s #t,
like in https://issues.guix.gnu.org/70341#2 ?
Best Regards,
Nigko Yerden
Hello Guix!
Recently the patch https://issues.guix.gnu.org/70341 has completely
disappeared from https://qa.guix.gnu.org/patches? list of patches
under consideration. I would like to know what does it mean? The patch
was considered inappropriate and rejected? Or some kind of error
in patch proces
Hello Guix!,
Consider a minimal test git repository [1] created in line with Cookbook
recommendations [2]. It has the following file structure:
.
├── content
├── .guix-channel
├── guix.scm → .guix/modules/test-repo-package.scm
└── .guix
└── modules
└─
Hello Florian,
There is no grievance for anything. Sorry, if my message has such
a tinge, it does not correspond to what was meant. Also I do not
think that 'local-file' is doing something wrong. I suspect that
introducing changes to such an ubiquitous procedure on such a minor
subject is not nec
pelzflorian (Florian Pelz) wrote:
Your intention was just that others with the
same use-case find your workaround in the mail archives. Then you did
right.
Not just this, but also that somebody may consider changing the cookbook.
The guix-cookbook describes a (local-file "."), not "../..",
Hello Florian,
pelzflorian (Florian Pelz) wrote:
Nonsense; it must have worked; 7.7 Wrapping Up lists
https://git.savannah.gnu.org/cgit/guile.git/tree/.guix/modules/guile-package.scm?id=cd57379b3df636198d8cd8e76c1bfbc523762e79
as proof.
I clearly did something wrong. Sorry.
What’s the diffe
Hello Attila Lendvai,
Thank you very much for your hints and references! Indeed I was puzzled by
weird behavior of 'current-filename'. I wrote another alternative
using '(module-filename (current-module))' based on
https://issues.guix.gnu.org/55464
See alt2 branch of [1].
This variant does not
Hello Florian!
Here is my version of your diff in attachment. With this patch
'guix build test-repo' works for the main branch of test-repo-channel,
i.e. it seems to solve one problem (I'm sure that 'guix build guile@3.0.99-git'
works as well for the guile channel, but didn't check this explicitl
On the second problem, I have found that while it is impossible
to obtain test-repo source code via 'guix build -S test-repo',
it can be retrieved programmatically from 'guix repl' REPL
using this commands:
(use-modules (test-repo-package)
(guix packages)
(guix gexp)
'guix build --debug=5 -S test-repo' build logs shows that the problem
sits in 'guix-build' procedure from 'guix/scripts/build.scm'. 'drv' local
variable is initialized by '(options->derivations store opts)' at line 762
to a list '("/gnu/store/...-test-repo-checkout") containing store
path to the t
pelzflorian (Florian Pelz) wrote:
Could you send a patch about the `guix build' diff you debugged to
guix-patc...@gnu.org? Preferrably you would drop the whole
absolute-dirname’s `if' and canonicalize unconditionally, I guess. Make
explicit in the docstring or in comments that symlinks are the
Hello Florian,
I am sure you have nothing to apologize for. I have been
thinking about your suggestion to submit `absolute-dirname' patch. In the
Guix manual at the end of the section 6.7 Creating a channel [1] there is the
following statement about Guix policy on changing API, not to mention ABI
Nigko Yerden wrote:
(dirname
(canonicalize-path
(string-append (current-source-directory) (current-filename
Sorry, this expression for source directory is wrong. The correct one:
(dirname
(canonicalize-path
(string-append (current-source-directory) "/"
pelzflorian (Florian Pelz) wrote:
But then again, the current current-source-directory
already does follow symlinks in nearly all cases, even in configuration
files
Don't see it neither in the code nor in our examples. 'syntax-source' doesn't
do this. The second branch of 'if' in 'absolute-dirna
Hello Florian,
pelzflorian (Florian Pelz) wrote:
While processing guile-package.scm,
(search-path %load-path "guile-package.scm")
returns an absolute path if and only if guile-package.scm is in the
load-path, like when using it from a channel. Then, your diff makes it
resolve symlinks.
That'
Hello Florian,
pelzflorian (Florian Pelz) wrote:
Still, when would your diff break someone else’s code?
Instead of answering this tough question I decided to make another
patch (see attachment) which is guaranteed not to break another's code:
- The patch adds 'follow-symlinks?' argument to 'c
pelzflorian (Florian Pelz) wrote:
Could you send this diff to guix-patc...@gnu.org?
Here it is: https://issues.guix.gnu.org/72867
Regards,
Nigko
17 matches
Mail list logo