mego / seriously

A golfing language that is probably terrible

Actually

A golfing language that is probably terrible. It is currently being developed. Actually is the spiritual successor to Seriously. Actually is stack-based, and is unique in its amount of single-character commands. Every character whose ordinal is in [0,255] does something (or will, once I finish development).

All strings are valid programs. There is no such thing as a syntax error, and anything that appears to be a runtime error is actually a no-op (NOP). If you call a command that expects a string on top of the stack, but the top value is an integer, instead of throwing an error, it silently performs a NOP. Eventually, this will not be seen, as the goal for Seriously is for every character and stack state to map to a command.

Seriously's name was inspired by this challenge. Actually followed because it is also an adverb, and it also looks funny in answer headers.

Try the online interpreter. This is part of Try it online! by @DennisMitchell.

Commands

See commands.txt

Examples

Hello, World!

H

If you prefer a more interesting version:

"Hello, World!"

Primality Test

p

99 Bottles of Beer

N

Coprimality Test

g1=

Counting the divisors of a number

w⌠iXu⌡Mπ

or

÷l

seriously's Issues

Vectorize functions

When a function whose first operand is not a list receives a list as its first operand, map the function over the entire list

cmp function

The cmp function is a Python 2 function, it allows easier comparison between numbers.

pop a,b: if a>b push 1; if a == b push 0; if a < b push -1.

Round by digits

Proposed:
pop a,b: push a rounded to b digits after decimal

Also consider:
pop a,b: push a rounded to b significant figures
pop a: push a in scientific notation (ex: 9000 => 9e3)
pop a,b: push a in scientific notation rounded to b significant figures

Also, I am currently experiencing roundoff errors.

Prefixes / Suffixes

The prefix/suffix function is a frequent function in 05AB1E. Stolen from 05AB1E

Get explanations from commands.txt

Flask has some great templating features so you don't need to put what each command does in both files.

Implicit input by default

Default behavior: read all bytes from STDIN, split on newlines, eval, and push to stack at start of program

With flag: operates as it does now (needing , to read from STDIN)

Excess input fails.

I dunno why, but, whenever I have something in STDIN that I don't use, the program errors out.

Try this one, for example, to get what I'm saying.

↑↓ - sort entire stack

• ↑ (18) sorts entire stack ascending, like APL
• ↓ (19) sorts entire stack descending, like APL

→ and ← might do something like "pop a; pob b; sort stack in range [a,b] ascending/descing."

If this is the wrong place to suggest these things... um, sorry.
[enhancement]

Actually eval

Actually seriously doesn't have an Actually eval? eval is also a very useful function.

Testing out Rultor

@rultor hello

Text encoding is a bad idea

No more text encoding. Everything needs to be hex.

w can't factorize negative numbers

current behavior: push []
expected behavior: push the result of [:-1:,1]@Awo instead.

Python 3 Conversion

Python 3 version needs to happen

f (format) hangs

╦#"{:.15f}"f causes the application to hang

Interacting with the system

Proposed:
pop "a","b": push "a" and write "b" to a file named "a"
pop "a" or [a]: push subprocess.call(a)

These would allow the original challenge to be done in Actually.

Also consider:
pop "a" or [a]: push subprocess.check_output(a)
41 (A): if stack is empty, push "Actually"
53 (S): if stack is empty, push "Seriously"

commands.txt link broken

In the master branch, I think the commands.txt link is broken and should be changed from:

https://github.com/Mego/Seriously/blob/master/commands.txt

to:

https://github.com/Mego/Seriously/blob/master/docs/commands.txt

⌐ and ¬ don't work on lists

Need to make a universal plus function with all of the overloading logic that other functions can use instead of +/__add__ - same with other operators

Maybe subclasses of built-in types?

Comments?

Please? That would be great. Thanks!

Online interpreter doesn't work

Throws 500s when trying to run code; related to recent source code attempted fixes

References to O persist

Some files (probably just those based on files from the O programming language) still make references to O. For example, from ide.py:

if __name__ == '__main__':
    print('Compiling O...')
    check_call(['gcc', 'o.c', '-DIDE', '-o', 'o-ide', '-lm'])
    print('Starting server...')
    app.run(host='0.0.0.0',port=80)

↨ - set SORTFLAG; ▬ - get SORTFLAG

↨ (17)

• pop a
• Set a variable I'll call SORTFLAG = a

SORTFLAG is used with [↑↓→←] to determine how sorting is done.
For example:

0. sort strictly by value. (default)
1. sort by value, strings first
2. sort by value, strings last
   etc...

-f --sortflag: set starting SORTFLAG.

▬ (16)

• push SORTFLAG

Strange bug with primality testing and repeating inputs

http://actually.tryitonline.net/#code=eVjDn-KImuKJiFJgcGDilpE&input=MTAw versus http://actually.tryitonline.net/#code=4oia4omIUmBwYOKWkQ&input=MTAw

Powerset

Powerset is a relatively useful built-in, e.g. for finding all pairs from a list.

