[bug#34378] update GNOME Planner to current

[Andrew Miloradovsky <andrew@interpretmath.pw>]

Hi Ludovic,

On 2/12/19 9:58 PM, Ludovic Courtès wrote:
> Hi Andrew,
>
> Andrew Miloradovsky <andrew@interpretmath.pw> skribis:
>
>> The release is fairly old and contains prebuilt scripts.
> What do you mean by “prebuilt scripts”?

By pre-built scripts I mean things like `configure`, generated by
auto-tools and gnome-common.

>> From cbe5c7a354381c665735085a5eb262404e901385 Mon Sep 17 00:00:00 2001
>> From: Andrew Miloradovsky <andrew@interpretmath.pw>
>> Date: Fri, 8 Feb 2019 00:03:48 +0000
>> Subject: [PATCH] gnu: gnome-planner: update 0.14.6 to 2019-02-08
>>
>> Build all from a modern version, enable Python bindings.
>>
>> * gnu/packages/gnome.scm (gnome-planner):
>>   - Fetch the sources from GNOME's own GitLab.
>>   - Run `autogen.sh` with the proper shell paths.
>>   - Enable Python 2 bindings, because we can now.
>>   - Add all the inputs necessary for all of that.
> Please see ‘git log’ and
> <https://www.gnu.org/software/guix/manual/en/html_node/Submitting-Patches.html>
> for examples of the commit message format.

The changes made didn't fit well in the recommended commit message
structure (documenting the changes to each field separately).

>> +  ;;; Warning:
>> +  ;;; The documentation generation mechanism is long broken. No Yelp yet.
>> +  ;;; So the output doesn't contain any docs, unlike the ancient release.
>> +  ;;; OTOH, that's probably not a huge concern, given the circumstances.
>> +  (let ((commit "fa7cbe309d5a705502ca46f808bcf78840804dbe")
>> +        (revision "2019-02-08"))
> How did you pick this particular commit?  In general we only package
> releases, unless there’s a compelling reason to build straight from the
> VCS.  Do you think that’s the case here?

I picked that commit just because it was the most recent: the release is
expected to be very difficult to make, a lot of things need to be
fixed/modernized first, and there are not many people to review the
patches (one to be exact, AFAICT).

I'm not sure there are really compelling reasons: most/all of the
changes made since the release are either about translations
(documentation generation is broken anyways), or fixes of deprecation
warnings (which otherwise would prevent it from building in Nixpkgs with
the default -Werror).

>
> Thanks for the patch!
>
> Ludo’.