Emil J (Dec 02 2024 at 08:59): Emil J (Dec 02 2024 at 08:59):Hi! I'm doing AoC in Roc (and others) and handling Results has taken up an incredible portion of the time solving 1a. I know the inputs are valid, so I'd like to crash on invalid inputs. However, when I do a match on an Ok Result, the crash in the Err would still evaluate. How should I use crash properly?
Emil J (Dec 02 2024 at 08:59):I ended up pulling off an absolute anti-pattern just to get it done
purifyStr = \x -> Result.withDefault x "AAAAA"
purifyInt = \x -> Result.withDefault x 999999
:confounded:
Emil J (Dec 02 2024 at 09:01):oh and this gives me
thread '<unnamed>' panicked at /build/source/crates/compiler/types/src/unification_table.rs:295:62:
index out of bounds: the len is 3561 but the index is 3561
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
want me to report it on the issue tracker?
Sam Mohr (Dec 02 2024 at 09:02):Normally, I'd write something like this, if I understand you correctly:
parseInput = \input ->
parseResult = doStuffWith result
when parseResult is
Ok val -> val
Err err -> crash "Failed to parse input with err $(Inspect.toStr err)"
This is also what I did last year. But with the new try syntax, it is so easy to handle the error case, that I don't feel the need to use crash any more. You can just write:
input
|> parseInput # This returns a Result
|> try # This Removes the Result
|> solveAoC
|> Ok
The only requirement is, that you are calling this code in a function that returns a Result.
Sam Mohr (Dec 02 2024 at 09:09):I'm working on improving try at the moment. Please feel free to try it out (harr harr), and let me know if you run into any issues!
Emil J (Dec 02 2024 at 09:09):oh, so that try simplifies things like result of results? That's good enough yep
Sam Mohr (Dec 02 2024 at 09:12):And also @Emil J I think that crash is because of your lefts_sorted
Sam Mohr (Dec 02 2024 at 09:12):When I changed it to camelCase, it worked
Sam Mohr (Dec 02 2024 at 09:12):We currently only do camelCase in Roc, but I'm surprised that you got such a deep error...
Sam Mohr (Dec 02 2024 at 09:13):I don't think it's worth reporting an issue on because we're planning on moving to snake_case soon
Sam Mohr (Dec 02 2024 at 09:14):@Emil J here's the recent announcement on usage of try:
#announcements>New `return` and `try` keywords
Sam Mohr (Dec 02 2024 at 09:14):Heads up, you can't currently use it in the same function as !
Sam Mohr (Dec 02 2024 at 09:14):But that'll get fixed in the next month or so, I think
Emil J (Dec 02 2024 at 09:23):ok thanks!
Last updated: Jul 26 2025 at 12:14 UTC