« first day (80 days earlier)      last day (497 days later) » 
00:00 - 18:0018:00 - 00:00

6:00 PM
@SnoringFrog Did you get everything to work? I'd like to start re-fixing my what code
late for a meeting, can't edit atm, but ai fir at fir "6" should get it working for yours @MDXF
@SnoringFrog Oh hm okay
@SnoringFrog Where?
Instead of ai my at my "6"?
@MDXF Probably
@WheatWizard Yeah that was it.
that seems reasonable
6:02 PM
Anyone want to test with the languages that aren't in the driver? I've checked what
@WheatWizard could you test the langs you have installed?
Yeah, evil was landing on the m and then running several things it shouldn't have been. There might be a 1-char fix but I don't have time to find it yet. might be too complex a fix to be worth it though
@MDXF Ok give me a second
@SnoringFrog I'm just swapping my for fir or whatever and it works fine
@WheatWizard Thanks
deadfish works
incident does not
but... but... the tokens...
6:06 PM
Moorhen works
Wait... prelude fails. Why?
almost certainly vertically aligned parentheses
that is nearly always the reason
there are other ways to break it too, but those are much rarer
(and typically involve adding parentheses in such a way that you screwed up its control flow)
@ais523 I'm just inserting <space><space>(????!?!??!??!!!!???!?!??!!?!?!!!!!?!!!!?????!???????????????????‌​?!) after the main ( )... line
you added a new parenthesis in column 3, that's pretty early
6:10 PM
if that isn't vertically aligned with another (although at this point it seems fairly likely it would be), it'd act as a jump, jumping the Prelude operation to the next matched parenthesis
So should I just add () at the end?
the thing about Prelude is that it runs code a column at a time
(it's not vertically aligned with another)
so you look for all the columns with opening parens, then all the columns with closing parens, and match them
where that match ends is the point at which Prelude will start
(this is assuming you don't have digits to the direct left of the leftmost parenthesis in the program)
I have two spaces to the leftmost parenthesis in my line
Okay do you want to take a look at it? I'm very confused
6:12 PM
think of it this way: each column is a Prelude command
( means "jump to the column with the matching ) if there are no numbers to the left of this ( on the same row, or if the number to the left on the same row is 0" (approximately, there are some exceptions but they're rarely relevant)
There are three ('s on the column where I put it
we use that technique to get Prelude through to the near the rightmost side of the program, where the actual Prelude code runs
what's the Prelude error message, anyway?
note that a column containing more than one parenthesis is invalid Prelude, full stop, and thus has ot be avoided
6:14 PM
try copying the code to run in tio for prelude only
"Error in the program: More than one bracket: position 23"
that seems pretty clear to me
Fixed it
I just stuck three extra spaces before
Now: Incident
I tokenized ar apparently
Wait... but I didn't add any ar's
I hate Incident
did you remove one? or remove a token overlapping it?
I think SnoringFrog messed that up too :I
the whole point of Incident is that it's basically impossible to permanently break as you have so much flexibility to control the tokens
note that all that matters is whether the tokens are the same or different, Incident doesn't care about the spellings
6:18 PM
It's done
6:37 PM
A: Add a language to a polyglot

MD XF71. what, 1877 bytes #16 "(}o+?23!@)-("//*\Dv;'[af2.q]PkPPX'#CO)"14";n4 #/*0|7//```"` [-'v][!(>77*,;68*,@;'1,@10␉␉11)(22)S␉␉(1 P''53'S^'q #>␉ # 36!@␉` # #_>++++.>++++++::@-----x-.+? #`<` #<]}} +<[<.>>-]>[ #{ #x} # #=x<R+++++[D>+++++++59xL+++<-][pPLEASE,2<-#2DO,2SUB#1<-#52DO,2SUB#2<-#32DOREADO...

now you're going to end up making @Feeds look stupid :-P
Oh yeah it automatically sends posts. Whoops :P
The TIO link is getting really huge. What's a good URL shortener that StackExchange doesn't recognize?
don't use URL shorteners
it's good to have all the relevant context in the post itself
and the TIO link is a good way to encode that
if the URL shortener goes down, we'd lose the content of the post
Blah. I hate the lag when trying to edit
add the TIO link last then
6:42 PM
I did
but anytime I try to, say, golf something, it takes me three minutes instead of thirty seconds
In my experience most of the lag actually comes from the rundown section
Do we really need to keep including that?
having a regularly updated rundown is really useful
to be honest, I didn't originally expect it to become a problem :-D
Couldn't we just have it in the question?
(also many early answers had a rundown full of verification links)
6:45 PM
Like, just have a list of all entries?
It would be nice if the rundown included links to the posts where the language was added
putting it in the question would require everyone to be able to edit the question
@ais523 ...most everyone can
A: Add a language to a polyglot

