Edouard Choinière

> It was designed as temporary(!) solution (written quickly, no tests available, only bugfixes, no improvements). A proof of the saying: there's nothing more permanent than a temporary solution

Maybe it is part of the test suite (either the code or fetching one not available). But in either case, it wasn't the only one that wasn't able to be...

Failures are weird, and Windows has way more, many tests exited right away, as instead of taking 50min, tests took 11 min.

> Mixing may not be good, especially when scripting comes into play. We'd need to verify by manually doing a single fix and seeing if this one is fixed, but...

Apart from the cygpath tool, what other conversions does msys2 do? There's this docs https://www.msys2.org/docs/filesystem-paths/ They say that they convert automatically env vars, process arguments. Native tools would work with...

Is it possible to have more context on this, to help out prioritizing (or give someone the will power to do it)?

@ninsbl I think it's ready, but just before, can you try (locally) if with ruff 0.5.3 there's no changes (PLR17xx something changed and covers a new case than before). I...

I wanted to reproduce with an nightly build of the upcoming 8.4, but I can't get an installation launching for now (bugs are filed in).

So it's still a valid bug to look into.

Hmm.. I'm not able to make the module run correctly, but it isn't because of loading a dll or segfault or the sort. I'm using grass-dev 8.4-313-f4d8c62ac-1 from OSGeo4W, without...