WHAT—

... "They point to how security researchers hated Visual Basic 6 binaries due to the complexity of reverse engineering the software, the presence of a Lua obfuscation layer in the 2012 Flame malware, and the Grip virus, which contained a Brainfuck interpreter coded in Assembly to generate its keycodes, as examples."

https://www.theregister.com/2025/03/29/malware_obscure_languages/?td=rt-3a

@cstross

Fun fact: VB6 uses its own internal libraries for GUI stuff. Like, every form & widget is rendered not by any standard Windows framework, but by VB6 ITSELF. That's because back when VB was conceived no such frameworks existed.

That's just one example. Don't even get me started on its native JIT interpreter, where you can (optionally) compile an exe from VB tokens that the interpreter processes at runtime. Do that, and there's no actual code to decompile! It's kinda neat.

@cstross I remember hearing Delphi also produces binaries that look as if an obfuscator went over them…

@cstross Lisp is the new Nazi salute {{{}}}

@cstross even just having a non x86 binary already made people ignore issues (Cf: intel firmware in the chipset back circa 2010 )

@cstross Seems eminently sensible to me.
If I were inclined to write malware (which I am not) I'd use Brainfuck too. Maybe not code the interpreter in assembler though, I mean, there's INTERCAL for these things.

@Uilebheist How about a dialect of Brainfuck in which the various symbols it uses for commands are replaced with unicode glyphs in a homograph attack (glyphs that display identically but are actually separate characters, eg. the latin, cyrillic, and greek letter "o" are all different in unicode)? Preferably whitespace homoglyphs or combinations: for example, you could encode all 8 Brainfuck instructions using three-character combos of space and zero-width space.

@cstross
Implementing BF in assembler would be easier than getting BF to do anything useful at all. 😄

@CppGuy Look, if you can write a CLC-INTERCAL compiler in INTERCAL (as indeed CLC did, back in the day), then ANYTHING is possible …

@cstross A. Brainfuck. Interpreter.
What’s next? And embedded Malbolge transpiler transforming Befunge to JavaScript?

(The worse part would be the JavaScript one)

@cstross Reversing Hadkell is a royal pain. Who needs function calls? Have a huge fucking jump table!

@dat @cstross As noted here: https://polyglot.city/@Uilebheist/114258519489252232

@dat @Uilebheist @edwinb Space/tab/linefeed combos are MUCH too easy to tell apart, though! They don't render the same way—you can edit whitespace source code in vim with :se list so you can tell tabs and spaces apart!

@cstross Rust was a similar issue recently because the tooling for RE'ing it wasn't great and it doesn't cleanly decompile to pseudo-C in most cases.

@dat @Uilebheist @cstross to be fair, I did this after a pub trip in 2003 while avoiding writing a thesis. I love that I still get these random mentions 22 years later.

Incidentally, its birthday is tomorrow (mostly a coincidence. April fool jokes could still be funny back then, at least we believed)

@cstross That sounds an eccellent improvement over Whitespace¹

There was talk of a dual CLC-INTERCAL and Whitespace compiler which would recognise the blackspace as INTERCAL and the whitespace as Whitespace. There had even been emails exchanged between the author of these two languages about that.

So...

I suppose it'll have to be done.

¹ https://esolangs.org/wiki/Whitespace

@cstross OK, there are enough zero-width space code points that one could use them and have zero-width Brainfuck programs. Then they could be inserted into unsuspecting text.

@edwinb @dat @cstross I know I emailed somebody at St. Andrews about including a whitespace compiler hidden in CLC-INTERCAL. Was that you?

It could still happen.

@Uilebheist @dat @cstross That is possible!

I can only apologise if I didn't reply (which is also possible) because in the deluge of all the Things I have to prioritise students and colleagues and end up missing fun things as a result.

@edwinb @dat @cstross You did reply.
But looking through my mailbox 20+ years later, I can't find it.

@dat @cstross @Uilebheist I am pleasantly surprised and also wondering if I met you at edlug several lives ago...

@cstross I think this was accidentally released a few hours too early. 3:O)>

@edwinb @dat @cstross That is definitely possible, likely even.

@cstross This... gives me an idea.

Or rather two ideas.

Zero-width INTERCAL, using different zero-width characters to encode the 5-bit BAUDOT encoding accepted by CLC-INTERCAL. Example: "‌‍"

But this isn't quite horrific enough

Combining INTERCAL, where all input is provided by combining characters, attached to visible (but ignored) text. Example: "᪰᷁𐰩"

Zero-width combining INTERCAL in which all combining characters are attached to zero-width characters. Example: "᪰᷁"

@Uilebheist @cstross Reminds me of Zalgol: https://leahneukirchen.org/caudex/zalgol.html

@leah @cstross Well, anyway, Charset::ZeroWidth is now in the repository for anybody who needs to write zero-width INTERCAL programs.

(Subject to testing when sober, which isn't now).

https://git.sr.ht/~uilebheist/CLC-INTERCAL/tree/main/item/CLC-INTERCAL-Base/INTERCAL/Charset/ZeroWidth.pm

@cstross My brain has just been hacked reading up on Intercal:
https://en.wikipedia.org/wiki/INTERCAL

Ive got James Brown's "please, please dont go" looping in my head.

...I darent wonder how many pleases it would take for someody to hand over their cryptowallet?

fwiw, you may find this Makefile interesting:
https://git.sr.ht/~indieterminacy/1q20hqh_oqo_parsing_icemaqr/tree/main/item/Makefile

@cstross @Uilebheist
Before COVID, I used to play Gõ with one of the people who invented the "Whitespace" language. It seemed obviously an extension from the black&white simplicity of the game.