Is of really needed?

[senhalil]

First of all, amazing idea! Always wanted to write regexp like a normal language, thanks for this awesome project.

It looks like nobody asked so I figure I might as well be that person. I feel like 2 <matches> is more expressive than 2 of match. (Not sure what would entail and/or if it would be possible but here are my 2 cents.

[yoav-lavi]

Thank you for the kind words!

Regarding of, waiting to see what other people think, but I personally see it as a sort of link between the number and what it's operating on.

Syntactically it makes more sense to me to have 2 * 5 or 5.times { ... } rather than 2(5) or 5 { ... }.

Will consider though!

[senhalil]

Thanks for the answer. I do understand your point of view.

I guess, I treated the syntax as ordinary English (which might not be expressive enough to handle all intricacies of regexp after all 🤔) and the class code example (i.e., // classname 1xx) read more nicely in my head:

some of <word>;
<space>;
"1";
2 of <digit>;

some <words>;
<space>;
"1";
2 <digits>;

Cheers!

[yoav-lavi]

I see, although that would require a plural form for each type of character (and that may not work for literals, e.g. 2 "A")

[senhalil]

Yeah, it would require plurals for the keywords -- and let me tell you, I had the "chance" to look at that problem recently and it is not really evident to do it programmatically for any word (fortunately there were libraries); however, I was thinking for the reserved keywords, since there are not many of them, it might even be okay to "hard-code" pluralize them.

For literals, I was leaning on 2 "A"s, some "A"s kind of logic -- i.e., if the multiplier/coefficient is greater than 1, then s itself is treated part of the syntax " "s.