"failing silently" (returning 0) when using 0 as divisor would be non-standard. panic as before is fine

Thanks, you bring up two issues:

• Behavior when n == 0 is not specified or tested
• Modulo values with greater than 7 bytes may silently overflow when doing shl (<<)

Issues with the PR:

• We should not return 0 for n == 0. 0 should instead be returned for n == 1. For n == 0 we should panic instead
• Your tests do not actually show the incorrect behavior for large values of n, nor do the code changes fix the issue

Closing in favor of #1123 which will fix the above issues

(Thanks for calling attention to this @reject-i 🙏 )

Thanks for your work!