[FEATURE]: Don't spell check URLs

[lmas]

Feature Description

Hi, a common annoyance for me is seeing warnings for "spelling issues" in URLs.
Being what they are, "fixing" a "misspelled" URL would break it, yes?
I would like to have spell check disabled (or at least have a flag to disable it) for all URLs no matter the context.

Example:

Use Case

Anything I copy an URL into one of my files that are spell checked using codebook.

Proposed Solution

Globally disabling spell check for URLs in any contexts.

Alternative Solutions

Feature flag to allow user to disable spell check of URLs.

Examples

See above.

Additional Context

No response

[blopker]

Hey! I rewrote the regex system (#80) to match regex line-by-line instead of after the word-split. I've also added default regex ignores to include URLs (and file paths). Once Codebook updates, this should be fixed! Let me know if it still gives you issues.

[lmas]

Thank you for that, the feature worked as intended!

But codebook will now run att full throttle, maximising the CPUs and making it unusable as a whole. I've had to to rollback from v0.3.0. Do you want me to open a new issue or should I provide more info here?

[blopker]

Ouch, sorry about that! I was worried about having too many default regex to process.

Could you share your hardware specs, and how large the files are that are causing issues? Realistically, I'll just need to do some profiling, but it would be good to have more info.

[lmas]

Running Helix and FreeBSD on amd64, with a Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz. 8 GB RAM.

# cat thesis.typ | wc
    1153    6406   44776

(that's lines, words, bytes being reported respectively for a file with a mix of prose and Typst syntax)

I was a bit concerned already with codebook@v0.2.13 sucking up around 30-40 % of the CPU time for about 5 seconds when I saved this file, but it was doable.

With the latest version codebook would max out all four CPUs while I was editing a single line (not when the file was saved), which sounds weird? Maybe that can give you a hint.

I'm sorry that I can't read Rust but I can at least try out new version.

[blopker]

Thanks for the extra info! Does this happen on supported code language files, like Python, as well? Or is it only happening on prose-like documents?

[lmas]

Stress testing with a bigger file:

# cat ../python/venv/lib/python3.11/site-packages/matplotlib/patches.py | wc
    4720   16929  163516

Arbitrary results after having opened the editor and started typing out a long line of text (+80 chars) of english words and then stop and wait.

v0.2.13: +300 % CPU time (~8 seconds until cooling down to ~0 %)
v0.3.0: +350 % (~10 seconds)

Re-testing the prose file from before:
v0.2.13: +80 % (~3 seconds)
v0.3.0: +300 % (~10 seconds)

I can't say much about these fluff numbers, new version ran noticeable heavier on the prose (compared to before) and about the same-ish on the large python file. Your best bet is profiling.

Edit: I don't use other LSP servers for Typst or Python.

[blopker]

Ok, I've had coffee now and realize this system is on FreeBSD. I'm surprised/delighted Codebook runs at all on FreeBSD as I've never tested it. On my older Mac (M1), Codebook can spellcheck Beowulf in a few milliseconds with less than 2% usage on one core.

I'll make a new issue for this to give visibility, but I'm thinking it might be an issue with FreeBSD and musl: https://raniz.blog/2025-02-06_rust-musl-malloc/ If you're open to it, I'd like to see if using a different allocator, as this post suggests, shows a significant difference.

[lmas]

Hahaha usually it's only Python that's throwing a tantrum on FreeBSD, so I'm quite surprised rust would have poor OS support too. I've compiled and run alacritty and jujutsu (both rust projects) pretty well too so far without performance issues for example.

I'm not sure I can dive into musl in the next weeks and test that though, what with being a student and working at the same time. I can continue running tests for you though.

Thanks again for looking into this issue and for running this project, appreciate it!

[blopker]

Awesome, thanks for the offer. I've put up a test build using jmalloc on Linux binaries: #81 When you get a chance, I'd love to know if it helps at all.