jmanuel1 / concat Goto Github PK

A poorly named Python-based concatenative programming language.

License: Other

Python 97.42% TypeScript 2.58%

concatenative concatenative-language concatenative-programming-language python stack-based

concat's Introduction

Hello there 👋

I'm a software developer at Imagine Learning and a recent graduate of Arizona State University. I studied computer science and was an officer of the Software Developers Association. I like using code to create, and I believe in technology's ability to help us do good things. Outside of coding, I like games such as Thomas Was Alone and Cities: Skylines, occasionally taking photos, math, and outer space.

🔭 I’m currently working on a transpiled concatenative language called Concat. Check out my Mailspring plugin, Todoer, too.
🌱 I’m currently learning programming language design & implementation (at least, what I need to know for Concat).
👯 If you want to contribute to any of my projects, I encourage you to look at Concat and Todoer.
🤔 I’m looking for help with Concat's type checker and exploring computational sustainability.
💬 Ask me about web development, programming languages, computational sustainability, Doctor Who lore, and my favorite Snarky Puppy tracks.
📫 How to reach me:
- LinkedIn
- Website
⚡ Fun fact: Once upon a time, I was on an ultimate frisbee team.

concat's People

Contributors

Stargazers

Watchers

concat's Issues

Have a way to get type at point

Type of a name
Stack at program point

Better static error handling for editors

Tokenization should continue even when a token is unclosed
Multiple parsing & type checking errors
Type check even when there's invalid syntax

Document standard library

Atom package: syntax highlighting is off-by-one for © character

Expected behavior: 2022 and 2023 are highlighted as numbers.

Observed behavior: 022- and 023o are highlighted as numbers.

Convert comments describing AST visitors and parsers to doc-strings

Atom package: Swallow `tokenize.TokenError` for syntax highlighting purposes

Test case: [

Expected: Syntax highlighting completes (although no style is attached to [)

Actual: This error in the console:

child_process.js:304 Uncaught (in promise) Error: Command failed: C:\Users\jamai\Documents\projects\concat\env\Scripts\python.exe -m concat --tokenize
Traceback (most recent call last):
  File "c:\users\jamai\appdata\local\programs\python\python37-32\Lib\runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "c:\users\jamai\appdata\local\programs\python\python37-32\Lib\runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "C:\Users\jamai\Documents\projects\concat\env\lib\site-packages\concat-0.1.2.dev1-py3.7.egg\concat\__main__.py", line 67, in <module>
    tokens = concat.lex.tokenize(code, should_preserve_comments=True)
  File "C:\Users\jamai\Documents\projects\concat\env\lib\site-packages\concat-0.1.2.dev1-py3.7.egg\concat\lex.py", line 39, in tokenize
    token = lexer.token()
  File "C:\Users\jamai\Documents\projects\concat\env\lib\site-packages\concat-0.1.2.dev1-py3.7.egg\concat\lex.py", line 68, in token
    return next(self._concat_token_iterator, None)
  File "C:\Users\jamai\Documents\projects\concat\env\lib\site-packages\concat-0.1.2.dev1-py3.7.egg\concat\lex.py", line 79, in _tokens
    for token_ in self.tokens:
  File "C:\Users\jamai\Documents\projects\concat\env\lib\tokenize.py", line 579, in _tokenize
    raise TokenError("EOF in multi-line statement", (lnum, 0))
tokenize.TokenError: ('EOF in multi-line statement', (2, 0))

    at ChildProcess.exithandler (child_process.js:304)
    at ChildProcess.emit (events.js:223)
    at maybeClose (internal/child_process.js:1021)
    at Process.ChildProcess._handle.onexit (internal/child_process.js:283)
exithandler	@	child_process.js:304
emit	@	events.js:223
maybeClose	@	internal/child_process.js:1021
ChildProcess._handle.onexit	@	internal/child_process.js:283
Promise.then (async)		
step	@	C:\Users\jamai\Docum…antic-grammar.ts:22
(anonymous)	@	C:\Users\jamai\Docum…antic-grammar.ts:23
__awaiter	@	C:\Users\jamai\Docum…antic-grammar.ts:19
module.exports.JavaScriptSemanticGrammar.markTokensForChange	@	C:\Users\jamai\Docum…ntic-grammar.ts:234
module.exports.JavaScriptSemanticGrammar.markTokens	@	C:\Users\jamai\Docum…ntic-grammar.ts:227
(anonymous)	@	C:\Users\jamai\Docum…ntic-grammar.ts:209
simpleDispatch	@	<embedded>:11
emit	@	<embedded>:11
emitDidStopChangingEvent	@	<embedded>:11
later	@	<embedded>:14
setTimeout (async)		
(anonymous)	@	<embedded>:14
emitDidChangeTextEvent	@	<embedded>:11
undo	@	<embedded>:11
(anonymous)	@	<embedded>:11
mergeSelections	@	<embedded>:11
avoidMergingSelections	@	<embedded>:11
undo	@	<embedded>:11
textEditor.undo	@	C:\Users\jamai\.atom…gle-selection.js:81
core:undo	@	<embedded>:11
n.<computed>	@	<embedded>:11
handleCommandEvent	@	<embedded>:11
KeymapManager.dispatchCommandEvent	@	<embedded>:11
KeymapManager.handleKeyboardEvent	@	<embedded>:11
handleDocumentKeyEvent	@	<embedded>:11

In this case, the issue is the unclosed bracket.

Atom package: Black lines in minimap

I think the minimap package does not understand that we're doing syntax highlighting. In fact, the Grammar subclass doesn't return any tokens in tokenizeLine which I think explains why the lines in the minimap are black. Atom's grammar tokens are supposed to collectively cover the whole text, but our grammar tokens cover none of the text because there are no tokens.

Possible solutions:

Have a textmate grammar. Use the Concat lexer to overlay extra information on top of it.
Give Atom the currently known tokens when it asks for them from the grammar. We get the tokens asynchronously, so maybe what we display will be out of date.
Plug into the minimap package.

I'm not of fan of the second option.

Document grammar and semantics

Write an LSP client to remove dependency on atom-languageclient

Use `trampoline` package for stack-safe continuations when possible

IndentationError: unindent does not match any outer indentation level

Reproduce with:

def remove_stack_polymorphism(f:forall `t *s. (*s i:`t -- *s) -- g:forall `t. (i:`t -- )):
  ()
 dfbfdbff

Atom package: Make command to run Concat configurable

Currently, the Atom package searches for Python in a few places and then runs python -m concat. However, it cannot possibly foresee every situation in every project. Therefore, there should be a project-specific or even directory-specific way to specify which command should be used to run Concat. The current behavior can be kept as a default.