@Pavel Does a "language" here don't need to be a "programming language"? (If so, I would request some trivial languages like HQ9+ and Hello in order to increase the language counter.)

@A_ It needs to not fit the PPCG definition of a language (Must be able to do addition and primality checking), must not be completely stupid (mostly stupid is ok)

There are several languages on TIO that I'm willing to bet haven't been used since a day after they were added

@Pavel Is a text editor like kilo stupid? (It is definitely a language.)

@A_ We've allready got ed as a text editor, as well as V (vim) if you count that, so why not

You do have to be able to do addition, as previously mentioned

And I'm not convinced Kilo actually allows anything intelligent

It seems to be less capable than nano

"mostly stupid is ok"

Unary addition can be done quite easily by appending text.

@Pavel Also, "It needs to not fit the PPCG definition of a language"

I think you understand what I meant there

@A_ Let's say, then, that the buffer contains two unary numbers. What keystrokes would you send to Kilo to add them?

@Pavel If they are in separate lines, I would just enter a backspace. (After I press PageUp and move the cursor down.)

...fair 'nuf

What about primality testing

I see what you mean here.

Can I use the file system as a lookup table?

Can Kilo even load a file without needing to be closed and opened again?

Whoops, my addition procedure was invalid. The correct sequence is PgDown, PgUp, move the cursor down, and backspace.

Well, although loading a file is unimplemented in the original Kilo, there is at least 1 that implemented this in the 522 forks.

By the way, mine does.

^ Then, maybe. BUt you have to actually demonstrate that it can be done.

Press Ctrl+R, type in the number, press enter. (The number will not be shown to the console.) Then, the results will be appended to the buffer.

how does this check primality

By using the file system as a lookup table.

I don't see how that alone allows solving the problem

Before that, you hard code the numbers into the file system by sequentially loading the results into the file system (not re-opening the editor).

Even if just looking it up in a table was a valid solution (I don't think it is, and either way you can't put a bunch of files into the filesystem before execution on TIO), I don't see how you transfer a number from the editor's buffer to the ctrl-r input

Anyways, how do you do primality checking in ed?

Ed supports regex

It's actually fairly powerful

@Pavel You type the number in a separate buffer, and then for the Ctrl-R, the command clears the other buffer and then reads a file to the main buffer.

@Pavel Right, we should add regex to TIO too.

I mean the regex that is in ed.

Hmm, does nano support regex?

nano seems to support regex, but I can't use it.

TIO already has too many text editors, and there is no need to add more.

(What's the point of getting all this fame for a server with a pile of useless languages?)

<hr/>

@Pavel Why is Malbolge on TIO when nobody has written a prime-checking program?

Malbolge is programming language for finite automata, the same as *, and, for example, Befunge-93. Therefore Malbolge is formal programming language, the same as *, technically the same as recursively enumerable languages when it comes to programming language definition (albeit formal languages are less powerful). – Krzysztof Szewczyk

@A_ Because it's interesting to a lot of people

There really aren't hard rules for what's allowed on TIO and what isn't

@Pavel - actually, I expect that there's one hard rule for what's allowed on TIO: Dennis has to be willing and able to get it working on the server.

@Dennis What the actual hell:

jimoto@DESKTOP-S9INNBV:~$ env -i coding/V/v <(echo -n "iHello world") -f /dev/null | xxd -g 1
00000000: 07 48 65 6c 6c 6f 20 77 6f 72 6c 64              .Hello world

I suppose I should get a fedora VM and try testing this on something that isn't WSL

Would it be possible for you to revert V while I'm testing this?

I'll do that when I get home.

@DJMcMayhem yeah, WSL is obviously inferior to VMs, especially since some of the distros aren't free...

like, €9.99 for sub-Linux stuff called "Fedora Remix"???

@EriktheOutgolfer But it's ridiculously convenient

Which version of fedora should I get?

This might be another clue: tio.run/##S0oszvj/PzWvTEE3U0E/v6BEv0y/TMFGIzU5I19BN09B/…

It should just be c381

looks like TIO uses Fedora 29

Cool

these are the complete release specs (if you ever wonder if anything changes)

Not for long, I hope. Fedora 30 is the latest version.

Is neovim one of the things holding you back?

It's a minor inconvenience on the severity scale.

@DJMcMayhem 0x07 is BEL, so it's supposed to produce a sound in your terminal.

Ha! No wonder I keep hearing that lol

it's the only one that you don't see with your eyes, but with your ears... :D

@DJMcMayhem Done.

Danke

@DJMcMayhem env -i turns UTF-8 off, so I guess it has something to do with that.

@Dennis Then why is this different? tio.run/##S0oszvj/PzWvTEE3UyE1OSNfQTdPQf1wo7pCjUJFRcr//wA

is the UTF-8 setting just for files?

The environment variable LANG (which is reset by env -i) is read and interpreted by vim. Á is converted to 0xc3 0x81 when you enter it in your terminal emulator (set to UTF-8) or TIO (also set to UTF-8).

@DJMcMayhem How does the hexdump work? EOF (-1) is 0xFF when converted to a byte, and the Unicode character U+00FF is 0xC3 0xBF in UTF-8.

Interesting

@Dennis The way I set it up was calling vim with the -w file flag (record keystrokes to a file), and then after the main vim had finished executing calling a secondary vim to open up the keystroke file and xxd it from within vim

This is the basics of how V works now: tio.run/…

Hm, no weirdness with env -i. Try it online!

But there is a newline.

Hmm, the encoding is wrong too: tio.run/…

No, that's fine. Those are the bytes that Bash passed to vim.

But there's no 0xc3 0xbf at the end.

@Dennis I do se fenc=latin1 so that it'll print the latin1 hexdump, not the UTF-8 hexdump

What hexdump do you get when you run the new V with iHello Wörld?

I'll have to check. I think I have an idea what could cause the EOF thingie, but I can't test it right now.

When you do, could you also leave the new V somewhere else on the server so that I can continue testing this?

@DJMcMayhem It depends.

On LANG, that is.

So setting LANG is what breaks it?

Hmm, they both have the trailing newline, but the first one has the right hexdump