[GRASS-dev] [release planning] GRASS GIS 8: create new release branch

[Vaclav Petras]

On Tue, 8 Jun 2021 at 16:10, Veronica Andreo <veroandreo at gmail.com> wrote:

> As you all know, grass birthday is approaching... July 29th is just
around the corner.

I would say the master branch looks good in terms of fixes versus
potentially breaking commits (grass.jupyter and grass.benchmark will both
be marked as experimental in the release (grass.benchmark already is),
possibly not even advertised):

https://github.com/OSGeo/grass/commits/master

Perhaps a good time to create the release branch for 8.0? Opinions?

>From what has the milestone 8.0 and is not mentioned below and is major, I
see only #1454 and #1501 which are related to band references. Maybe we
should again consider reverting the band references in order to get 8.0 out
or how do you see the status of #1454 and #1501, Martin and Maris?

https://github.com/OSGeo/grass/milestone/4
https://github.com/OSGeo/grass/pull/1454
https://github.com/OSGeo/grass/pull/1501

See the status of what was discussed in this thread below.

On Wed, Jul 14, 2021 at 5:24 PM Vaclav Petras <wenzeslaus at gmail.com> wrote:

>
>
> On Sat, Jul 10, 2021 at 4:20 AM Maris Nartiss <maris.gis at gmail.com> wrote:
>
>> Hello all.
>>
>> 2021-07-09 21:39 GMT+03:00, Markus Neteler <neteler at osgeo.org>:
>>
>> > So, from my point of view, as you suggest:
>> >
>> > - get g.extension working, then
>>
>
> #1700 seems to be close to being finished. Please test.
>
> https://github.com/OSGeo/grass/pull/1700
>

Merged by Stefan Blumentrath 6 days ago. Seems to be working well.


>
>
>> There are a few PRs to merge but they are stuck due to failing tests –
>>
>
> I fixed the g.download.location part in #1715 and all or most "randomly"
> failing tests in #1681.
>
> https://github.com/OSGeo/grass/pull/1681
>

Both of these seem to be stable as well.


>
>
>> Windows build & test is just bad (no bug or PR so far?),
>
>
> Can someone give an update on that, please?
>

#1683 is not merged yet, but missing sh in the test environment is not a
blocker in this case.

https://github.com/OSGeo/grass/pull/1683


>
>
>> CentOS is too
>> old (#1709) & needs auto tools update to fix it (or at least env
>> variable in build script as I just have done to enable testing in
>> #1501),
>
>
> ...or CentOS 8, CentOS Stream, or remove the CentOS test.
>
> https://github.com/OSGeo/grass/issues/1709#issuecomment-880215427
>

The CI now contains environmental settings which are applicable to users
using older systems with old compilers. It would be better if it is
automatic, but it works now.

>
>
>
>> Ubuntu based test scripts use g.download.location extension
>> that is not operational at the moment (fixed by #1715?).
>>
>
> #1715 is merged, so the tests are fixed.
>
>
>>
>> I would also strongly advocate merging ctypes upgrade before 8.0
>>
>
> Seems to be close to being finished too.
>

Closer, but not quite finished yet. The not working piece seems to be just
the unusual use of NULL as a value for RASTER3D_DEFAULT_WINDOW.

https://github.com/OSGeo/grass/pull/1651

Best,
Vaclav
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-dev/attachments/20210727/d4eaed49/attachment.html>