« first day (41 days earlier)      last day (536 days later) » 

00:19
here's the japt refactor
I couldn't clean up C's "puts" line without crashing Japt's parser, so I left it as is and took most of the byte saving from Brain-Flak
 
12 hours later…
12:21
@Chance Perhaps for cops you should require them to write a different polyglot. For instance cops do task X in some set of languages and robbers have to do task Z in the same languages.
You would need to show that each language can do task Z too, otherwise you would find a single language that can do X but not Z and then polyglot it with as many languages as possible to make an answer that can't be cracked
This way cops have to do work to earn more difficult to crack submissions
 
3 hours later…
15:06
I worry that this challenge has made a lot of "most languages" polyglots trivial to produce.
you could literally remove half of the languages in here, and it would still be the biggest on PPGC (I believe).
And quite a few of our languages are outputting strings too, which are trivial to change in most cases.
Personally, I want to task the community with creating"poly-nots" because I don't believe there is a resource out there that provides information on language incompatibility.
Quite a few times in this project, I've found myself wondering if I was opening more doors than I was closing.
@Chance What do you mean by that
oops, my last patebin was the wrong link. This is the Japt refactor. pastebin.com/4jgqn5xn
@WheatWizard For example, when I added C++ for example, I had to change the top 2 lines a lot (and make them very restricted). I worried that the number of C like languages that might be added to the ployglot is less than the number of 2D languages I closed off the polyglot to more 2D languages.
15:22
Oh so you are closing more doors then opening
right.
this refactor is looking great!
down to 1282 bytes.
1277
are you adding a language in the refactor? or just creating a base for future work
just a base so far.
I suspect it opens Ksh, but I haven't proven it yet.
 
1 hour later…
17:05
@ais523 got a ksh/zsh/bash polyglot
@WheatWizard But it breaks the Brain-Flak code tio.run/nexus/…
here's the ksh verision
^ can you do that?
^ same thing but shorter
looks like that was the secret sauce
thanks
I stared at that for waaaay to long to find the difference
I might be able to make it chorter
I think the python code intentionally has too many parens for old Brain-flak stuff that are just getting in the way now
@Chance could you do case argv[1]+${p:u} to save bytes?
17:20
Oooh. that's a good idea
You can get rid of a {} in the Brain-flak too
actually you can't get rid of a {}
@Chance: neat; looks like we're going to get a working polyglot soon
I'm going to hold off until you've posted it, as this sort of mass rearrangement likely changes which languages are viable
prelude is proving challenging
17:35
@Chance Have you tried removing parens from the python code? I went back and read some of the old posts and they seem to be there to aid Brain-Flak code I destroyed. Might help with prelude but it should certainly make it shorter
Oh. are you dumping all those values first?
I didn't realize.
yeah I have to empty the stack so that I can figure out which language I'm in
cool. reverted the script-lang print statement to it's pre-B-F glory without error.
Nice!
Did that help with Prelude?
yeah. only cubix now
I threw a superfluous parens into the B-F to detokenize a couple.
ok, this is weird
cubix runs when I look at only it in tio.
but the test driver shows an error
17:52
copy-paste issue, perhaps?
those ESC characters can be pesky
I'm not seeing it.
maybe a bug in the cubix wrapper?
try the cubix debugger for a third opinion?
It works fine on the online cubix interpreter
based on expanding the debug view, it looks like the Cubix interpreter is having problems lexing the code
wait, no
it's in the A command
which takes input
I bet the issue is to do with the way we're providing input to the polyglot, somehow
maybe I can just swap that Fac for a Do
18:00
err, are the Cubix and Japt wrappers meant to be identical?
let me check a previous entry
I just dumbly copied the japt wrapper when I submitted cubix.
it worked so I shrugged
it's believable
but I bet what's happening is that the Cubix wrapper is providing an undefined value rather than an empty value as input
because Cubix and Japt take input differently
echo -n 'require("/opt/cubix/cubix.js").run(' > cubix-wrapper.js
cat polyglot.escaped >> cubix-wrapper.js
echo ', "");' >> cubix-wrapper.js
@Chance: change the Cubix wrapper to that
the Japt wrapper doesn't provide input correctly
cool. that did it
100%
1292 bytes
VIP score: .006621
now all I have to do is wait for the post
glad I could help, and glad it was just a driver change so that I'm unspoiled for the next entry
(I didn't even read the code you submitted; it was onscreen but I didn't try to mentally parse it)
can you check it in Deadfish~?
18:12
sure, give me an updated link (or will the existing link still work?)
I'll check it in Incident too
here's a new one for safty
I need to verify that an old enough version of Japt had the parsing error too.
err, Incident's broken
it appears to print infinitely many 3s
I must have obliterated a business token somehow.
looks like you obliterated several, in fact
lol
oops
18:15
the culprits appear to be 01 and 10 overlapping tokens that are wanted
works in Deadfish~, though
oh, cool that's easy
I had to tweak whirl.
cool. all re-solved
give me a new link so that I can verify?
what, you don't trust me. :P
works in Deadfish~ and Incident now :-)
18:31
yay!
 
2 hours later…
20:58
does anyone know why ksh doesn't match this case? a=a;case $argv[1]+${a:u} in 1)echo 50;;
echo $argv[1]+${a:u} returns [1]+a

« first day (41 days earlier)      last day (536 days later) »