Seriously Penguin

Very serious penguin 10/10 icon

Make function literal delimiters not the same character for start and end

Because ` is used for both the start and the end of a function literal, nesting function literals is incredibly difficult. A pair of characters need to be chosen for the start and end delimiters.

Smarter overloading with iterables

• If at first we don't succeed, try vectorizing
• Other stuff?

Comparing the results of O and r

= doesn't play well with O's map objects and x's range objects. One needs to # everything to get comparisons to work.

Edit:
What doesn't work with O:
= and x (range)
Σ (sum)
σ (cumulative sum)
▼ ([x//gcd([a]) for x in a]; only returns one value)
H (head)
t (tail)

What does work:
π (product)
∩ (intersection
∙ (Cartesian product)

Duplicate doesn't work on a zip

After zipping two lists with Z, ; doesn't work. I need to use Z♂#; because just Z#; doesn't convert the inside lists to lists.

Complex values are not yet supported

Get to work, you lazy bum!

Lucas numbers

I guess this is more of a reminder than an actual issue, but please implement Lucas numbers in Seriously just as you promised to do.

index function is not working

Based in the commands.txt, ¡ is supposed to be the character to do the index function (like b.index(a)) but when I tried it, it didn't work. For this code:

[1,2,3]2¡
I got an output of:
2 [1,2,3]

Clearly, this is not expected. In the description, the ¡ character is said to "pop a,b: push a string representing a in base b" So even the description is messed up. Clearly this is a bug.

99 Bottles Text does not match the most recent challenge's.

Dude. Fix yo' language so we can use it for evil eval! Muahahaha!

Numeric list commands that don't work on a list of complex numbers

See this code here: http://actually.tryitonline.net/#code=zqM&input=Wy0xNS4yMSswLjhqLCAxMC4xLTAuM2osIC0wLjA3KzIzLjU1al0

Printing without trailing newline?

I want a feature so that I can print without a trailing newline, like sys.stdout.write. Right now, using . doesn't suffice. Is there any such function alerady?

Implement Miller-Rabin primality testing for prime-related functions

Ideally, this will be a standalone implementation - no reliance on numpy or sympy. I'd like to avoid adding those as requirements, as they're quite large libraries.

The implementation must be deterministic.

Bonus points for using the current-best set of witnesses to ensure accuracy for very large numbers.

Use generators more

Really, there's no real reason to explicitly cast to list every time an iterable is pushed on the stack. Keeping iterables in generator form on the stack will reduce memory usage and increase performance. Lists should only be used when needed (like when printing).

NUL - HALT_AND_CATCH_FIRE

Cause the interpreter to crash, just for fun. I'm sure we'll find a use for this.

split for lists not supported

Can you possibly add a split() function to Seriously? it would help with various different challenges on code golf.

List literals break implicit input

The check for explicit input needs to be more robust - [1,2] causes input to not be taken implicitly.

Function values are not yet supported

Get to work, you lazy bum!

Dot product doesn't work on lists of floats

Example:

Stack: [-0.3, -10.1], [-15.21, 0.8]

Tries * for dot product. Nothing happens.

Insufficient documentation

Sure, I can look in the online interpreter, but 'aint nobody got time for that.

Bug with ╡

╡ command (sublist; chunk by number) doesn't appear to work on TIO.

Registers not being saved with O

O╗⌠╜k⌡M

https://tio.run/##S0wuKU3Myan8/9//0dTpj3oWPJo6J/tRz0Lf//@jlRKTklNSlXQUlDJSc3LyQYyS1OISsEB@XrZSLJeSpaGREgA

combinations and permutations

need commands for nCr and nPr

Add a for loop

I propose to have a for loop enhancement.

Factoring fails for large inputs

Code: :1965593254291461501637330902918203684832716283082y (w produces similar results)

Error:

Traceback (most recent call last):
  File "/opt/actually/seriously/seriously.py", line 231, in eval
    self.fn_table.get(ord_cp437(c), lambda x: x)(self)
  File "/opt/actually/seriously/SeriouslyCommands.py", line 1349, in <lambda>
    0x79:lambda x:x.push(factor(x.pop())),
  File "/opt/actually/seriously/SeriouslyCommands.py", line 579, in factor
    return [a for a,b in full_factor(n)]
  File "/opt/actually/seriously/SeriouslyCommands.py", line 560, in full_factor
    init_primes_up_to(int(rmath.sqrt(n))+1)
  File "/opt/actually/seriously/SeriouslyCommands.py", line 259, in init_primes_up_to
    temp=[1]*(n-max_tested)
OverflowError: cannot fit 'int' into an index-sized integer

The factoring algorithm needs to be more memory-efficient. Perhaps there should be a threshold value set for switching over to a slower-but-memory-efficient algorithm instead of the current somewhat-fast-but-memory-intensive approach.

@kckennylau Any thoughts?