Ignore 'misspellings' due to string escapes

[Zac-HD]

This issue has been reported many times - #17, #111, #233 (comment), #1421, #1774, #2508, #2809, and #2871 at least. It would also let me enforce codespell at work, which would be nice!

Design: in order to make this super-fast, we do an O(1) check of each detected misspelling, with no new configuration or dictionary entries. If the match is preceeded by a backslash and begins with a letter which (with backslash) constitutes a common escape string, and the remainder of the word is not a misspelling, we skip it; otherwise we report exactly as we would now, with no special handling for the escapes. It's a minimalist change meant to address a common complaint, accepting a very few missed alarms in exchange for lower complexity for both maintainers and users.

Fixes #17, fixes #1774, and closes #174.

[DimitriPapadopoulos]

This change would be a change of paradigm: codespell is currently file type-agnostic. This change introduces a feature that is specific to files types that that support \ as an escape character, without even relying on the current regexp mechanism. That said, such file formats are common enough that we can give this a thought.

Also, this feature cannot be disabled, it is not optional.

[larsoner]

Agreed @DimitriPapadopoulos this might cause some misses in some rare circumstances (binary files maybe?) but 1) it should help with false alarms and 2) it seems like 99% of people will want it "on" so I don't think we need to add an option yet. I'm inclined to fix the test (will commit shortly) then merge

[larsoner]

Okay let's give it a shot, thanks @Zac-HD !

[Zac-HD]

Thanks all! Any idea when the next release will be, so I can use it?

[larsoner]

In theory https://github.com/codespell-project/codespell/releases/tag/v2.2.5 should cause a push to pip soon

[peternewman]

This test relied on ndoes being in the dictionary, which I don't think it is.

I also found when adding a positive test that it doesn't match/trigger anyway because of the apostrophe at the start.

I've added some more thorough tests in #2879 .

[peternewman]

This won't deal with anything where someone has put in e.g. \ to render a backslash (or \\ etc) and the following (unrelated) word is a typo.

[peternewman]

This change would be a change of paradigm: codespell is currently file type-agnostic. This change introduces a feature that is specific to files types that that support \ as an escape character

Agreed @DimitriPapadopoulos this might cause some misses in some rare circumstances (binary files maybe?)

RTF seems like it might possibly be the biggest/first hit (or maybe it's delimiters save it):
https://en.wikipedia.org/wiki/Rich_Text_Format#Code_syntax

Also, this feature cannot be disabled, it is not optional.

2. it seems like 99% of people will want it "on" so I don't think we need to add an option yet.

Personally I think it would be nice to be able to turn it off, if someone finds an issue with it, or they just want to be more thorough. It wasn't very hard to find an edge case in #2880 .