Comment author: @lpw25

Just a note to point out that zoep has attached a patch for this request

Comment author: @gasche

The code looks good, but I'm a bit wary of having to review sys_nat_of_string to make sure it didn't somehow make base-10 assumptions. I assume this would be catched rather quickly by testing; would it be possible to add relevant tests in testsuite/tests/lib-num to exercise the new code paths?

Comment author: @gasche

After having a closer look, sys_nat_of_string is in fact fine but should be updated, along with base_digit_of_char (still in nat.ml), to accept the lexical conventions of OCaml integer literals ( http://caml.inria.fr/pub/docs/manual-ocaml/lex.html#integer-literal ):

• both lowercase and uppercase letters should be accepted in hexa (currently only uppercase)
• underscores in the number should be ignored (at any base)

Comment author: zoep

I made a new patch based on gasche's notes and the lexical conventions described in the link.

• Underscore characters are ignored anywhere in the number, except before the first digit
• Lowercase letters are accepted in hexa
• Both uppercase and lowercase characters are accepted in base prefixes, instead of just lowercase in the previous patch
• Some simple tests were added

Comment author: @gasche

This new patch looks fine, thank you!

Comment author: @gasche

The patch is now committed in trunk, thanks! If you wish to appear in the changelog under a different name than "zoep", this can still be changed.

(To keep in mind for future patch submitters: at some point we pick a name to give credit to.)

Comment author: zoep

Happy to see that! Also I would prefer if you could use my real name, Zoe Paraskevopoulou, instead of zoep. Thanks!