bug#40006: [core-updates] Merge wip-hurd

bug#40006: [core-updates] Merge wip-hurd

[Jan Nieuwenhuizen]

Hello Guix'y supporters of the Hurd,

This is a follow-up to a discussion on guix-devel
https://lists.gnu.org/archive/html/guix-devel/2020-03/msg00081.html to
keep track of merging recent work for the Hurd by Efraim and others
that started on the Guix Days @fosdem20

    https://gitlab.com/Efraim/guix  @wip-hurd-bootstrap

The `wip-hurd' branch @savannah has now been reset* and now contains a
set of patches that allow for Hurd development, either by building
packages natively or by cross-building.

Most interesting here are the hurd bootstrap binaries and patches to get
through "commencement".  Efraim built a first set of bootstrap binaries
but I found we at least need some minimal set of glibc patches

    glibc-hurd-clock_t_centiseconds.patch
    glibc-hurd-clock_gettime_monotonic.patch
    glibc-hurd-signal-sa-siginfo.patch

which I added and so I built a new set.  While this mostly works, we may
want to look at/include more patches from the Debian glibc delta

    https://salsa.debian.org/glibc-team/glibc/-/tree/glibc-2.31/debian/patches/hurd-i386

...but how to choose?

This patch series may be especially interesting as I have been also
using a number of hacks to get to build `hello' natively, avoiding
failing/hanging tests here and there

    HACK gnu: acl: Add patches for the Hurd, disable tests.
    HACK gnu: gettext-minimal: Expect test-raise test to fail on the Hurd.
    HACK gnu: check: Skip hanging tests on the Hurd.
    HACK gnu: guile: Disable tests on Hurd.
    HACK gnu: coreutils: Disable tests on Hurd.

I haven't re-evaluated the need for these yet after my last glibc
patches and I haven't included these on the wip-hurd branch; these live
on wip-hurd-system on my gitlab

    https://gitlab.com/janneke/guix  @wip-hurd-system

That's also where my scratchbook on the Hurd lives

   https://gitlab.com/janneke/guix/-/blob/wip-hurd-system/THE-HURD

As discussed on guix-devel, we need to build gnumach-headers and
hurd-headers from a tarball release.  Initially I tried the latest
official releases but they are too old.  So I have created unofficial
source tarballs for Gnumach and Hurd tarball by running `make dist' on
the a git checkout on the Debian Hurd and uploaded them here

    http://lilypond.org/janneke/hurd

wip-hurd is branched off core-updates about two weeks ago; I will now
start with rebasing and re-evaluating on latest core-updates.

Greetings,
janneke

*) Everything from Manolis' old wip-hurd since been merged or ported to
   the new wip-hurd.

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com

bug#40006: [core-updates] Merge wip-hurd

[Manolis Ragkousis]

Awesome work Jan!

bug#40006: [core-updates] Merge wip-hurd

[Jan Nieuwenhuizen]

Jan Nieuwenhuizen writes:

> Hello Guix'y supporters of the Hurd,

`wip-hurd' is now pushed to core-updates as

    3a1c3642d4d611c5516a8ba5b6bc7e39bdc1c9ae

As discussed on IRC yesterday, we did it in two stages: we first
merged all work necessary to build sensible bootstrap binaries
to core-updates:

    3342a1182b15ec031f0ec6f602fd96c1dca3d4b0

then we build bootstrap binaries and verified the result, producing
this commit message

--8<---------------cut here---------------start------------->8---
On 3342a1182b15ec031f0ec6f602fd96c1dca3d4b0
   gnu: make-bootstrap: Use _IOLBF on Guile 2.0 only.

Run
    ./pre-inst-env guix build --target=i586-pc-gnu bootstrap-tarballs --verbosity=1

Producing

    /gnu/store/lhca65c997pvic5cfrpm0dasniwqlg2a-bootstrap-tarballs-0

With guix hash -rx /gnu/store/lhca65c997pvic5cfrpm0dasniwqlg2a-bootstrap-tarballs-0

    07jnq2by98f2a45k8wd2gj62iazvwfa4z7p3w3id4m1g0fdsvc3b
--8<---------------cut here---------------end--------------->8---

Next up is to work towards getting a bootable guix image for the Hurd
and other porting work in the Debian VM.

