Closed dreamer-coding closed 1 month ago
Also I'm interested in why the checksums were originally wrong...
Aparently used sha224sum insted of sha256sum tool as suggested probably should have double checked that.
For the python
issue with fscl-xcore-c
, use find_program('python3')
(Cf. documentation).
is it possible to set an option for a subproject from a root project via the command-line?
Yes, same syntax as in ci_config.json
: -Dsubproject:option=value
.
Is it possible to set up a similar sanity-checking system for the sake of saving time when creating these wrap files?
I'm not sure I understand what you're trying to do with that repo.
Isn't this already what tools/sanity_checks.py from this repository does?
The idea would be to have a redundant layer for testing new upstream wraps before considering adding to the WrapDB project. But that's off-topic so let's see, can I use just the python3 command for the test generator step in my projects or am I required to use find_program? How can I skip MSVC?
The idea would be to have a redundant layer for testing new upstream wraps before considering adding to the WrapDB project.
I run tools/sanity_checks.py from a terminal shell all the time. That's how I test wraps before considering adding them to the WrapDB.
But that's off-topic so let's see, can I use just the python3 command for the test generator step in my projects or am I required to use find_program?
This is a misunderstanding of what commands are, what meson's find_program()
is, and how the CPython project distributes itself.
Your repository uses the python
executable, which doesn't always exist, there are a number of ecosystems that only provide python3
.
Switching from "python" to the version with a 3 at the end, fixes the problem.
(meson's find_program()
will be internally run, if you don't manually specify it, on the first argument to run_command or custom_target if it is a string instead of a found/built executable/target.)
How can I skip MSVC?
The error message indicates that the reason MSVC is aborting is because you're using an unofficial experimental-phase std, which implies... yeah.
Just changed it for both the main and release branches and it's still complaining about the python executable being undetectable
Because it's still using https://github.com/fossil-lib/fscl-xmock-c/archive/refs/tags/v2.4.0.tar.gz
?
You have not made a new release, the wrap for fscl-xmock-c
is still using v2.4.0, which does not have your last 2 commits, no?
Shouldn't it point to the release branch? I prefer not to make a bunch of new releases just because of some small change for every issue. Shouldn't there be a checklist for the requirements of what to check when preparing a project for upstream?
Use a temporary overlay (patch_directory = …
in your wrap) until you have something that work, then make a new release.
What am I testing against?
Is it possible to have some of my projects depend on other projects? Algorithms and Structures depend on ToFu.
Just adding the Fossil Logic C API, guide me if I missed something.