Update codespell dictionary?

[Delgan]

Hi!

If I understand well, typos builds part of it's database using a dictionary sourced from codespell: https://github.com/crate-ci/typos/tree/master/crates/codespell-dict

However, this dictionary doesn't seem to have been updated for 2 years. Is there any reason for this?
Codespell has added over 10 000 fixes to its dictionary since then.

[Delgan]

Hum... There's something else I must be missing actually.

I expected "usefull" to be replaced with "useful".
This is listed here for example:

typos/crates/wikipedia-dict/assets/dictionary.txt

Line 4117 in 20b36ca

usefull->useful

However, this is not detected as a spelling mistake.
Why that? Are others dictionaries actually used by typos?
Or maybe it's because "usefull" is actually used as a correction suggestion here, thus disabling the typo?

[epage]

Our primary dictionary is in typos-dict. The others exist mostly for comparison purposes.

[Delgan]

Ok I understand, thank you!

I've compared the two tools on my codebase, and although typos finds more errors than codespell, the latter also finds interesting errors that are unfortunately overlooked by typos.

I can imagine that maintaining such a list is very complicated and can hardly be exhaustive.
Had you considered merging the various dictionaries to cover more possible typos?

[szepeviktor]

codespell works with regular expressions and does not chop up variable names, e.g. MissspellingChecker

[Delgan]

@szepeviktor Yes, I know, I prefer typos for this reason. I'm mainly talking about dictionary data that could be reused (without needing to change the implementation).

[Delgan]

Thanks for the update!