MD XF71. what, 1877 bytes #16 "(}o+?23!@)-("//*\Dv;'[af2.q]PkPPX'#CO)"14";n4 #/*0|7//```"` [-'v][!(>77*,;68*,@;'1,@10␉␉11)(22)S␉␉(1 P''53'S^'q #>␉ # 36!@␉` # #_>++++.>++++++::@-----x-.+? #`<` #<]}} +<[<.>>-]>[ #{ #x} # #=x<R+++++[D>+++++++59xL+++<-][pPLEASE,2<-#2DO,2SUB#1<-#52DO,2SUB#2<-#32DOREADO...

@WheatWizard actually, I do like that idea
6:45 PM
And you could easily approve the edits
we clearly can't make the question CW because then all the answers would be
Yeah. And that would suck
right; Chance wouldn't even be able to downvote
due to losing so much rep
Yeah lol
What is going on? Still no votes on what answer? (*ba dum crash*)
I haven't had a chance to look at it in detail yet
6:47 PM
I love that language's name
also, I really ought to go get food, so bye for now everyone
I don't really care, I just took an opportunity to make a pun
Well, time to search for obscure languages, preferably ones with punny titles
I'm planning on adding an even older version of Python later today, so you'll be able to post if I finish that.
Sounds good
Driver improvement suggestion: Make FAIL more distinct from PASS, it's so easy to mistake the two
They're the same size font, same color that's not very distinguishable from the background, same string length
They've even got the same 2nd letter
I just do <⌘+f>FAIL every time
6:54 PM
Me too, but still
Yeah that's how I missed that evil fail at the end
we could probably just mess up it's alignment
@SnoringFrog I've already posted my answer so do whatever you want to yours
Well I haven't got to even testing literate haskell, but i worked out how the test driver works so i can actually test now
@MDXF I've been thinking the same thing about PASS/FAIL/????. you have the last answer, so feel free to update the driver. :)
maybe lowercase pass?
@Chance That would be good
7:00 PM
and maybe single char ?
@Chance Good idea!
Well actually, I don't think the ? matters
The XX throws me off, though
(Might just be dyslexic me)
Oh, yeah I had a lot of that stuff in there for alignment
just my aesthetic preference.
All right, changed PASS to pass, ???? to ?<3 spaces> and XX to --
aaand I done broke it
WHY?!? Changing XX to -- makes it print pass instead of ????
7:05 PM
I just changed "PASS" to " PASS". Makes the fails and ??? stand out. literally
Yeah it does
But I'd still like to make the XX more clear
Oh I fixed it
There was a part checking if the first character was X
CMC: Make the polyglot test driver a polyglot
Testing the new driver has me wondering how few characters I could use to break ALL the languages.

That could almost be a separate challenge itself.
@SnoringFrog Probably just one
The right Unicode character in the right spot could screw everything over
@MDXF It should be scored in bytes
how few characters I could use to break ALL the languages.
characters != bytes
7:19 PM
Thats my point, it makes more sense to score in bytes
But it's harder :P
that's a great idea for a challenge. And it would definitely not be just one.
ooh I didn't think to use unicode. Adding a . as the first character did quite a bit of damage though.
I broke 75% of the languages by adding ?"72":END before the code
And as for the challenge, it could be an challenge
Like, answer #10 would have to break all of the languages used in answer #10 of the polyglot development challenge
Adding a " at the start breaks all the non-esoteric non-rk languages
about 27 were left standing after a `.` at the beginning.

