João Távora [2023-04-06 13:49 +0100] wrote:
> On Thu, Apr 6, 2023, 12:22 Michael Albinus <michael.albinus@gmx.de> wrote:
>
>> > Instead of using your distribution package manager, I recommend
>> > installing pylsp via the official Python package manager, pip, and
>> > doing it like so:
>> >
>> > pip install "python-lsp-server[all]"
IME this may not be sufficient to get the latest version of pylsp's
dependencies, such as autopep8, which I presume is the component
responsible for the differences in formatting across development
environments.
IOW it may be necessary to explicitly 'pip install -U autopep8'.
If that were true, GitHub CI would be failing, since it used this command, and it's not. Tests run on my laptop and I've never run that command, just the one I suggested.
>> I disagree. Running regression tests towards bleeding edge development
>> version of pylsp is not the intention. You cannot expect, that everybody
>> running Emacs tests has installed pylsp like this. But she could have
>> installed the Debian pylsp package.
>
> Sure, or she has a 'pylsp' that prints the complete works of
> Shakespeare. But these tests are designed for the number one
> recommended pylsp installation method. It's AFAIK not the
> "bleeding edge" (which would be a Git installation). And tests
> have been running fine with that installation method for a
> number of months now, maybe even years.
FWIW the autopep failure in bug#61637 was with latest pip-installed
pylsp and autopep8, and the patch in that report worked around it by
accepting one of a number of possible formattings.
A version check for autopep8 (rather than pylsp) may have been an
alternative solution
Who knows? Did you try it? If it works i would have been happy to accept it Noone proposed it. Instead, some truly vague and complex vapourware things were proposed.