💡: `caseSensitive` word by word

[JounQin]

Problem

We want IPv4, not ipV4, ipv4 nor Ipv4. But we allow install, Install, INSTALL.

Solution

Extending the caseSensitive option into list:

caseSensitive:
  - IPv4

Alternatives

No response

Additional Context

No response

Code of Conduct

• [x] I agree to follow this project's Code of Conduct

[Jason3S]

@JounQin,

Words list are case aware, but they generate non-case sensitive versions on load. There are directive you can add to your word list file.

my-words.txt

# This is a comment

# The following directives are supported: 
# - 'no-split' - do not split words (default)
# - 'split' - split words on spaces 
# - 'generate-alternatives' - generate non-case sensitive and accent alternatives (default)
# - 'no-generate-alternatives' - do not change the case.

# Example
# cspell-dictionary:split

This line will be split into individual words. 
# The words will be: `This`, `line`, ... `words.`
# By default, lower case versions are generated for when `caseSensitive` is `false`.

# cspell-dictionary:no-generate-alternatives

IPv4 # IPv4 is allowed, but ipv4 is not in this list.

!ipv4 # use `!` to forbid a word, this is useful if it is in another dictionary and you do not want it.

~english # `~` prefix is used to specify use only in non-case sensitive checking.

# cspell-dictionary:generate-alternatives
# Start generating again.

Resumé # will also generate `~resume`

[JounQin]

Thanks @Jason3S first!

Your solution should work absolutely, but IMO it's not that convenient when building a high level linter built on top of ESLint + cspell like https://github.com/alauda/doom/blob/47379bb0459377c003abc931ea0a8b8bfabc45bf/src/cli/lint.ts#L47-L59

Is there any downside to support what I'm proposing? I believe the interface is just simple enough.

[JounQin]

Friendly ping @Jason3S.

[Jason3S]

I'm very reluctant to add another type of word list, there are already 4 (words, ignoreWords, flagWords, and suggestions).

All word list are by definition case sensitive. For speed purposes, they also make a parallel list with case and accents removed. This second list is used when caseSensitive is false, which it is by default.

#7351 Adds the ability to specify supportNonStrictSearches: false when specifying a dictionary. This will prevent the spell checker from auto generating the second list.

[Jason3S]

Example: https://github.com/streetsidesoftware/cspell/pull/7352/files#r2095017131

[Jason3S]

I'm going to close this as completed. There is a way to achieve the desired outcome. So far the only issues seen is with respect to words found in other dictionaries.

[github-actions[bot]]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.