For the driver: I still think we should have Fail's misaligned to make them standout. Not a huge deal either way though.
7:23 PM
We'd need to fix the infinite stderr bug in the test driver before we could really test breaking everything.
If anyone can break rk without changing anything inside of the rk:start to rk:end I'll be very surprised
(or the space before rk:start and after rk:end)
@Chance Is that why the output keeps getting truncated?
not totally sure.
@MDXF easy enough to do when you add an extra rk:start and rk:end with their own bug inside :P
@SnoringFrog Oh whatever :P
I've found one way to cause a syntax error / SIGSEGV in rk-lang
the test driver is basically hardened to demonstrate a good polyglot, but not necessarily accurately display all the fails.
7:29 PM
Have we tried adding Ook?
Well looks like i wont be getting literate haskell in till at least tomorrow if it is possible. My computer keeps eating all my mobile data when I connect through mobile hotspot. I did get a rough test done with the current languages, only 3 or 4 broke. But I didn't get the haskell running correctly. I think it was ghc erroring with an exit code of 127, but it could have been the executable generated too. Will have to check tomorrow when I have proper internet.
I've looked at Ook a lot, but that was back when we were still trying to beat the orignal VIP challenge's winning score, and it was always too verbose for that
@SnoringFrog So if we weren't trying to beat the VIP score it might be ok?
Also I'm pretty sure the guy who wrote what has about 10 other languages
Personally I think we should save the brainfuck equivalents until we are really running short on other languages.
But we should use 6 different Python versions? :P
7:32 PM
@Potato44 only breaking 3/4 languages is VERY good. looking forward to adding haskell.
They are really easy to add (relative to other langs) so they can get us out of a tight spot
Ook still could potentially be too verbose per this challenge's rules (I don't even remember what the limits are now, check the question), but I'm not sure. Most trivial bf equivalents were the same way
But back then we were stuck with ~70-80 characters per submission, maybe
And I think the minimal code in BF to print what we needed was about 25 characters, so having to replace each BF command with 3 characters or more meant a lang wouldn't fit
@Chance it broke evil and alphuck, I can't remember what else it broke though
alphuck is really easy to fix, and evil shouldn't be that hard
@Potato44 Alphuck isn't too hard to un-break ime
and evil is pretty simple
7:38 PM
Which of the characters in ```
<[linefeed] > main=print 72 would be the ones that nroke them?
i think m breaks evil and p needs a preceeding s for alphuck
Oh, i remember s.i.l.o.s broke too
but the isn't a problem depending on the line you put it on.
silos could be trickier
Just before the commercial is where i was trying
evil and alphuck are usually pretty simple fixes
e will break both, f/m and maybe others will break evil, p will braek alphuck
7:42 PM
i think trying after the line #sseeeem might be better
I think it might be the = upsetting silos if i remember discussion from earlier correctly
silos hasn't broken before, so you'll be blazing some new territory on that one.
well, it did break once when the interpreter got updated, but wet changed to a simpler answer.
Excuse to learn a new language i guess, considering i don't know any of the esolangs yet
@Chance Yeah, main would be a pain to work with otherwise. m is the landing point for evil's jumps (usually, there's an alternate mode that could get around this) and the a will break things unless shortly followed by a u.
Actually main print is just a nasty combo for evil. p/r also break evil...
I could try using putstrln . Show instead of print
If that would be easier
Lower case s
7:47 PM
@Potato44 r is the read command, so dodging that is super helpful. p is only an issue if you do this before the main evil line
also, if you can capitalize things, then evil can't see them; then you tend to generate Fission issues though
Well, haskell is case sensitive
I think I have a good evil fix though.
I cant think of any haskell commands to print that dont contain an r off the top of my head
So the important commands for this are f x m j
Does echo possibly work? (never used Haskell except for quines, pardon me if I don't know what I'm talking about)
7:51 PM
@MDXF dont think so
m and j are markers, which are landing points for jumps; f initiates a forward jump, and x changes the marking mode (which determines if you land on m or j). There aren't really any lowercase j in the latter half of the program, so you could add an x near the end of the main evil line, then add a j near the end of the program (I'd probably put it with rk's filler s) and that should make the whole second half invisible to evil.
I've avoided using alternate marking mode (with j as landings) thusfar because j is alphuck's output command and, if I recall correctly, alphuck haphazardly executes a whole bunch of random code that by some miracle happens to not cause infinite loops
@SnoringFrog thanks, will try that when i get back to workinng on the polyglot.
If we're lucky, that spot near the end won't be executed by alphuck and it'll be an easy switch. Otherwise it'll require more detailed tracing of what alphuck does these days.
Alphuck is just a relabelled brainfuck, right?
Yep. And the test driver output actually includes a transpiled version if you want to track the brainfuck code itself

