Add split_windows_volume_prefix() function and tests.

[aznashwan]

WARN: this code may NOT run correctly on OTP because of this compiler bug: gleam-lang/gleam#2733

[aznashwan]

I just wanna say I really personally find what gleam format did with these nested tuples a bit less readable that the original version...

[lpil]

Thank you.

What's the motivation for this addition? I've not come across a function like this before- what are some scenarios in which I would use it?

[aznashwan]

It's generally useful to be able to split the drive of a path you're dealing with on Windows for a variety of reasons like determining if it's a local drive, remote UNC path, registry hive path, etc..., or processing/reusing later if it's a UNC share (like extracting the hostname/IP from it).

Python's stdlib has one FWIW: https://docs.python.org/3/library/os.path.html#os.path.splitdrive

Even if you don't want to have this function be public, it's still useful as an internal function to be reused in split() and is_absolute() as seen in #5, as currently both of those do NOT work correctly on Windows.

[lpil]

Makes sense! Thank you. I've left some notes inline

[aznashwan]

@lpil changes done, I've even squashed the commit into one (minus the version bump and changelog update which is still a separate commit)

Thanks for you patience, and please re-review at your leisure!

PS: considering it's important to test with both /, and \, there's about 90 individual tests now 🤣

[aznashwan]

@lpil just pushed the latest iteration on it with the added edge case you suggested.

Note that after more consideration of split_windows_volume_prefix's integration into is_absolute() it became apparent to me that drive-relative paths (e.g. C:Users\Administrator, which is interpreted as C:\$CWD\Users\Administrator) will need some form of differentiation from drive-absolute paths (C:\Users\Administrator), so I have elected to make the change of having the separator between the drive and the rest of the path remain in included in the path.

This will allow the later update of is_absolute() to work correctly in both cases, whereas without the included separator it would have potentially been ambiguous.

[aznashwan]

Hey there @lpil, glad to have some time for Gleaming again.

Following the Erlang pattern matching fix reaching GA in Gleam 1.2, I have went ahead an updated this PR, PTAL whenever you can! 😄