Anton (Dec 03 2025 at 10:17): Anton (Dec 03 2025 at 10:17):I feel like we should disable running if there are check errors for now or use a separate subcommand for "run anyway" so it stays actively tested in CI. The Roc crashed: ... messages can be confusing and distract from the true cause like here.
Fabian Schmalzried (Dec 03 2025 at 10:23):This might be generally a good idea. Or make it interactive: "I found errors, would you like to run the application anyway? (it will crash if it reaches one of those errors)". And have an run on error flag to set it explicitly?
Anton (Dec 03 2025 at 10:26):Flag sounds good. I would personally avoid the interactive approach, because most often I don't want to run on errors, and then I would have to confirm that every time.
Richard Feldman (Dec 03 2025 at 12:07):yeah we went through this cycle with the original compiler. I think the curren design is what we eventually want, because eventually errors will be visible in your editor, so by the time you've decided to run, you already knew about all the errors.
Richard Feldman (Dec 03 2025 at 12:07):I don't want to end up with like roc, roc run, roc dev again :sweat_smile:
Richard Feldman (Dec 03 2025 at 12:08):that said, I think it's ok to use a flag for now as a stopgap until the editor tooling is there
Anton (Dec 03 2025 at 16:24):I'll add the flag right now
Anton (Dec 03 2025 at 17:57):https://github.com/roc-lang/roc/pull/8557
Last updated: Dec 21 2025 at 12:15 UTC