fix/emojis-and-vulgar-fractions: hunspell should not check it

[laysauchoa]

What does this PR accomplish?

• 🩹 Bug Fix
• 🦚 Feature

Closes #141 .

Changes proposed by this PR:

Notes to reviewer:

📜 Checklist

• Works on the ./demo sub directory
• Test coverage is excellent and passes
• Documentation is thorough

Test for emojis

error: spellcheck(Hunspell)

  --> /home/tmhdev/Documents/rust_samples/recursion_rust/src/fibonnacci.rs:2
   |
 2 |  😘🤗🦀
   |  ^^^
   |   Possible spelling mistake found.

error: spellcheck(Hunspell)
  --> /home/tmhdev/Documents/rust_samples/recursion_rust/src/fibonnacci.rs:2
   |
 2 |  😘🤗🦀⅔⅔
   |  ^^^^^
   |   Possible spelling mistake found.

error: spellcheck(Hunspell)
  --> /home/tmhdev/Documents/rust_samples/recursion_rust/src/fibonnacci.rs:3
   |
 3 |  ⅔⅔
   |  ^^
   |   Possible spelling mistake found.

I ran again and the possible mistake is not seen.

This PR refers to #141

[laysauchoa]

Just checking if the changes will be in this part of the code. Still need to find something for fractions, maybe https://crates.io/crates/unicode-segmentation

[drahnr]

The code location is correct 👍 , it would be nice to make it a separate module, since other backends might want to allow whitlisting emojis and graphemes as well.
Note that graphemes are multiple unicode characters, so a more fully future lib might make more sense i.e. unic

[laysauchoa]

ok, not sure if the other unicode should also be included here besides emojis and vulgar fractions. I only focused on those two. Feel free to drop any tips about it. I could not find in unic a better way to handle with vulgar fractions, maybe I am missing something? Another possibility, would be to find if the char is in the ranges of vulgar fractions but maybe not so readable.

[drahnr]

There are a few performance concerns, the general approach looks really good 👍

[drahnr]

Small set of changes, but other than that good to go, I took the liberty to fix the botched merge commit.

[laysauchoa]

Hi @drahnr, thanks for fixing the test back to working 🙏 . So if regex is a too expensive operation, and also convert to the name as it was done before, then another option would be have the VULGAR FRACTIONS* hard-coded and then simple check if a char is on that list and use unic emoji for check of the emoji. What do you think? Which approach you propose and do you mean to add a flag to enable the check of chars and emoji?

[drahnr]

The expensive part is the refer compilation, not applying the compiled regexp :) so if you wrap that in a lazy static it should be fine.

If you have a better/faster idea, go for it. If unsure, measure with i.e. a criterion bench or a simple cputime crate.

[drahnr]

I tend to prefer unic since the general plan would be to move in that direction anyways

[drahnr]

Ping 🙃

[drahnr]

@laysauchoa finally got around to taking another look at this, and I'd say this is good to go as is now :)

Old

[drahnr]

The CI issue is unrelated to these changes, it's just the first time it popped up - filed bminixhofer/nlprule#68 but is most likely a lazycell issue.