Do a bit of explanation on implementation choice we made WRT range expressions

[yuripv_icloud.com]

Summary

• add comment in regcomp.c
• add the same in regex(3)

[kevans]

lib/libc/regex/regex.3

615

This belongs up above in the "IMPLEMENTATION CHOICES" section (see line 532); the explicit reference to IEEE Std 1003.2 can be dropped there, because that section is dedicated to how we handle behavior explicitly defined as undefined.

[yuripv_icloud.com]

lib/libc/regex/regex.3

615

It *is* in implementation choices section, and I did use the reference to distinguish between what specifications says and what we have chosen to implement -- see a "backslashed character" paragraph a bit above making the same reference.

[kevans]

lib/libc/regex/regex.3

615

Whoops, misread a cross-reference up above. My bad. =)

I don't care much for the redundancy up there either, but it makes sense: it wants to say "given a magic meaning by [The Standard]"; this is implied in most of the other implementation choices as well, but we don't tend to repeat that- this would've been written to basically chop off everything before "expression ranges" following the structure of these.

I don't feel strongly enough about it to request it to be changed, though.

[yuripv_icloud.com]

lib/libc/regex/regex.3

615

Then I'll need your help here as I can't properly rephrase without the standard reference :-) (sorry not a native speaker).

[yuripv_icloud.com]

update wording in man page a bit

lib/libc/regex/regex.3

615

Hopefully it looks better now?

[kevans]

lib/libc/regex/regex.3

615

Looks good to me

[kevans]

LGTM; tag new account @yuripv