Start 6.1 release process

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

Start 6.1 release process

Rik-4
Could someone merge the default branch on to the stable branch so that the
6.1 release can start?  While working on bug #57799 (mkdir and rmdir always
return status output), I realized that there is a common code pattern which
is not Matlab compatible.  Fixing it will hit more than 6 functions and
that in turn will require changes to many other files which use these
functions; fixing rmdir alone required changes to 15 other files.  This
feels like too much change immediately before a release so I want to
continue this work, but on the development branch that will become 7.1.

Thanks,
Rik

Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

John W. Eaton
Administrator
On 2/16/20 10:51 AM, Rik wrote:
> Could someone merge the default branch on to the stable branch so that the
> 6.1 release can start?  While working on bug #57799 (mkdir and rmdir always
> return status output), I realized that there is a common code pattern which
> is not Matlab compatible.  Fixing it will hit more than 6 functions and
> that in turn will require changes to many other files which use these
> functions; fixing rmdir alone required changes to 15 other files.  This
> feels like too much change immediately before a release so I want to
> continue this work, but on the development branch that will become 7.1.

Seems good to me.  I also have some pending changes that are probably
best done early in the development cycle instead of just before a
release.  I will merge and update version numbers, buildbot, etc.

jwe


Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

Rik-4
On 02/17/2020 09:33 AM, John W. Eaton wrote:

> On 2/16/20 10:51 AM, Rik wrote:
>> Could someone merge the default branch on to the stable branch so that the
>> 6.1 release can start?  While working on bug #57799 (mkdir and rmdir always
>> return status output), I realized that there is a common code pattern which
>> is not Matlab compatible.  Fixing it will hit more than 6 functions and
>> that in turn will require changes to many other files which use these
>> functions; fixing rmdir alone required changes to 15 other files.  This
>> feels like too much change immediately before a release so I want to
>> continue this work, but on the development branch that will become 7.1.
>
> Seems good to me.  I also have some pending changes that are probably
> best done early in the development cycle instead of just before a
> release.  I will merge and update version numbers, buildbot, etc.
>
Thanks for the series of changesets.  Do we need to update API_VERSION in
configure.ac as well?  I'm wondering after looking at "Versioning Hints" at
the bottom of https://wiki.octave.org/6.1_Release_Checklist.  Maybe also
the library version numbers?

--Rik


Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

John W. Eaton
Administrator
On 2/17/20 3:17 PM, Rik wrote:

> Thanks for the series of changesets.  Do we need to update API_VERSION in
> configure.ac as well?  I'm wondering after looking at "Versioning Hints" at
> the bottom of https://wiki.octave.org/6.1_Release_Checklist.  Maybe also
> the library version numbers?

Is the following the correct thing to do?

   * on stable, update to api-v54

   * on default update to api-54+

   * on stable, just before release, update library versions, merge
stable to default

If so, I can do that.

Also, after

   changeset:   28087:5e79401cbc35
   branch:      stable
   parent:      28085:a3b40e48c069
   user:        Rik <[hidden email]>
   date:        Mon Feb 17 12:31:16 2020 -0800
   summary:     Set default "paperposition" property to correct value
(bug #57651)

I'm seeing three test failures in orient.m and I'm not sure what the
correct fix is.

jwe




Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

Mike Miller-4
On Mon, Feb 17, 2020 at 16:48:14 -0500, John W. Eaton wrote:
>   * on stable, just before release, update library versions, merge stable to
> default

The Debian maintainers, and probably other distro packagers, would be
very happy if the library sonames were already updated, at the very
latest before making the first release candidate.

--
mike

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

John W. Eaton
Administrator
On 2/17/20 4:54 PM, Mike Miller wrote:
> On Mon, Feb 17, 2020 at 16:48:14 -0500, John W. Eaton wrote:
>>    * on stable, just before release, update library versions, merge stable to
>> default
>
> The Debian maintainers, and probably other distro packagers, would be
> very happy if the library sonames were already updated, at the very
> latest before making the first release candidate.

OK.

jwe


Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

Rik-4
In reply to this post by John W. Eaton
On 02/17/2020 01:48 PM, John W. Eaton wrote:

> On 2/17/20 3:17 PM, Rik wrote:
>
>> Thanks for the series of changesets.  Do we need to update API_VERSION in
>> configure.ac as well?  I'm wondering after looking at "Versioning Hints" at
>> the bottom of https://wiki.octave.org/6.1_Release_Checklist.  Maybe also
>> the library version numbers?
>
> Is the following the correct thing to do?
>
>   * on stable, update to api-v54
>
>   * on default update to api-54+
>
>   * on stable, just before release, update library versions, merge stable
> to default
>
> If so, I can do that.

Looks right to me.

>
> Also, after
>
>   changeset:   28087:5e79401cbc35
>   branch:      stable
>   parent:      28085:a3b40e48c069
>   user:        Rik <[hidden email]>
>   date:        Mon Feb 17 12:31:16 2020 -0800
>   summary:     Set default "paperposition" property to correct value (bug
> #57651)
>
> I'm seeing three test failures in orient.m and I'm not sure what the
> correct fix is.
>

Running 'test orient' by itself doesn't fail for me.  I'll try running
'make check' and see if it is due to prior initialization.

--Rik

Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

chris_frick
Dear all

What's the state of the 6.1 release? Is there already a RC to test for
Windows?

Best regards
Christian



--
Sent from: https://octave.1599824.n4.nabble.com/Octave-Maintainers-f1638794.html

Reply | Threaded
Open this post in threaded view
|

Re: Start 6.1 release process

mmuetzel