Switch to using unicode when parsing the command line on windows

[Rageoholic]

This should help out with issue #534. Just using GetCommandLineW and then adding in some fixups to make it compile, return utf8 still, and make the tests pass.

[LemonBoy]

The tests are failing because mips is a big-endian machine, the result from nextCodepoint needs to be byte-swapped or the tests disabled on everything but Windows.

[Rageoholic]

Changes are applied. Looks like the W functions should give you back little endian codepoints. MS has AFAIK never released a version of windows that runs on big endian so I don't care but doing the fixup should be really inexpensive or free and really who puts argument parsing on the fast path anyways. Also we can just run the tests on any given machine and you'll know if you broke it.

[LemonBoy]

Changes are applied.

Tests are still failing because skip/next are not using littleToNative.

Looks like the W functions should give you back little endian codepoints.

UTF-16 or UCS-2 ?

Also we can just run the tests on any given machine and you'll know if you broke it.

That's a great argument, I like it 👍

MIPs

The lowercase s is making me extra sad for no reason :(

[Rageoholic]

The lowercase s is making me extra sad for no reason :(

I wrote that early in the morning and now it's making me sad.

UTF-16 or UCS-2 ?

It's definitely UTF-16. I crawled through MSDN to check. Frankly though they can't port to a big endian architecture without either breaking the people who assumed little endian architecture (given I just did I can't blame them) and people who actually do the fixup on big endian architectures. Hopefully enough people were like "I'll just use a library" and the library did the right thing so that MS can keep it's word without too much pain.
https://docs.microsoft.com/en-us/windows/win32/intl/using-byte-order-marks

[daurnimator]

It's definitely UTF-16. I crawled through MSDN to check

Where/how? MSDN usually fails to note that when they say UTF-16 they really mean UCS-2.

[Rageoholic]

It says utf-16 in the link I posted

…

-------- Original Message --------

On Nov 28, 2020, 6:26 PM, daurnimator wrote: > It's definitely UTF-16. I crawled through MSDN to check Where/how? MSDN usually fails to note that when they say UTF-16 they really mean UCS-2. — You are receiving this because you authored the thread. Reply to this email directly, [view it on GitHub](#7241 (comment)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/ACC47ZJO6WFSJREUDEPR4CDSSGILVANCNFSM4UFOMV7Q).

[andrewrk]

Thanks!