Hackathon: fix build errors on the "master" branch.

Hackathon: fix build errors on the "master" branch.

[Ricardo Wurmus]

Hi Guix,

for the past weeks the "master" branch has been in pretty poor state
according to ci.guix.gnu.org.  It keeps hovering at around 56% progress,
which is a far cry from the 80+% we used to have in the old days.

A lot of packages are in fact broken.  Many more are marked as broken
due to a bug affecting ci.guix.gnu.org.

It would be a good idea to do a little spring cleaning and go through
the dashboard for the master branch and work on fixing builds that are
currently broken on x86_64-linux.  (Other architectures are in worse
state, but fixing x86_64-linux may be something we can actually
accomplish.)

What do you think?

-- 
Ricardo

Re: Hackathon: fix build errors on the "master" branch.

[Efraim Flashner]

[-- Attachment #1: Type: text/plain, Size: 1490 bytes --]

On Mon, Apr 29, 2024 at 04:52:29PM +0200, Ricardo Wurmus wrote:
> Hi Guix,
> 
> for the past weeks the "master" branch has been in pretty poor state
> according to ci.guix.gnu.org.  It keeps hovering at around 56% progress,
> which is a far cry from the 80+% we used to have in the old days.
> 
> A lot of packages are in fact broken.  Many more are marked as broken
> due to a bug affecting ci.guix.gnu.org.
> 
> It would be a good idea to do a little spring cleaning and go through
> the dashboard for the master branch and work on fixing builds that are
> currently broken on x86_64-linux.  (Other architectures are in worse
> state, but fixing x86_64-linux may be something we can actually
> accomplish.)
> 
> What do you think?

Using the QA¹ page we can see another data point of how the different
architectures are doing.  Also, assuming there are 30,000 packages, the
4.3% difference on bayfront between x86_64 and i686 would be 129
packages.  Given how easy it is to build for i686 on x86_64 I'd suggest
also going through some of the i686 packages which don't build and
checking if there's a build failure that can be patched or if it's a
case of the software not supporting 32-bit architectures.

¹ https://qa.guix.gnu.org/branch/master

-- 
Efraim Flashner   <efraim@flashner.co.il>   רנשלפ םירפא
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: Hackathon: fix build errors on the "master" branch.

[Christopher Baines]

[-- Attachment #1: Type: text/plain, Size: 2189 bytes --]

Efraim Flashner <efraim@flashner.co.il> writes:

> On Mon, Apr 29, 2024 at 04:52:29PM +0200, Ricardo Wurmus wrote:
>> Hi Guix,
>> 
>> for the past weeks the "master" branch has been in pretty poor state
>> according to ci.guix.gnu.org.  It keeps hovering at around 56% progress,
>> which is a far cry from the 80+% we used to have in the old days.
>> 
>> A lot of packages are in fact broken.  Many more are marked as broken
>> due to a bug affecting ci.guix.gnu.org.
>> 
>> It would be a good idea to do a little spring cleaning and go through
>> the dashboard for the master branch and work on fixing builds that are
>> currently broken on x86_64-linux.  (Other architectures are in worse
>> state, but fixing x86_64-linux may be something we can actually
>> accomplish.)
>> 
>> What do you think?
>
> Using the QA¹ page we can see another data point of how the different
> architectures are doing.  Also, assuming there are 30,000 packages, the
> 4.3% difference on bayfront between x86_64 and i686 would be 129
> packages.  Given how easy it is to build for i686 on x86_64 I'd suggest
> also going through some of the i686 packages which don't build and
> checking if there's a build failure that can be patched or if it's a
> case of the software not supporting 32-bit architectures.
>
> ¹ https://qa.guix.gnu.org/branch/master

I get 1290 as 4.3% of 30000, and that matches up better with the 1551
results I see here for things that are available from bordeaux, but not
from ci:

  https://data.guix.gnu.org/revision/ddf3759645ba76ef532658a257ae74fe0b1788b5/package-derivation-outputs?search_query=&substitutes_available_from=2&substitutes_not_available_from=1&output_consistency=any&system=x86_64-linux&target=none&field=no-additional-fields&after_path=&limit_results=&all_results=on

You can also ask the data service about failing builds, although it only
knows about bordeaux builds:

  https://data.guix.gnu.org/revision/ddf3759645ba76ef532658a257ae74fe0b1788b5/package-derivations?search_query=&system=x86_64-linux&target=none&minimum_builds=&maximum_builds=&build_status=failing&field=builds&after_name=&limit_results=&all_results=on

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 987 bytes --]

Re: Hackathon: fix build errors on the "master" branch.

[Felix Lechner via Development of GNU Guix and the GNU System distribution.]

Hi Chris,

On Tue, Apr 30 2024, Christopher Baines wrote:

> I get 1290 as 4.3% of 30000

Thanks for that!  I do not understand the underlying math but read the
other number with some skepticism.

Kind regards
Felix