Stream: bugs

Topic: Num.atan gives inconsistent values


view this post on Zulip Guillaume Sarisson (Feb 17 2026 at 20:18):

Hi there! It looks like the Num.atan function always returns the same two values in the REPL. Has anyone encountered this behavior before ? It looks something like this :

>> Num.atan(1/1)
0 : Frac *
>> Num.atan(1)
2550673372347179.51480058619310904 : Frac *

view this post on Zulip Luke Boswell (Feb 17 2026 at 20:26):

Can you clarify which Roc version you have? -- I assume this is from one of the nightlies

view this post on Zulip Luke Boswell (Feb 17 2026 at 20:29):

I haven't seen this before... though tbh I haven't looked at the Rust version of the compiler in about a year now :sweat_smile:

view this post on Zulip Guillaume Sarisson (Feb 17 2026 at 20:33):

roc --version gives me roc nightly pre-release, built from commit d73ea109 on Tue 09 Sep 2025 09:02:08 AM UTC

view this post on Zulip Luke Boswell (Feb 17 2026 at 20:47):

Is this blocking you in particular.. or more of an observation?

view this post on Zulip Luke Boswell (Feb 17 2026 at 20:48):

We haven't been spending much time supporting the old compiler, unless it's blocking someone who is actively using Roc

view this post on Zulip Guillaume Sarisson (Feb 17 2026 at 20:55):

It’s not really blocking me but I’m working on a simple scientific library in pure roc and the atan function appears quiet often in some area. I can just work on some other part of the library or update my compiler if needed.

view this post on Zulip Luke Boswell (Feb 17 2026 at 22:34):

Should be fixed in https://github.com/roc-lang/roc/pull/9186


Last updated: Feb 20 2026 at 12:27 UTC