Request for merging "go-team" branch

Hi,

Placing to the merge queue go-team branch after shifting default Go to 1.24

which requires rebuilt all the packages with go-build-system.

As the bonus - kubo is updated to the latest version with 100%

unvednored inputs! Time for restic now.

Merge requests included:

- Refresh all packages from (gnu packages golang-build) https://codeberg.org/guix/guix/pulls/381

--

Oleg

Hello the go team,

our queue of merge requests has gone down quite a bit, so this might

not be as far away as one could fear. Please communicate on the state of

the branch, in particular if it needs postponing.

After the emacs-team branch, I have rebased on commit

ced31f8dd156e4202a2c7115fc003608a541388c

Andreas

Hi Andreas,

Thanks for rebase.

I would love to have this brunch merged.

The main interest is to bring go-1.24 as default compiler.

I've tested locally with:

guix build -P1 go@1.24 -P1 go@1.23

The only package which failed is lxd, see open issue

Brunch is ready, and I'm not intend to make any other

large maneuvers.

Thanks,

Oleg

VCS: https://github.incerto.xyz/;https://git.sr.ht/~hellseher/

GPG: 9847 81DE 689C 21C2 6418 0867 76D7 27BF F62C D2B5

… наш разум - превосходная объяснительная машина которая способна найти

смысл почти в чем угодно, истолковать любой феномен, но совершенно не в

состоянии принять мысль о непредсказуемости.

Rebased on 5d6604a88e1dae62f56a196d5a8140d272f8e78c to digest the

ruby-team merge.

Andreas

Rebased on a88d6a45e422cede96d57d7a953439dc27c6a50c.

Included PRs:

- [go-team] build-system/go: Fix substitutable? inconsistency.

https://codeberg.org/guix/guix/pulls/857

--

Thanks,

Oleg

Hi,

This branch is really and doesn't include any regression but full

Golang stack rebuild due to change to default compiler.

Should we wait core-packages-team, nss-team and c++- team?

Thanks,

Oleg

VCS: https://github.incerto.xyz/;https://git.sr.ht/~hellseher/

GPG: 9847 81DE 689C 21C2 6418 0867 76D7 27BF F62C D2B5

… наш разум - превосходная объяснительная машина которая способна найти

смысл почти в чем угодно, истолковать любой феномен, но совершенно не в

состоянии принять мысль о непредсказуемости.

Hello,

Am Sat, Jul 05, 2025 at 09:34:02PM +0100 schrieb Sharlatan Hellseher:

which seems to be quite a lot of packages - when doing a

"guix refresh -l go", I do not even see how many since the list of

packages rolls off my screen.

I think the core-packages-team is almost ready to go, there is a lot of

work in a milestone

https://codeberg.org/guix/guix/milestone/21092

to iron out the last problems.

Then I do not see a particular reason to prefer one branch over the

other; all of them will have to be rebased. There is also the qt-team

branch, part of which has been moved to core-packages-team.

The QA data service has been broken for a week now with no branches

being built; CI usually errors out with 504 when clicking on a revision,

but the jobset page for a given branch and the dashboards seem to work.

So to be honest, we currently do not have the tooling to comfortably

decide which branches to merge...

Andreas

Andreas Enge <andreas@enge.fr> writes:

If possible, I would prefer rust-team to prioritize merging, so that the

subsequent branches can reduce the compilation of many unnecessary rust

packages

Hi,

I've rebased on 2e628ad02c5b3efa4e6e4359cd866c0d57c18bff, resolved

conflicts and managed to build the whole chain of the first level

dependent packages:

There are only 2 packages which faild to build.

- lxd : old package, there is a new version which requirs 60+ new

packages, tracking refresh progress here

https://codeberg.org/guix/guix/issues/552

- demlo : old package, depends on go-1.17 (!) and unmaintained

dependencies: go-github-com-wtolson-go-taglib,

go-github-com-aarzilli-golua, go-github-com-stevedonovan-luar,

go-gitlab-com-ambrevar-golua-unicode

--

Oleg

Rebased on 6cf00e6cdec63d141cfc716446b8c0eac464ab36.

