Hm, we have been adding quite a few additional compilers over the last years. Is this really useful? Just want to prevent that our compilers directory becomes a dumping ground of anything that someone found useful over the times at some time

It is supposed to be as useful as mypy/pyright as Python type checkers, an analog to ruff as an alternative to established linters; released a few days ago (15k stars so far). To avoid becoming a dump, would a cleaning pass checking for the popularity of older compiler files help?

Offering a configuration variable to change the string value of a builtin option, even one contained in a plugin, seems to suggest we either don't have a useful compiler plugin or we have several. I don't think this level of configuration makes much more sense in a compiler plugin than it would in an ftplugin, for example.

Historically any compiler submission was accepted, they're usually stable outside the Javascript hellscape and not much of a maintenance burden. However, this is a new tool. Will it still be around in five years?

Removing any of the compiler plugins is a backward compatibility failure.

seems to suggest we either don't have a useful compiler plugin or we have several

I'm not sure if I completely follow, but naturally the idea is rather the latter, it can be useful not only as a globally installed tool, but also if specifically installed inside a project (uv run) or as a one-short tool (uvx)

Historically any compiler submission was accepted, they're usually stable outside the Javascript hellscape and not much of a maintenance burden.

Yes, as with the addition of new file types, there will be many who appreciate and few who suffer by finding another compiler file included.

However, this is a new tool. Will it still be around in five years?

It has already now a large user base. Similar to uv, ruff, ... they are recent but many times faster variants, besides other selling points, of existing tools in the Python landscape and therefore find quick adaption.

Of course, one can wait, and I'm in no hurry. The idea behind submitting it here shortly after was maximizing outreach to Vim Python users interested in using this fast type checker as a complement to the existing linters; every other mode of distribution felt roundabout and redundant.