vdoc: cleanup get_module_list, add unit tests

[ttytm]

Cleans up the get_module_list function and adds simple unit tests for it and get_ignore_paths. Hoping this creates a better base for further development of such features. E.g. I'd like to add the functionality for a .vfmtignore - similar to a .prettierignore, but for now with the simple functionality of the .vdocignore - in a follow up.

[JalonSolov]

Did you mean .vdocignore? Alex does not want .vfmtignore - he doesn't want ANY options to vfmt, as it is supposed to stay opinionated.

[ttytm]

The commit is related to cleaning up tests the .vdocignore functionality.

A feature I would like to implement is a .vfmtignore. There already is // vfmt off which is the same functionality that a .vfmtignore would do.

[ttytm]

I'm coming to the .vfmtignore as you currently can't run v fmt on the. v-analzyer project.
V is ignoring formatting for some internal files hardcoded. For other projects and for the V project itself, instead of hardcoding / adding vfmt off to every problematic file, a .vfmtignore would be just logical for me. There would be no new options.

But I'm not too attached to the feature, so if it's not wanted improving .vdocignore is also fine 👍

[spytheman]

The commit is related to cleaning up tests the .vdocignore functionality.

A feature I would like to implement is a .vfmtignore. There already is // vfmt off which is the same functionality that a .vfmtignore would do.

No, it is not. The difference is scope.
A feature like .vfmtignore will shift the balance towards making it extremely easy to just put a * pattern in it. I want the balance to be towards most projects using vfmt, and thus towards consistent formatting across the entire V ecosystem, since that is the easiest.

To do that using // vfmt off, you would have to put that in every single .v file, since it was designed for local changes, not for global ones, like a .vfmtignore file would allow.

[spytheman]

V is ignoring formatting for some internal files hardcoded. For other projects and for the V project itself, instead of hardcoding / adding vfmt off to every problematic file`

It is better for the ecosystem to gradually fix those, not to make it easier for projects to use a global .vfmtignore file.

[ttytm]

The decision to not have a .vfmtignore is comprehensible. I'll won't submit a PR. The issue of not beeing able to easily run v fmt -verify/-c . on a project like v analyzer and tell if there are formatting errors next to internal errors was solved by a recent PR extending the exit codes.

[spytheman]

v-analyzer can just use .vv for those files, similar to what V itself does.
For example, after:

rename s/.v$/.vv/gm metadata/stubs/*.v
rename s/.v$/.vv/gm editors/code/syntaxes/tests/*.v
rename s/.v$/.vv/gm tests/testdata/documentation/*.v
rename s/.v$/.vv/gm tests/testdata/types/*.v

vfmt does not complain anymore:

#0 11:58:29 ᛋ main /v/v-analyzer❱v fmt -c .
#0 11:59:28 ᛋ main /v/v-analyzer❱