Warnings about letterlike characters in identifiers about codeinspector HOT 5 CLOSED

Yes, this is working as intended.

For example, a few weeks ago this caught a real bug in the Mathematica startup code when this code was mistakenly added:

аnimationRepetitions = foo[]

Can you see the problem? It is subtle. (I deliberately did not turn on mathematica coloring)

I do agree that there might be value in allowing some ranges of characters but I'm afraid it would be a slippery slope and I'm not sure what that design would look like.

In the meanwhile, feel free to programmatically filter the output of CodeInspect.

from codeinspector.

Lints can always be turned off. I find it useful when it reports more rather than less, and I simply turn off whatever is not relevant to my own use case. (In fact, for every package on which I use this tool, I have a list of lints to turn off.)

Just don't fall into the trap of trying to change your code just to shut up the linter. It's a tool meant to help you find issues, not an overlord to satisfy 😄

from codeinspector.

I would turn the whole UnexpectedLetterlikeCharacter off, but seems that it shares the same name when the characters fall into Utils::isMBStrange. The only difference is that the confidence is higher for the latter case. I wonder if it is true that for isMBStrange characters, a Unexpected character warning will also be thrown? So that I don't need to worry that supressing UnexpectedLetterlikeCharacter will ignore the non-characters.

from codeinspector.

The idea is that UnexpectedLetterlikeCharacter lints are at the "token" level and UnexpectedCharacter lints are at the "character" level.

Right now, any character that gives UnexpectedLetterlikeCharacter will also give UnexpectedCharacter.

isMBStrange is for the "character" level and isMBStrangeLetterlike is for the token level.

from codeinspector.

Thanks for the explanation. Closing this issue.

from codeinspector.