@Dennis Thanks! Would you mind pulling again?

@Pavel I'll take a look. Probably tomorrow.

@MDXF Done.

@Dennis The page for downloading prebuilt binaries seems to be broken, so you might need to build from source.

Building from source requires Python, OCaml, and C++. What could possibly go wrong. :P

I'm building it right now, I could upload a tarball myself.

Nah, it's fine. Just tell me if there are any tricks/prerequisites I should know about.

Not really, other than that it takes forever. I finally understand xkcd.com/303.

@Dennis Thanks :)

This is a fairly clean Fedora install that I got two days ago, everything needed seems to be in the gcc-c++ and ocaml system packages, since I haven't installed any *-devel packages or anything like that.

Well, there are a bunch of warnings, but I can't really understand what any of them mean and they aren't fatal.

@Dennis LibSDL is required for an optional component, but since running using a graphics library on TIO is pointless, it shouldn't be needed.

Also, I don't know if you want to call them "Command Line Options" or "Compiler flags", since most of the available flags are just passed to gcc when building and running the produced C++ code.

Also, file must have .flx extension, so ln -f .code.tio .code.tio.flx

Mm, I just discovered that on my installation the repl doesn't work. Running anything at all causes a compiler error.

Actually don't mind ^, I was using the REPL wrong. Unlike sane REPLs, the felix one actually requires provideing semicolons and using print statements.

@Dennis Emojicode 0.6 will invalidate all existing Emojicode ≤0.5 programs, when it's out can you change the current "Emojicode" to "Emojicode 0.5" and add 0.6 as "Emojicode"? (On a second thought that'd invalidate current links too...)

@betseg It's possible for one language to have multiple names, for example MDXF's SimpleStack and Implicit.

@betseg The language's name and ID are fairly independent, so renaming Emojicode to Emojicode 0.5 is not a problem, as long as the ID remains emojcode. Thanks for the heads-up!

@Mego Done.

@Dennis Can you pull Cubically please?

@MDXF Done.

@Dennis Thanks!

@Dennis Could you pull Husk as well, please?

@Zgarb Done.

Thanks!

@Dennis Do you have a way of testing if a language on TIO is up to date with the repo?

Same as everywhere else: git show

Could you test whether Flipbit is up to date?

@cairdcoinheringaahing Just execute git -C /opt/flipbit show on TIO.

@Dennis Bash, presumably?

Or any of the shells, yes.

@Dennis Thanks, it seems to be up to date

@Dennis maybe add it here?

@Dennis I've just noticed that Clean doesn't use UTF-8 (or any particular encoding) for its source files or output, which isn't reflected on TIO at the moment. It handles source files as bytes instead of characters, and doesn't care about what bytes are in strings (except that it expects codepoints 0-127 to be ASCII).

This leaves pretty much every ANSI variation as valid, but whatever TIO is doing at the moment doesn't work properly. Specifically, compiling the output of tio.run/##S85JTcz7/z83P6U0J1UhNzEzjysztyC/… on both my linux and windows installs doesn't error. It does when I copy the output from TIO into the codebox.

@Οurous You never see the output encoding. The output textarea contains characters, so when you copy them to your clipboard, either your browser or your OS (not sure which) decide the encoding.

Source files are UTF-8 encoded though. That's a problem for quite a few languages, and sometimes not desired at all. The only solution is a user-selectable encoding.