Rebased on b22edc407e34848745106ce29040bbfa29aeeec3

--

Thanks,

Oleg

Rebased on 4e9d7a590f364782a170d0f419ebf0a339b02469

Hello,

after the qt-team merge, it is time to rebase again; normally you are

third in line after c++-team and nss-updates, but I have not heard from

nss-updates in a while, so this order could be swapped.

Andreas

Hi Andreas,

Thanks for the ping.

I tried to rebase on curent master HEAD

b740fe801e072094c6262ed71a30a0d953b0886b, but I need to resole some

conflicts first.

Will push later tonight.

---

Oleg

block 78689 by 78781

thanks

Am Fri, Jul 25, 2025 at 04:33:02PM +0100 schrieb Sharlatan Hellseher:

I have rebased again on commit bc2625db7dcc9f988426de567161615a6d255f81,

which went through without any conflict; this is a commit known to data.qa.

Andreas

Hello,

after c++-team was merged, let's go for go-team!

I have rebased on

https://data.qa.guix.gnu.org/revision/6cf6feee30db2392de3e1abbef1b92036e641d9e

which is a master branch commit being treated by the data service.

There was a merge conflict in go-github-com-docker-docker, which

hopefully I have resolved correctly.

Andreas

Hi Andreas,

Thanks for the rebase, I've been away for a while. Last time I've

checked the QA status it was on 69/75%, hoping it will be picked quickly

this time.

Thanks,

Oleg

Well, the commit on which I had rebased never finished its evaluation by

the data service; probably my restarting the service in the middle of it

did not help... I have rebased again on a green master commit so that we

get numbers. However, last time the go-team branch took almost 11 hours

for its evaluation, so we will need a little patience.

Andreas

Hi Andreas,

Which coverage are we expecting to pass QA?

It's 83% on Bordeaux CI for i686 architecture.

Thanks,

Oleg

VCS: https://github.incerto.xyz/;https://git.sr.ht/~hellseher/

GPG: 9847 81DE 689C 21C2 6418 0867 76D7 27BF F62C D2B5

… наш разум - превосходная объяснительная машина которая способна найти

смысл почти в чем угодно, истолковать любой феномен, но совершенно не в

состоянии принять мысль о непредсказуемости.

Am Wed, Aug 06, 2025 at 03:38:07PM +0100 schrieb Sharlatan Hellseher:

This is probably cheating...

I think the percentage is rather high because not that many packages are

concerned by changes. With the core-packages merge, the percentage was

essentially the same as before, but then people experienced a number of

packages close to the leaves they were interested in, and which did not

build anymore.

Maybe Go is and was not much available on i686?

https://qa.guix.gnu.org/branch/go-team

shows only about 3000 packages altogether concerned by the change on i686,

but about 13000 on x86_64 and still over 10000 on aarch64.

The idea would be to arrive at about the same number of "succeeding"

packages as before; right now most of them are unknwon. We want to not

only be assured that not much breaks, but also that substitutes will be

available immediately after the merge.

I think not a single package has been built since this morning.

There has been a "guix gc" running since 2am (!), and I dare not

interrupt it. Last time I did it I lost access to the machine, even

though the causal connection is unclear. But before this is finished,

I think nothing will happen.

And on CI I get 504 errors most of the time, so it is of not much help

either.

Andreas

Am Wed, Aug 06, 2025 at 04:59:32PM +0200 schrieb Andreas Enge:

So if I understand the output of "ps" correctly, 30 hours of "guix gc",

and not a single go-team package built. I have killed the gc process,

for good measure restarted the daemon, the build coordinator and

qa-frontpage, and hope things will pick up.

I see errors in the log file of qa-frontpage, trying to talk to the

build coordinator:

2025-08-07 09:49:40 In knots/timeout.scm:

2025-08-07 09:49:40 203:4 11 (with-port-timeouts _ #:timeout _ #:read-timeout _ #:write-timeout _)

2025-08-07 09:49:40 In guix-qa-frontpage/manage-builds.scm:

2025-08-07 09:49:40 600:29 10 (_)

2025-08-07 09:49:40 In guix-build-coordinator/client-communication.scm:

2025-08-07 09:49:40 1028:15 9 (fold-builds "http://127.0.0.1:8746" #<procedure 7f5b0ba074b8 at guix-qa-frontpage/manage-builds.scm:552:5 (build-details result)> () #:tags _ #:not-tags _ #:systems _ #:not-systems _ #:processed _ #:canceled _ #:priority-> _ #:priority-< _ #:created-at-> _ #:created-at-< _ #:relationship _ #:after-id _ #:limit _ #:port _)

