trilinos-serial-xyce source disappeared

  • Done
  • quality assurance status badge
Details
2 participants
  • Marius Bakke
  • zimoun
Owner
unassigned
Submitted by
Marius Bakke
Severity
normal
M
M
Marius Bakke wrote on 15 Mar 2020 13:48
(address . bug-guix@gnu.org)
87sgi9lr77.fsf@devup.no
The source URI for 'trilios-serial-xyce' serves a 404 page. Following
download links from the home page only leads to the git repository.

This package should probably be changed to download a tagged git
revision instad of attempting to fetch a defunct tarball.

ci.guix.gnu.org has a copy of the previous source tarball so it's not
very urgent (anyone coming here can add "--fallback" to their guix
invokation to get it).
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCgAdFiEEu7At3yzq9qgNHeZDoqBt8qM6VPoFAl5uJAwACgkQoqBt8qM6
VPoP+gf+P9RTEM0FKCxaBCRiMMGHxj7q/pqzhg5R2PP17ZaAgv78mNhYs3p6glor
jm89touRMvulPfH/HL4G6pah9czC2vbfHc9azgGe6jDA35TAMWKtjl2lYh/jJe3a
l27RJKJZAdDGsYRIRFDtPRN6rx9ZtTCvId8Tic+pKJYBfa70sX582X7xlkwZCAPx
4MwD0KUX0ECF4LBU10rKm/r8Nrett4k5rJS1WTWtUFIl9IKdQJEzPd4ab39lCLgj
0McJ/75CgaNJTqCQz3IjPJWaSmNeYDRM+FSb1k5EQudyTj29wuTFu9Eo8S09vibO
mRu8LFFev/bCpE673ExGNLP81P1gSQ==
=RNbf
-----END PGP SIGNATURE-----

Z
Z
zimoun wrote on 15 Oct 2020 21:38
(name . Marius Bakke)(address . mbakke@fastmail.com)(address . 40074@debbugs.gnu.org)
868sc71dr9.fsf@gmail.com
Dear,

On Sun, 15 Mar 2020 at 13:48, Marius Bakke <mbakke@fastmail.com> wrote:
Toggle quote (10 lines)
> The source URI for 'trilios-serial-xyce' serves a 404 page. Following
> download links from the home page only leads to the git repository.
>
> This package should probably be changed to download a tagged git
> revision instad of attempting to fetch a defunct tarball.
>
> ci.guix.gnu.org has a copy of the previous source tarball so it's not
> very urgent (anyone coming here can add "--fallback" to their guix
> invokation to get it).

The commit 74cb33c22818b25d1b9c6d576352fc8249832d34 [1] transitioned
from ’url-fetch’ to ’git-fetch’ for missing upstream (even disappeared
one). I am proposing similar strategy:

First patch, replace ’url-fetch’ by ’git-fetch’ with creating a Git repo
from the build farm tarball and falling back to SWH. This does not
change the integrity checksum. And it avoids the “pressure“ on the
build farm.

Second patch, update the upstream source to the new good ’git-fetch’.
And this change the integrity checksum. Maybe upstream released a new
version.


WDYT?



All the best,
simon
Z
Z
zimoun wrote on 18 Oct 2022 17:37
(name . Marius Bakke)(address . mbakke@fastmail.com)(address . 40074@debbugs.gnu.org)
86ilkhywgx.fsf@gmail.com
Hi,

On Thu, 15 Oct 2020 at 21:38, zimoun <zimon.toutoune@gmail.com> wrote:
Toggle quote (11 lines)
> On Sun, 15 Mar 2020 at 13:48, Marius Bakke <mbakke@fastmail.com> wrote:

> First patch, replace ’url-fetch’ by ’git-fetch’ with creating a Git repo
> from the build farm tarball and falling back to SWH. This does not
> change the integrity checksum. And it avoids the “pressure“ on the
> build farm.
>
> Second patch, update the upstream source to the new good ’git-fetch’.
> And this change the integrity checksum. Maybe upstream released a new
> version.

Commit b729be989f137749430779499b8c0dd1d54f2a18 did the replacement by
’git-fetch’ using the upstream repo.

It is hard to compare the both sources but it does not really matter
since the new one is the official upstream one.

Well, what remains is the guarantee that ’guix time-machine’ will still
work even if the Guix project looses the tarball.

This case should be covered by Disarchive and SWH. Well, is it worth to
keep open this issue when the question about the time-machine coverage
is bigger than just this example.

Closing?

Cheers,
simon
Z
Z
zimoun wrote on 3 Nov 2022 10:55
(name . Marius Bakke)(address . mbakke@fastmail.com)(address . 40074-done@debbugs.gnu.org)
86leosbbvc.fsf@gmail.com
Hi,

On Tue, 18 Oct 2022 at 17:37, zimoun <zimon.toutoune@gmail.com> wrote:

Toggle quote (4 lines)
> This case should be covered by Disarchive and SWH. Well, is it worth to
> keep open this issue when the question about the time-machine coverage
> is bigger than just this example.

After 2 weeks without an answer, I am closing since this specific case
is already covered by the ongoing work around Disarchive and SWH.

Feel free to reopen if I miss something.

Cheers,
simon
Closed
?