Now that the parser tool is moved out of the repository, it makes less sense to use the taskfile-based approach for the
CI infrastructure. In order to make the repository more contributor-friendly, the documentation and configuration
checking system is now confined to the .github subfolder.
Now that the parser tool is moved out of the repository, it makes less sense to use the taskfile-based approach for the
CI infrastructure. In order to make the repository more contributor-friendly, the spell checking system is now reduced to
a single workflow file.
This approach allows the diff to be written directly to a file, rather than needing to sanitize the contents of the
output from the octokit/request-action before writing it to the file via the shell.
The index source file contains the normalized URL and the metadata that can't be derived from the library releases
("types" data and locked name) for each of the libraries in the Library Manager index. It's most appropriate to store it
in the same repository as the submission list.
I had intended to test these out but didn't get around to it. The workflow_dispatch might be convenient, but it's not at
all essential since the same thing can be accomplished by commenting on the PR.
Rather that committing directly to the index source file repository, save the index entry to a workflow artifact, which
will be consumed by the system that updates the index source file.
With the current submission system, it's common for people to provide the tag/release URL rather than the repository
URL. I think it likely the same sort of thing will occur with the new system. Previously, a very naive check was done
for the path component of the URL to have two levels. It turns out that the Gitlab repository URLs have more levels than
that. The superior approach is to use `git ls-remote` to verify that the URL is a Git repository.
Submissions of libraries to the Arduino Library Manager index can be done by submitting a pull request adding repositor
URLs to the list. The submissions are checked for compliance wih the requirements for addition to the index and as soon
as they are passing the pull request is merged and the entries are pushed to the index source file.
