Whole word scanning

[jamesmaa]

Fixes #889

• : Popup flickers when moving within characters in a word, probably because yomitan considers the lookups to be different
• : Doesn't work on google.com apparently according to @Casheeew

[github-actions]

⚠️ Visual differences introduced by this PR; please validate if they are desirable.

View Playwright Report (note: open the "playwright-report" artifact)

[Kuuuube]

Is there any upper bound for how long a word is/how far backwards it can seek? I'm not seeing anything and this could lead to quite a lot of trouble if a user tries to scan an enormous string. Even a huge limit like 1000 characters would probably be a good idea.

[jamesmaa]

@Kuuuube it uses this._scanLength as the max limit for looking back. Basically it tells yomitan to try to set the startOffset back by this._scanLength characters and indicates for it to stop going back if it sees a whitespace.

Reference
https://github.com/themoeway/yomitan/pull/1330/files#diff-255481c4f025ab785ed8746ef13607c1fcf4ad6fc4e7ee7aecd69a25244bf992R474

[Kuuuube]

Looks good, will approve when tests are fixed. test/options-util.test.js needs the version bumped from 50 to 51.

[StefanVukovic99]

Words in quotes might not be scannable. Also not sure why the first word in the example here can't be scanned:
Screencast from 08-22-2024 09:13:40 PM.webm

Probably seeking back into other HTML elements and concatenating. Layout-aware scanning doesn't seem to impact it.

[Kuuuube]

Might want to have it pull the sentence termination characters and break on any of them. Though there is also the issue of should it stop on ' because some words contain ' but it could also be used to quote things.

[Casheeew]

Probably seeking back into other HTML elements and concatenating. Layout-aware scanning doesn't seem to impact it.

Might be the same or similar issue to what i demonstrated on discord.

[Casheeew]

Might want to have it pull the sentence termination characters and break on any of them. Though there is also the issue of should it stop on ' because some words contain ' but it could also be used to quote things.

I think it's safe to just look one character before the quote and break if it's a whitespace, and not otherwise.

Or rather, look for a whitespace and any combination of quotes, to accommodate the case if the string is doubly quoted.

[Casheeew]

It now works on double quoted words but not on single quoted words.

[Casheeew]

Otherwise lgtm

[jamesmaa]

Okay done I followed your advice and peeked at the character before the single quotes to see if it's an apostrophe or wrapped in single quotes