Request for merging "world-rebuild" branch

  • Done
  • quality assurance status badge
Details
3 participants
  • Andreas Enge
  • Andreas Enge
  • Christopher Baines
Owner
unassigned
Submitted by
Andreas Enge
Severity
normal

Debbugs page

A
A
Andreas Enge wrote on 21 Jul 02:55 -0700
(address . guix-patches@gnu.org)
aH4OoKAQDWqcbhSr@jurong
This is for the time being an empty bucket. The goal is to reserve a
space in the queue, and then to try out the suggestion developed in
my message to guix-devel: "Branches and other updates"
The idea is to collect a few updates that cause many rebuilds, but are
not covered by a team. The corresponding commits will be collected
in a milestone on codeberg, and decided upon once this branch moves
closer to the front of the queue.

Andreas
A
A
Andreas Enge wrote on 22 Aug 01:11 -0700
Block
(address . control@debbugs.gnu.org)
aKgmOqunaxtB71Ua@jurong
block 79051 by 79071
thanks

# Since the emacs-team branch currently cannot be evaluated by the data
# service, I let world-rebuild go to the front.
# Andreas
A
A
Andreas Enge wrote on 25 Aug 02:50 -0700
Re: Request for merging "world-rebuild" branch
(address . 79071@debbugs.gnu.org)(address . mail@cbaines.net)
aKwx_M_hGrOfz8ZG@jurong
The branch has been started on August 23.

Over the last day it has ground to a halt; I see error message like this
in the build coordinator:
2025-08-25 11:43:01 @ download-progress /gnu/store/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv https://data.guix.gnu.org/nar/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv776 776
2025-08-25 11:43:01 message: "derivation `/gnu/store/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv' has incorrect output `/gnu/store/x8mbs7k5ynwjd16aly14h49bphgv43qf-tvm-0.20.dev0-1.d3a2ed6', should be `/gnu/store/2z33z4qzxxyllm5a9i7acsha53a3p9n0-tvm-0.20.dev0-1.d3a2ed6'"

for a number of packages.

And another type of error:
2025-08-25 11:49:48 message: "path `/gnu/store/x5p6qihyhiiggp9vcy0jbklq7cml0wva-cl-spatial-trees-0-1.81fdad0.drv' does not exist and cannot be created"

I do not see how the first situation can occur.

Andreas
A
A
Andreas Enge wrote on 25 Aug 06:25 -0700
(address . 79071@debbugs.gnu.org)
aKxkVke2ZM3dZVna@jurong
Am Mon, Aug 25, 2025 at 11:50:52AM +0200 schrieb Andreas Enge:
Toggle quote (7 lines)
> Over the last day it has ground to a halt; I see error message like this
> in the build coordinator:
> 2025-08-25 11:43:01 @ download-progress /gnu/store/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv https://data.guix.gnu.org/nar/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv 776 776
> …
> 2025-08-25 11:43:01 message: "derivation `/gnu/store/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv' has incorrect output `/gnu/store/x8mbs7k5ynwjd16aly14h49bphgv43qf-tvm-0.20.dev0-1.d3a2ed6', should be `/gnu/store/2z33z4qzxxyllm5a9i7acsha53a3p9n0-tvm-0.20.dev0-1.d3a2ed6'"
> for a number of packages.

These are frightening, but I realise that they cannot be related to the
branch: This comes from data and not data.qa. So it must be related to
building the master branch, which seems to work: I see recently being
updated packages that are built.

So the problem must be somewhere else...

Andreas
A
A
Andreas Enge wrote on 29 Aug 02:09 -0700
(address . 79071@debbugs.gnu.org)
aLFuUSvQvHfcKGKR@jurong
I have rebased on 06f79dc2bd44ca292dade633d344c3f70749b9e7 to capture
the latest update to qtbase on master.

Andreas
C
C
Christopher Baines wrote on 29 Aug 03:26 -0700
(name . Andreas Enge)(address . andreas@enge.fr)
877bymbf02.fsf@cbaines.net
Andreas Enge <andreas@enge.fr> writes:

Toggle quote (23 lines)
> Am Mon, Aug 25, 2025 at 11:50:52AM +0200 schrieb Andreas Enge:
>> Over the last day it has ground to a halt; I see error message like this
>> in the build coordinator:
>> 2025-08-25 11:43:01 @ download-progress
>> /gnu/store/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv
>> https://data.guix.gnu.org/nar/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv
>> 776 776
>> …
>> 2025-08-25 11:43:01 message: "derivation
>> `/gnu/store/55ay81kd2gdh5xjcwlzl9g1nmwih91al-tvm-0.20.dev0-1.d3a2ed6.drv'
>> has incorrect output
>> `/gnu/store/x8mbs7k5ynwjd16aly14h49bphgv43qf-tvm-0.20.dev0-1.d3a2ed6',
>> should be
>> `/gnu/store/2z33z4qzxxyllm5a9i7acsha53a3p9n0-tvm-0.20.dev0-1.d3a2ed6'"
>> for a number of packages.
>
> These are frightening, but I realise that they cannot be related to the
> branch: This comes from data and not data.qa. So it must be related to
> building the master branch, which seems to work: I see recently being
> updated packages that are built.
>
> So the problem must be somewhere else...

Yeah, there's something wrong with the data service processing revisions
in that sometimes some of the derivation information (inputs/sources)
can be missing from the database, which manifests as failures to
substitute the derivations.

This shouldn't have been possible in the first place, so it's really
hard to tell if I've fixed whatever caused these data issues. I've been
looking in to trying to find and fix the broken derivations on
data.guix.gnu.org today but I've got stuck as Guix is refusing to
substitute a store item, and I can't tell why, which hints at some
potential new problem with the data:


Unfortunately as I say, there's no information as to what is going
wrong, so I've given up for now.
-----BEGIN PGP SIGNATURE-----
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=CqaM
-----END PGP SIGNATURE-----

A
A
Andreas Enge wrote 5 days ago
aLYAPHJZLiTtMNU6@jurong
For the last two days, QA has stopped building packages.
As aarch64 has been successfully finished, I am confident that the
patches on the world-rebuild branch are good to go. On x86_64, there
are still 3500 packages to be built on QA, but they look good on the
CI dashboard. i686 has noticeably more packages being blocked by build
failures on QA, but the biggest blocker there, ghc, appears to be
available on CI.

So I have rebased the branch and pushed it to master.

Over the next two weeks, I will mostly be away from keyboard, and after
that probably spend less time than over the summer to work on branches.
So please organise yourselves to follow the branches on
and merge one after the other when they are ready; depending on the
state of QA, this may require to also have a look at CI.

Andreas
Closed
?
Your comment

Commenting via the web interface is currently disabled.

To comment on this conversation send an email to 79071@patchwise.org

To respond to this issue using the mumi CLI, first switch to it
mumi current 79071
Then, you may apply the latest patchset in this issue (with sign off)
mumi am -- -s
Or, compose a reply to this issue
mumi compose
Or, send patches to this issue
mumi send-email *.patch