yes-repro

BTW, staging.ato.pxeger.com is working, for anybody else like me who just needs to try some code out

Welp, now staging has the same issue

uh-oh, this isn't good

this happened yesterday, but I just restarted it and it fixed itself

but if it keeps happening then something is clearly broken

@PyGamer0 Let me rephrase that: On average it has shorter programs than Vyxal.

@DLosc Now I really don't understand... even though a restart fixed it on ato., it isn't working on staging.ato.

[root@atostaging ~]# reboot
Connection to staging.ato.pxeger.com closed by remote host.
Connection to staging.ato.pxeger.com closed.
~[attempt-this-online] $ ssh root@staging.ato.pxeger.com                                              255
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@    WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!     @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY!
Someone could be eavesdropping on you right now (man-in-the-middle attack)!

highly suspicious

ok this was unrelated

I really don't think Unrelated String would do something like this :P

@pxeger Hey, yeah I wasn't able to get it to work with GHC 9, but any GHC 8 should work

@pxeger I'm not able to see the error message because it says unknown language then brings select language menu over the button for error/etc

@DarrenSmith yeah, sorry, because of the issues I've been having the past few days I had reverted to an older version

The error message was "cannot allocate memory" which seemed unlikely to be right

Hmm, I haven't seen that error. Sounds like you got it to compile though, curious why it didn't see the packages in the directory by default, what did you have to do to fix that?

I didn't fix it; that error message appeared anew when I started running it on ATO instead of on my local computer

I think the reason it doesn't see the packages is because I deleted them though ;)

If you run it with a flag -c (which should just save a file but not generate and run a haskell program) does it work? There might be something weird about it compiling haskell source (which is how it executes nibbles programs). Similar to Husk in that regard, but Husk hasn't been done too it looks like

Ah ok, there is a fix in the memoize one which isn't in the official repo yet (it get's really slow for large integers without it)

doesn't matter much though, only affects recursive fns which are rare

oh, hmm

Ok, I think I need to run nibbles from in the nibbles working directory for it to detect the modules?

it should be able to be ran from anywhere, but I think it would need to be compiled from there

(the nibbles binary compiled from there that is)

otherwise I still get "Could not find module Data.Function.Memoize" errors etc

nibbles -hs <<< '|_~`^+`@ 2' && ghc -v out.hs gives these logs:

gist.github.com/pxeger/5340640147fe10618af1f8ae83e9c021

ah actually you are right. I didn't realize that issue

so yeah it would need to be ran from there too... I really regret compiling to HS instead of just interpreting... premature optimization :(

so running it from the nibbles directory works fine in the medium term, but I don't know why it doesn't work on ATO...

on ATO does the -c or -hs option work? I suspect it is nibbles itself trying to call GHC and start processes that causes the issue

(those options will not call any other process and just generate .nbb or .hs files)

looks like the out of memory error was probably a genuine one: the maximum memory usage of nibbles <<< '|_~`^+`@ 2' 4 is 196MiB

(ATO limits you to 128MiB)

I should probably try a simpler example program

I suspect it calling GHC uses a ton :(

yup :/

ato.pxeger.com/…

I guess I'll raise the memory limit...

nibbles: out: callProcess: posix_spawnp: does not exist (No such file or directory)

so I think in nibbles.hs you'll need to change "out" to "./out" I already did that in latest version but not in stables release yet

ah ok, that should be simple enough to fix

I wish this was easier... thanks for taking the time to work through it

yes, hopefully that'll be the ironing out of most of the issues

line 158 I think

Would it be better to install Nibbles from main in the GitHub repository rather than a release?

Otherwise I can just do a patch

The easiest would probably be golfscript.com/nibbles/nibbles-latest.tgz (it has the latest changes). That should be easier than github since it includes the libraries it depends on (those aren't in github)

ah ok

normally that link could contain some unvetted changes, but I haven't done anything wild to it since 0.25