patches in
> > the GDB 9.1 release.
> 
> But they are included in the pretest tarball?

They are not going to be included at all, neither in the pretest,
nor in the release. Basically, this patch was identified as a candidate
for having it in the 9.1 release. However, upon realizing what the patch
entails, it was deemed to disruptive to be considered for inclusion in
the GDB 9 release cycle.

> Thanks, I'm okay with testing the first pretest, and fixing any
> problems in the release branch.  I was asking whether the current
> pretest is a good approximation for the release, as I don't enjoy
> solving problems I don't need to solve, and Gnulib has been
> historically an important source of problems for MinGW.
> 
> If there are significant changes (not bugfixes) expected before the
> release, I'd prefer a pretest after those changes.  I'm not sure I
> understand what was the decision regarding the specific issue
> mentioned above, if there was a decision.  Am I missing something.

I believe you can do your pretest now. As far as I know, there aren't
any patches known to be pending (we cleared the "want" list prior to
creating the branch).

-- 
Joel