Introduce MIX_OF_LOWER_AND_UPPER_CHAR_CASE_IN_RANGE warning

[KvanTTT]

Consider the following set: [A-z]. Most likely it's an incorrect definition because the range

ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxy

contains nonimplied characters:

[\]^_`

Correct and clear definition: [A-Za-z].

If these characters are implied, they can be added explicitly:

[A-Za-z[\\\]^_`]

I suggest adding at least a new warning MIX_OF_LOWER_AND_UPPER_CHAR_CASE_IN_RANGE. It's especially actual if use the caseInsensitivity option. Or name it as RANGE_PROBABLY_CONTAINS_NOT_IMPLIED_CHARACTERS.

@parrt what do you think?

[parrt]

Hmm...yes, this does seem like it could cause strange errors. On the other hand, has this ever come up? Usually I try to avoid work unless it's solving a known problem haha.

[KvanTTT]

I've encountered the error during fixing caseInsensitive option. Moreover, I think it is quite a real case, it's quite intuitive but wrong definition. That's why it should be fixed. BTW it's incorrect for Latin characters but is correct for Cyrillic characters. [А-я] does not contain intermediate characters.

[parrt]

Understood. Well I would welcome a PR :) I'd definitely love to get the case insensitive option merged so I can try to review some of your other work...I start work for my new galactic overlord on 4 January and it would be nice to get a stable set up...

[KvanTTT]

It would be good to take a look at this PR while I'm fixing the rest case insensitive issues: #3349 It resolves a lot of issues including character ones (but it should be merged after case insensitive PR).

[KvanTTT]

I started testing our grammars-v4 and encountered some subtle cases with the current warning that related to Unicode ranges. Probably I was too fast about the correct solution here, but I'll resolve it (restrict to ANSI characters) or revert it tomorrow. Sorry.

Yet another point for including grammars-v4 to the test infrastructure in some way (integration testing).