8:09 PM
Just thinking out loud because I have no idea whether it would be a good idea or not. What do you guys think of adding archway2, assuming we can find the interpreter? Looks to be a 2D brainfuck with loops replaced with conditional reflectors. The IP starts in the lower left.
Page on esolangs is just called archway
Wouldn't be our first 2D BF-variant (Minimal-2D). IP starting in the lower left would be interesting. It's worth nothing that the s 8 in the lower left has thusfar been very malleable (before the 8, it's Prelude-padding that's doubled as alphuck/Incident help from time to time), so we have a little leeway there already
@WheatWizard how about 73?
@all we could add Ook but it's Ook! Ook! Ook! Ook? Ook. Ook. Ook. Ook. Ook. Ook? Ook. Ook. Ook! Ook? Ook. Ook. Ook? Ook. Ook? Ook! Ook. Ook? Ook? Ook! Ook. Ook? Ook! Ook! Ook! Ook. Ook! Ook! Ook! Ook! Ook! Ook! Ook! Ook! Ook! Ook. (200 characters). How bad is that?
YABALL is in the "languages to be added" queue as well
8:25 PM
Could we add Morse code?
@MDXF -[++>+[+<]>]>-.----.
@WheatWizard Yeah I figured it out
200 characters is pretty bad
Yeah. But we could add morse
11 characters: --... ...--
I'll look for an interpreter that ignores non-Morse characters
If you can find an interpreter
8:33 PM
@MDXF The limit's technically 20% of current size, which gives us ~360 about now. So it would technically fit. And since it requires all the punctuation it actually doesn't even affect Moorhens, which I think it now the biggest challenge for adding real-word based BF variants (like There Was Once A Fish Named Fred)
@SnoringFrog So it might be an option?
Seems like it.

And good luck on that interpreter hunt...similar searches have led me into some weird places on the internet. Finding that dang alphuck interpreter...
Either Cow or alphuck kept taking me to German geocaching sites. One of those did have a working online interpreter, but it was buggy in really weird ways so it seemed like a maintainability nightmare since we didn't have access to the source.
Was it an online interpreter?
Yeah, but server-side
I wanted to suggest REing the executable if it was local, but this is worse
8:43 PM
That's why the Cow solution is 20-30 bytes longer than it could have been. Cow can print as an int, but every interpreter except that German one outputs a space after printing as an int.
Well, every interpreter that existed at the time. I've since forked a couple and remedied the problem lol.
That also reminds me I need to finish my trivial bf substitution interpreter, so ANY of the simple variants will always have an interpreter. That also means that I'll have an interpreter that, weirdly, predates some of the languages it interprets.

And since a language is defined by the interpreter...maybe that means you could create bf-variants on the fly for challenges until that loophole is closed? lol.
While I've been searching all over the non-English corners of the internet for a Morse interpreter, I forgot that I wrote one last year
@SnoringFrog Found it lol
@SnoringFrog That's the case with my rk-lang interpreter, and that's why I had to redo the whole thing
9:00 PM
@MDXF We've been bitten by weird interpreter issues a number of times in this challenge. That's why we're stuck with the one case insensitive alphuck interpreter that seems to exist; however, we've managed to use that to our advantage. On the downside, though, it's cells are way bigger than the usual 8 bytes, so we can't abuse overflow for shorter solutions
That's also why the current driver test for alphuck is considered an abstraction: it works with 8 bit cells and we just make sure we don't actually depend on that at all
9:14 PM
Well, does anyone want to add another lang so I can add Morse or Ook?
I won't have time to for a while, probably. Only managed to slip Commercial in because I'd worked on it for a bit before and because it broke less than I expected it to
Oh okay
Actually since @WheatWizard will add one tonight I'll wait until he does that
I probably don't have enough time to rn anyway
@stasoid could probably add his C++11 answer.
Oh btw nobody will take either of my lang ideas, right?
I have to leave for ~5 hours in 30 minutes so...
I've already figured out how to work both languages into the existing code so it'd be a waste of time for anyone else to do it
and I want to post them :P
9:35 PM
I think you're safe.
Okay lol
10:41 PM
Finally made a C/Fortran/asm polyglot
now I just have to close out the code
10:57 PM
@Chance Wait really?!?!
Oh nvm I thought you said polyquine
I've been working on a polyquine for ~3 weeks
11:22 PM
A: Sandbox for Proposed Challenges

MD XFTrivial Brainf**k Substitution Interpreter code-golf brainfuck interpreter Brainf**k is the most famous esoteric programming language and is the inspiration for hundreds of other esoteric languages. In fact, there are quite a few languages that are so heavily based off of Brainf**k that the only...

11:32 PM
Alright. I believe I've taken Fortran far enough to prove it'll be a valid addition.
@ais523 let me know if this will hose your Lua plans.
I had to make some moves on line 1 & 2
11:49 PM
@Chance if it obeys C comments, it won't
if not, it might
oh, I see
I don't think that will hose it, but I'm not 100% sure
at least, it looks less obnoxious than some of the things I have to work round, such as <><
(in case you're interested, <>< is the language that made me ask "can I change interpreter?"; I'm trying to work around an arguable bug in the current one)
00:00 - 18:0018:00 - 00:00

« first day (80 days earlier)      last day (497 days later) »