2025-08-07 09:49:40 813:11 8 (send-request "http://127.0.0.1:8746" GET "/builds?tag=branch:go-team&not_tag=revision:7ca1cb9db6c3e6c6ca8158a02275ef204957180d&processed=false&canceled=false&relationship=no-dependent-builds" #:request-body _ #:port _)

2025-08-07 09:49:40 In web/client.scm:

2025-08-07 09:49:40 533:24 7 (http-request _ #:body _ #:verify-certificate? _ #:port _ #:method _ #:version _ #:keep-alive? _ #:headers _ #:decode-body? _ #:streaming? _ #:request _)

2025-08-07 09:49:40 In web/response.scm:

2025-08-07 09:49:40 196:32 6 (read-response #<input-output: socket 122>)

2025-08-07 09:49:40 In web/http.scm:

2025-08-07 09:49:40 1198:15 5 (read-response-line _)

2025-08-07 09:49:40 159:10 4 (read-header-line _)

2025-08-07 09:49:40 In ice-9/suspendable-ports.scm:

2025-08-07 09:49:40 681:11 3 (read-delimited _ _ _)

2025-08-07 09:49:40 184:27 2 (fill-input #<input-output: socket 122> _ _)

2025-08-07 09:49:40 67:33 1 (read-bytes #<input-output: socket 122> #vu8(110 115 115 45 117 112 100 97 116 101 115 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 …) …)

But debugging this is beyond my knowledge.

Andreas

Hello,

nice progress in the last 24 hours: x86_64 has gone from 1345 built

packages to 5235; at the same time, part of the build farm is still

turned off because already the running machines are not running at full

capacity. At this speed, we should be ready to merge in two days.

Spurious build failures on aarch64:

https://bordeaux.guix.gnu.org/build/1189d555-b8f4-4335-8af6-569aaa8fc8e9

due to lack of free disk space on mayflower.

Chris, maybe more aggressive "guix gc" would be in order?

Andreas

Am Fri, Aug 08, 2025 at 09:42:50AM +0200 schrieb Andreas Enge:

Using the bug as a notebook: a bit of slow-down over the last 27 hours,

7842 succeeding on x86_64, 5071 to go. But we are progressing.

Andreas

Midnight check

- https://ci.guix.gnu.org/

- i686-linux 71.7%

- x86_64-linux 55.8%

- https://bordeaux.guix.gnu.org/

- aarch64-linux 84.5%

- i686-linux 84.8%

- x86_64-linux 88.4%

- With branch changes

- x86_64-linux

- Succeeding 9174

- Failing 237 (less than on base)

- Blocked 431 (less than on base)

- Unknown 3603

--

Thanks

Oleg

Am Sat, Aug 09, 2025 at 12:29:33PM +0200 schrieb Andreas Enge:

And since the last 23 hours: 10580 succeeding, 2068 to go.

I have rebased on 82e0408b0e79a0beebe06b27809d643453ef29b1 and pushed.

Concerning aarch64, I will resubmit the blocking builds, which had

failed due to out of memory errors; maybe this will restart the

depending packages.

I suppose we can push to master tomorrow.

Andreas

Build failure of go@1.24 on armhf:

https://data.qa.guix.gnu.org/gnu/store/gd07vhc7cg63s1d66b1zm91xchi0sp8b-go-1.24.3.drv

While this is not blocking, maybe you have an idea on how to repair it.

Andreas

Hi Andreas,

Thanks for letting me know.

I'm finalising python-team after migrating to pyproject-build-system, high

impact packages.

Will take a look tomorrow

VCS: https://github.incerto.xyz/;https://git.sr.ht/~hellseher/

GPG: 9847 81DE 689C 21C2 6418 0867 76D7 27BF F62C D2B5

… наш разум - превосходная объяснительная машина которая способна найти

смысл почти в чем угодно, истолковать любой феномен, но совершенно не в

состоянии принять мысль о непредсказуемости.

