Luke Boswell (Jan 23 2025 at 23:25): Luke Boswell (Jan 23 2025 at 23:25):Instead of the pipe here...
"Default color" |> ANSI.color({ fg: Default, bg: Default })
will we be able to dispatch on a Str?
"Default color".color({ fg: Default, bg: Default })
Thinking about this more... it's not a sensible question. Str isn't going to be a Custom Type.
Luke Boswell (Jan 23 2025 at 23:29):I'm just looking at my ANSI examples... and debating if I refactor these to remove the |>
ANSI.color("Default color", { fg: Default, bg: Default })
Thinking about this more... I would probably refactor the API to have a ANSI.styled : Str -> ANSI or similar and then have methods hanging off that to style the string segment. Sounds like a job for future Luke. :smiley:
Sam Mohr (Jan 23 2025 at 23:42):I think Str will be a custom type
Sam Mohr (Jan 23 2025 at 23:42):And that you can call methods on a string
Sam Mohr (Jan 23 2025 at 23:42):In the same way List would work
Sam Mohr (Jan 23 2025 at 23:43):We need methods like List.map to work
Luke Boswell (Jan 23 2025 at 23:43):Wait so "str".color(...) might work? I won't be able to write my own methods because Str is a builtin.
Sam Mohr (Jan 23 2025 at 23:45):That would be my expectation yes
Richard Feldman (Jan 23 2025 at 23:45):no, that would not work
Sam Mohr (Jan 23 2025 at 23:45):In general, methods only get defined in the module of the custom type
Richard Feldman (Jan 23 2025 at 23:45):right
Richard Feldman (Jan 23 2025 at 23:45):because literally what the . is doing there is saying "go look up the type of the thing before the . and call a function by this name in its module"
Sam Mohr (Jan 23 2025 at 23:45):If we didn't require that, then we'd have to do some expensive Rust style trait resolution
Richard Feldman (Jan 23 2025 at 23:46):so "str".color(...) does exactly the same thing as Str.color("str", ...)
Brendan Hansknecht (Jan 24 2025 at 02:34):That said, assuming we support .(fn) you would be able to do:
from ANSI import color
"Default color".(color)({ fg: Default, bg: Default })
This is part of the reason I really want to support .(fn)(...)
Anthony Bullard (Jan 24 2025 at 03:04):Just without the python import syntax @Brendan Hansknecht :wink:
Sam Mohr (Jan 24 2025 at 03:05):We need a Zulip LSP!!!
Anthony Bullard (Jan 24 2025 at 03:05):First we need Zulip to support Roc syntax highlighting
Anthony Bullard (Jan 24 2025 at 03:05):I wish we had a way to patch that in
Anthony Bullard (Jan 24 2025 at 03:05):Even if it was just TextMate grammar
Sam Mohr (Jan 24 2025 at 03:06):Someone was working on it a while back
Luke Boswell (Jan 24 2025 at 03:10):@Isaac Van Doren I think
Anthony Bullard (Jan 24 2025 at 03:14):I know that @Eli Dowling owns the Treesitter grammar, right?
Sam Mohr (Jan 24 2025 at 03:14):Yep, but it's not Treesitter
Anthony Bullard (Jan 24 2025 at 03:14):Right, I know
Anthony Bullard (Jan 24 2025 at 03:14):And the TS grammar is woefully out of date now I think
Sam Mohr (Jan 24 2025 at 03:14):https://docs.zulip.com/help/code-blocks#language-tagging
Luke Boswell (Jan 24 2025 at 03:15):I think Eli is travelling for a month or so, he said he'd pickup the grammar one syntax madness settles a little
Sam Mohr (Jan 24 2025 at 03:15):
Isaac Van Doren (Jan 24 2025 at 03:17):I did highlight-js support for roc but have not done textmate
Brendan Hansknecht (Jan 24 2025 at 03:22):Anthony Bullard said:
Just without the python import syntax Brendan Hansknecht :wink:
I have been writing too much python at work
Sam Mohr (Jan 24 2025 at 03:24):You mean any at all?/s
Sam Mohr (Jan 24 2025 at 03:26):Small Python Good, Big Python Bad
Brendan Hansknecht (Jan 24 2025 at 03:28):Indeed.
Last updated: Jun 16 2026 at 16:19 UTC