Root Access

Mazura

12:03 AM

Which one of us is doing it backwards? The OS goes on the SSD, games on HDD, or vice versa?

@Burgi Try crouch. What game?

Burgi

@Mazura tried that, crysis 2

Mazura

sry, don't have it

rahuldottech

1:59 AM

@Burgi I'll try

@djsmiley2k me, mom, sister. Pretty high usage

allquixotic

2:27 AM

BOLTR: Anatomy of a FAIL. | Makita not you too!?

►

@JourneymanGeek AvE loves his woof

bwDraco

3:17 AM

@allquixotic woof

allquixotic

@Bob tell the foxes to go away D:

bwDraco

...I need to play more strategy games.

Bob

@allquixotic I want C++ and UTF-16 to go away :(

=> en.wikipedia.org/wiki/UTF-16#Description

> It is possible to unambiguously encode an unpaired surrogate (a high surrogate code point not followed by a low one, or a low one not proceeded by a high one) in UTF-16 by using a code unit equal to the code point. ... Windows allows unpaired surrogates in filenames and other places, which generally means they have to be supported by software no matter how much the Unicode consortium protests.

That's just pure wtf.

what is an unpaired surrogate

why do they exist

what is life

6

bwDraco

@Bob The Windows NT kernel uses UTF-16 internally. Getting rid of it would require rewriting large portions of the core system.

Bob

@bwDraco that is totally not the point

unpaired surrogates shouldn't be a thing

at all

bwDraco

3:28 AM

It's not valid Unicode.

Bob

pretty sure that violates the UTF-16 spec

but Windows uses them

WHY

WHY IS THIS HAPPENING

5

bwDraco

But I suspect this is, again, ingrained into the kernel.

Bob

you know what

bwDraco

Again, the Windows NT kernel uses UTF-16, and I suspect there may be code deep inside the OS which relies on this behavior.

Bob

I'm just gonna ignore any characters outside the BMP

this is insane

allquixotic

3:30 AM

that's ... a lot of dead foxes prepping for butcher

Bob

they probably died from Unicode poisoning

stabbed by a surrogate pair

allquixotic

lol

Bob

then clobbered with a bunch of unpaired surrogates

allquixotic

specifically that

Bob

if (cmd[cmdEnd] > 0xD7FF && cmd[cmdEnd] < 0xE000)
		{
			// ignore all surrogates, high, low, paired, unpaired, whatever
			// we don't care. seriously. fuck unpaired surrogates.
			continue;
		}

allquixotic

3:33 AM

(see log on left)

@Bob why are you even dealing with them? this is what libraries are for! :P

Bob

@allquixotic project goal: write a really simple wrapper that enables VT processing and calls launches a child program

allquixotic

also, if Windows allows unpaired surrogates in file names, does that mean a file name in NTFS can technically be any arbitrary stream of bytes (as long as it's divisible by ... what? 2? 4 bytes?) up to the max file length?

Bob

pulling in a unicode library is not really simple

also this is mostly C code

because I really really cbf converting from LPTSTR to std::string/wstring depending on _UNICODE and WHY IS THIS SO HARD

allquixotic

and I guess you can blame C++ for not enforcing a sane standard for string formatting but this really sounds like an impl detail of Windows/its VFS layer than a C++ problem

Bob

it was meant to be a simple CreateProcess

but then I realised I need to strip the first param from the command line before I passed it in

which means string processing

allquixotic

3:38 AM

and are your users downstream regularly passing you unpaired surrogates?

Bob

@allquixotic oh, I fully blame Windows for making unpaired surrogates a thing at all, but I also blame Unicode for making surrogate pairs (and generally variable-length chars) a thing. And I blame C++ for not providing a standardised API for dealing with them.

allquixotic

or are you just worrying about that remote possibility?

Bob

@allquixotic The latter. This is mostly an internal use wrapper (that I'll probably chuck on github at some point)

Oh of course array lengths can't be constant

ffs C

allquixotic

variable length chars are historically necessary, and I think if we standardized on fixed length always using 32-bit, we'd eventually exhaust it with an emoji of a dark haired woman with South Asian complexion feeding artificial milk formula from a bottle to a baby goat, and switch up each aspect of that description with every other possible thing that could be

Bob

hm... now I need to decide if I want to go C and malloc or go C++ and new

technically this is a C++ compiler...

allquixotic

3:42 AM

you'd need blond haired woman, blond haired man, dark haired man, etc. then Caucasian complexion, dark complexion, etc., ... feeding natural milk, ... from a bowl, ... to a calf, etc.

Bob

@allquixotic Even if that were how it's encoded in memory, I want a nice easy-to-use built-in wrapper that lets me write my code without ever knowing there's multi-byte codepoints.

allquixotic

and the full factorial of all possibilities

Bob

@allquixotic technically? those are handled with separate code points

which are perfectly fine

as in, it might mis-render, but nothing will explode if you split them up

if you split up a surrogate pair ... welp. actually, no, apparently that still works fine on Windows because reasons

allquixotic

soon we won't have any more languages, we'll just have Unicode emoji and communicate entirely in pictograms

we'll have an equivalent of the Japanese radical keyboard for emoji and construct a scenario based on the component parts to build the right emoji

😕➡️🏇 🏇 🏇 ➡️ 😐 😊 👍

Bob