unofficial mirror of help-guix@gnu.org 
 help / color / mirror / Atom feed
From: tobtoht <tob@featherwallet.org>
To: "help-guix@gnu.org" <help-guix@gnu.org>
Subject: Guix 1.4 unable to pull with --no-substitutes
Date: Sat, 31 Aug 2024 20:44:25 +0000	[thread overview]
Message-ID: <-F5boi4fqKBzG9H0-EZANSRgFehli6AFc86G88G23S_IkT4OJ17hlexi9d3L7yvQml9MFMU7eIWHVH8_4s5PSryCm3j82TXS83h4Av0rhho=@featherwallet.org> (raw)

Hi,

Using guix 1.4.0 as a starting point on a foreign distro, I'm unable to
complete a pull without substitutes.

Steps to reproduce:

$ install guix using the installer script (or apt install guix)
$ export GUIX_BUILD_OPTIONS="--no-substitutes"
$ guix pull

It then proceeds to build a number of packages, including guile 3.0.7. Tests 
for this package fail as described here: https://issues.guix.gnu.org/48392

I have tried building it at least a dozen times and the tests fail each time.

Context:

A project I am involved with uses guix time-machine + guix shell to set up a
build environment that is used to build portable binaries for a desktop 
application. For security and auditability reasons, the build process is 
bootstrappable and verifiably reproducible.

Multiple builders build the software using guix. The results are signed, 
submitted to a central repository, and compared.

To ensure the build is in fact bootstrappable, and to detect potential 
non-determinism in guix packages, I want to encourage builders to build the 
manifest without substitutes.

I'd like to provide mostly distribution-agnostic documentation. Instructing 
builders to build a more recent version of guix from source would make this
process less accessible.

Questions:

1. Why does it attempt to build guile 3.0.7? The version of guile defined in
   the latest version of guix (or the time-machine pinned commit) is 3.0.9.

2. Is there a way to avoid having to build guile 3.0.7? (While still using
   the guix provided by the installer script as a starting point.)

3. If not, is it possible to skip the test phase for this specific package?

4. If not, is there anything I can do to make the tests less likely to fail?



             reply	other threads:[~2024-09-01 11:48 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-31 20:44 tobtoht [this message]
2024-09-04 16:52 ` Guix 1.4 unable to pull with --no-substitutes Simon Tournier

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://guix.gnu.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='-F5boi4fqKBzG9H0-EZANSRgFehli6AFc86G88G23S_IkT4OJ17hlexi9d3L7yvQml9MFMU7eIWHVH8_4s5PSryCm3j82TXS83h4Av0rhho=@featherwallet.org' \
    --to=tob@featherwallet.org \
    --cc=help-guix@gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).