Ruslan Kuprieiev

@Eugeny gcloud CLI is an important part of the ecosystem and it is a bit odd to try to use names that are not supported by it. There was some...

@Eugeny Makes sense. Would it be an option to just go through those files through API and rename them to comply with limitations? Merging this PR will break versioning support...

@Eugeny You are correct, # is chopped if it makes it all the way down to `_request`, but if fs is `version_aware`, it gets interpreted as generation in `_split_path`. I...

I don't know your situation, but overall, having # in filenames is probably going to cause you more problems later on, so to me, if possible, just renaming them sounds...

Just to be clear, we totally understand that this is coming from the need and don't want to discourage you. We might even run into the same problem ourselves later...

This is my first time seeing `local_file`, but looks like dirfs could indeed simply pass it through similar to how we do that with `sep` in https://github.com/fsspec/filesystem_spec/blob/02c557c11d738eaf9d14ff5f350055dd2dd127ff/fsspec/implementations/dirfs.py#L48 @jdonnerstag Maybe you...

My 2c: we've also introduced `fs.path` for convenience methods like `name/parts/parent/parents/join/etc` in dvc https://github.com/iterative/dvc/blob/master/dvc/fs/path.py Compared to using path-like objects as we did before, it has better performance when dealing with...

After using our `fs.path` for almost a year and making attempts (private) to create a patch (and finding a lot of `fs.path` already used and also some existing path-manipulation methods...

Other stuff got in the way, so I didn't manage to contribute it yet 🙁

@agrinh Finally getting around to moving fs.path to plain fs methods in dvc, so hoping to get around to contributing it to fsspec around new years 🙂 (said that before,...