See:

    https://gitlab.com/janneke/guix/-/blob/wip-hurd-system/THE-HURD

but that's a whole other stage of effort, I'm tempted to close this bug
and open a new one (or two?) for that.  WYDT?

Greetings,
janneke

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com

bug#40006: [core-updates] Merge wip-hurd

[Efraim Flashner]

[-- Attachment #1: Type: text/plain, Size: 1834 bytes --]

On Thu, Mar 26, 2020 at 01:25:32PM +0100, Jan Nieuwenhuizen wrote:
> Jan Nieuwenhuizen writes:
> 
> > Hello Guix'y supporters of the Hurd,
> 
> `wip-hurd' is now pushed to core-updates as
> 
>     3a1c3642d4d611c5516a8ba5b6bc7e39bdc1c9ae
> 
> As discussed on IRC yesterday, we did it in two stages: we first
> merged all work necessary to build sensible bootstrap binaries
> to core-updates:
> 
>     3342a1182b15ec031f0ec6f602fd96c1dca3d4b0
> 
> then we build bootstrap binaries and verified the result, producing
> this commit message
> 
> --8<---------------cut here---------------start------------->8---
> On 3342a1182b15ec031f0ec6f602fd96c1dca3d4b0
>    gnu: make-bootstrap: Use _IOLBF on Guile 2.0 only.
> 
> Run
>     ./pre-inst-env guix build --target=i586-pc-gnu bootstrap-tarballs --verbosity=1
> 
> Producing
> 
>     /gnu/store/lhca65c997pvic5cfrpm0dasniwqlg2a-bootstrap-tarballs-0
> 
> With guix hash -rx /gnu/store/lhca65c997pvic5cfrpm0dasniwqlg2a-bootstrap-tarballs-0
> 
>     07jnq2by98f2a45k8wd2gj62iazvwfa4z7p3w3id4m1g0fdsvc3b
> --8<---------------cut here---------------end--------------->8---
> 
> Next up is to work towards getting a bootable guix image for the Hurd
> and other porting work in the Debian VM.
> 
> See:
> 
>     https://gitlab.com/janneke/guix/-/blob/wip-hurd-system/THE-HURD
> 
> but that's a whole other stage of effort, I'm tempted to close this bug
> and open a new one (or two?) for that.  WYDT?
> 
> Greetings,
> janneke
> 

There's bootstrap binaries now existing, I'd say that counts as merging
in Hurd support.

Congrats!

-- 
Efraim Flashner   <efraim@flashner.co.il>   אפרים פלשנר
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 --]

bug#40006: [core-updates] Merge wip-hurd

[Jan Nieuwenhuizen]

Efraim Flashner writes:

>> Next up is to work towards getting a bootable guix image for the Hurd
>> and other porting work in the Debian VM.
>> 
>> See:
>> 
>>     https://gitlab.com/janneke/guix/-/blob/wip-hurd-system/THE-HURD
>> 
>> but that's a whole other stage of effort, I'm tempted to close this bug
>> and open a new one (or two?) for that.  WYDT?
>
> There's bootstrap binaries now existing, I'd say that counts as merging
> in Hurd support.
>
> Congrats!

Thank you => closing.
janneke

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com

bug#40006: [core-updates] Merge wip-hurd

[Ludovic Courtès]

Hi!

Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> Jan Nieuwenhuizen writes:
>
>> Hello Guix'y supporters of the Hurd,
>
> `wip-hurd' is now pushed to core-updates as
>
>     3a1c3642d4d611c5516a8ba5b6bc7e39bdc1c9ae

Congrats!  \o/

> Next up is to work towards getting a bootable guix image for the Hurd
> and other porting work in the Debian VM.
>
> See:
>
>     https://gitlab.com/janneke/guix/-/blob/wip-hurd-system/THE-HURD
>
> but that's a whole other stage of effort, I'm tempted to close this bug
> and open a new one (or two?) for that.  WYDT?

Yeah, probably use one bug for each focused issue.  It’s easier to
follow & review, and also more rewarding because you get to close issues
more often.  :-)

Ludo’.