[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: 01/01: hydra: Work around import problem.
From: |
Mark H Weaver |
Subject: |
Re: 01/01: hydra: Work around import problem. |
Date: |
Tue, 30 Jan 2018 21:12:50 -0500 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/25.3 (gnu/linux) |
Danny Milosavljevic <address@hidden> writes:
> On Mon, 29 Jan 2018 09:47:01 -0500
> Mark H Weaver <address@hidden> wrote:
>
>> I strongly suspect that this evaluation error occurred for a commit
>> _before_ you added (use-modules (gnu bootloader)), unless you have
>> reason to believe otherwise.
>
> Ah, might be. How do I find out which commit has been evaluated?
The commit of _successful_ Hydra evaluations is recorded, but I don't
know of any record of the commits of failed evaluations. However, from
my own memory of the timeline, I feel reasonably confident that the
failed Hydra evaluation was started in the ~10 hour period between
commit de2bc8d6d (hydra: add arm image builder) and commit 1fe805b29
(hydra: Add (gnu bootloader) import.)
I think it would be okay to simply revert commit ebd6a6445 (hydra: Work
around import problem.) and see what happens.
However, it also seems that the current 'flash-image' job isn't working.
Its first build on Hydra got stuck at a guile prompt:
https://hydra.gnu.org/build/2481717/nixlog/6/tail-reload
https://hydra.gnu.org/build/2481717
The error message was "ERROR: In procedure copy-file: Success". I
believe this was caused by the fact that Guile's 'copy-file' procedure
incorrectly assumes that an error occurred if 'write' writes fewer bytes
than we asked it to. This can happen if interrupted by a signal, or if
the file system was full. I'll look into fixing this in Guile, although
in this case it seems likely that the file system was full.
Mark