Yes, the issue still need fixing.
Here is my suggestion https://issues.guix.gnu.org/70341
Regards,
Nigko
Vincent Legoll wrote:
Hello,
On Thu, Jun 20, 2024 at 4:49 AM Nigko Yerden wrote:
Indeed, they are related. My message from 70332 was intended
to go to 70302 and didn't get there by my
Hi Vincent,
Indeed, they are related. My message from 70332 was intended
to go to 70302 and didn't get there by my mistake, thus creating
new issue 70332.
Regards,
Nigko
Vincent Legoll wrote:
I think the two issues are related, and the second one may have been
created by typoing the issue
Hello Florian,
From testing old Guix System system generations, I believe September
2023’s new daemon feature “builtin:git-download” has caused the change
in hashes
As far as I remember, approximately at that time, in the late
autumn, I have started to experience very long 'guix pull'.
I’m
Hello Florian,
These are great news! Thanks a lot for investigating this!
Indeed, foreign distros still run guix-daemon from Guix v1.4,
which is very old now. It makes sense that old guix-daemon + recent
remaining parts of Guix might have produced modified derivations.
For foreign distros
I had not understood that you used the --commit option, sorry for the
misunderstanding.
Technically I wasn't because at the time of experiment 08360a2cd
was the latest commit (not only the latest commit with substitutes).
Foreign distros are not treated specially, but configuration for
stitutes.
[1] https://ci.guix.gnu.org/api/latestbuilds?nr=1=guix.x86_64-linux
Regards,
Nigko
pelzflorian (Florian Pelz) wrote:
Hello Nigko Yerden. You could definitely test if substitutes are never
available if you used
https://guix.gnu.org/manual/en/html_node/Channels-with-Substitutes.html
wh
Hello Guix!
I have two systems: first (A) is a full Guix system and second (B) is a
foreign distribution with Guix package manager installed onto. I have
noticed when I'm performing 'guix pull' system B always build 'guix
pull' derivations by itself while system A most often just download