On Sun, 10 Aug 2025, 20:30 Andreas Enge, <andreas@enge.fr> wrote:

One test failed

--- FAIL: TestElfBindNow (4.83s)

--- FAIL: TestElfBindNow/bindnow-pie-linkmode-external (4.37s)

elf_test.go:388: test .dynamic: "bindnow-pie-linkmode-external"

section not in readonly segment

elf_test.go:390: section .dynamic location: st=0xf7e10 en=0xf7f28

elf_test.go:392: sec .dynamic found in these segments:

elf_test.go:394: "PT_LOAD"

elf_test.go:394: "PT_DYNAMIC"

elf_test.go:394: "PT_GNU_RELRO"

elf_test.go:396:

all segments:

elf_test.go:398: 0 t=PT_PHDR fl=PF_R st=0x34 en=0x194

elf_test.go:398: 1 t=PT_INTERP fl=PF_R st=0x194 en=0x1e3

elf_test.go:398: 2 t=PT_LOAD fl=PF_X+PF_R st=0x0 en=0x9be94

elf_test.go:398: 3 t=PT_LOAD fl=PF_W+PF_R st=0x9dc28 en=0xf9f46

elf_test.go:398: 4 t=PT_DYNAMIC fl=PF_W+PF_R st=0xf7e10 en=0xf7f28

elf_test.go:398: 5 t=PT_NOTE fl=PF_R st=0x1e4 en=0x28c

elf_test.go:398: 6 t=PT_GNU_EH_FRAME fl=PF_R st=0x9be8c en=0x9be94

elf_test.go:398: 7 t=PT_GNU_STACK fl=PF_W+PF_R st=0x0 en=0x0

elf_test.go:398: 8 t=PT_LOPROC+1 fl=PF_R st=0x9be80 en=0x9be88

elf_test.go:398: 9 t=PT_TLS fl=PF_R st=0x9dc28 en=0x9dc28

elf_test.go:398: 10 t=PT_GNU_RELRO fl=PF_W+PF_R st=0x9dc28

en=0xf8000

elf_test.go:401: test bindnow-pie-linkmode-external failed

FAIL

FAIL cmd/link/internal/ld 7.272s

VCS: https://github.incerto.xyz/;https://git.sr.ht/~hellseher/

GPG: 9847 81DE 689C 21C2 6418 0867 76D7 27BF F62C D2B5

… наш разум - превосходная объяснительная машина которая способна найти

смысл почти в чем угодно, истолковать любой феномен, но совершенно не в

состоянии принять мысль о непредсказуемости.

Am Sun, Aug 10, 2025 at 11:26:10AM +0200 schrieb Andreas Enge:

It did!

However, over night a "guix gc" job has started, which I have just killed.

Hopefully we can continue like this until the hard disk will have been

upgraded...

I have also restarted guix-build-coordinator and qa-frontpage just in case.

Job submission has picked up again, visible for x86_64. On the positive

side, it looks as if the aarch64 machines have continued to build in the

absence of new submissions and have almost emptied theire queue, so the

architecture has caught up over night.

This looks still feasible, but maybe rather in the evening than in the

morning. I have rebases nss-updates on top of go-team, so the data

service has evaluated the branch already, and it should start right away.

Andreas

Am Mon, Aug 11, 2025 at 10:00:39AM +0200 schrieb Andreas Enge:

I think we are ready to go! If you are around for pushing to master

tonight, Oleg, I will let you enjoy the pleasure, if not I will go ahead

and push myself tomorrow morning.

Andreas

Hi,

Great news!

Sorting out some family business and would enjoy to merge ^.^

Thanks,

Oleg

VCS: https://github.incerto.xyz/;https://git.sr.ht/~hellseher/

GPG: 9847 81DE 689C 21C2 6418 0867 76D7 27BF F62C D2B5

… наш разум - превосходная объяснительная машина которая способна найти

смысл почти в чем угодно, истолковать любой феномен, но совершенно не в

состоянии принять мысль о непредсказуемости.

On Mon, 11 Aug 2025, 19:16 Andreas Enge, <andreas@enge.fr> wrote:

Hi,

go-team branch has been merged to master!

--

Oleg

Am Mon, Aug 11, 2025 at 08:13:32PM +0100 schrieb Sharlatan Hellseher:

Time to celebrate, thanks for seeing it through!

Andreas