help-guix
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: scm_lreadr error on guix pull


From: Phil Beadling
Subject: Re: scm_lreadr error on guix pull
Date: Mon, 23 Aug 2021 18:52:04 +0100

On Mon, 23 Aug 2021 at 18:18, zimoun <zimon.toutoune@gmail.com> wrote:

>
> Could you try with the broken account this:
>
>   /usr/bin/guix pull -p /tmp/test
>   /usr/bin/guix pull -p ~/.config/guix/current
>   hash guix
>   which guix
>   guix pull
>
>
Thanks for the suggestion - I think I did a similar test which has solved
the initial problem, but there is an underlying problem I'm still perplexed
by.  I will try your suggestion too, but below is my progress so far.

Say we have 2 users:
good-account
bad-account

One bad account I did:
cd
mkdir vault
mv * vault/
mv .* vault

logout/login

Then with a bare account I did a guix pull and it worked!

I thought I'd solved the problem - presumably some stale setting somewhere
in bad-account's config/env/etc - but then I tried a 'guix environment'
command and noticed that the bad-account was still trying to build things
the good-account would take from the local gnu store.

I isolated an example below - the 2 accounts retrieve the same
channels/branches/commit id.  The good-account picks up the version
previously built in the store, but the bad account (presumably) comes up
with a fingerprint that disagrees with the other account and decides to
build from scratch.

___________________________


good-account@app-server:~$ guix build myproject

updating checkout of 'ssh://git@git:81/proj/myproject.git'...

retrieved commit *422427c60ebeb9dc725e13892876bf64476509eb <- commit id is
identical*

*/gnu/store/0z1aax8g3r3bxjfnbf2p4rffkvzy3bz8-myproject-7.7.1  <- Take the
version in the store*


good-account@app-server:~$ guix describe

*Generation 40 **Aug 23 2021 15:09:52*    (current)

  my-packages 99af600

    repository URL:  ssh://git@git:81/proj/myproject.git

    branch: master

    commit: *99af600245d284dcb4b53cc5000b98ec35432c6f <- CHANNEL COMMIT IDs
match below *

  guix 6755408

    repository URL: https://git.savannah.gnu.org/git/guix.git

    branch: master

    commit: *675540892719387e1e4e76f097ff8e4ee4b559f7 <- CHANNEL COMMIT IDs
match below *




____________________________


bad-account@app-server:~$ guix build myproject

updating checkout of 'ssh://git@git:81/proj/myproject.git'...

retrieved commit *422427c60ebeb9dc725e13892876bf64476509eb <- commit id is
identical *

The following derivation will be built:

   /gnu/store/a25k61g9c4p6cvjzfppzg1wn5ppra9qv-myproject-7.7.1.drv

<SNIPPED OUT BUILD LOG FOR BREVITY>

successfully built
/gnu/store/318cq2099mgxjsikrrsxrzn6y5nzp24q-lchsmartpy-7.7.1.drv

*/gnu/store/6cnf47k37ng9zwlg8m63dprziwg2p1gh-myproject-7.7.1   <- Ignores
the version in the store and rebuilds *



bad-account@app-server:~$  guix describe

*Generation 5*    *Aug 23 2021 15:09:52*    (current)

  my-packages 99af600

    repository URL:  ssh://git@git:81/proj/myproject.git

    branch: master

    commit: *99af600245d284dcb4b53cc5000b98ec35432c6f <- CHANNEL COMMIT IDs
match above*

  guix 6755408

    repository URL: https://git.savannah.gnu.org/git/guix.git

    branch: master

    commit: *675540892719387e1e4e76f097ff8e4ee4b559f7 <- CHANNEL COMMIT IDs
match  above *




There is probably is still some rogue config operating on behalf of the
bad-account that produces the different fingerprint - but no idea what this
could be?


reply via email to

[Prev in Thread] Current Thread [Next in Thread]