Support non-English characters

[kantord]

fix #9

[kantord]

I'm not sure how to test it, so I haven't tested it in the browser, but I've tested the regexp itself.

[kantord]

Also not sure if it could be somehow extended to languages that don't need to put non-word characters between words

[voldikss]

Hi @kantord thanks for this but I feel that this would make it more disturbing for completion.

Because I am a chinese, the regex in the patch will grab all the chinese characters, which are not what we want.

[kantord]

What would be the expected behavior for Chinese text?

[voldikss]

Chinese text should not be grabbed as it is too disturbing

[voldikss]

Actually East Asian people like me usually doesn't use completion from their native language text.

[kantord]

So to understand better: your ideal use case is grabbing completion only from source code? Or is it: only source code + text in foreign languages?

If it's more about code then perhaps it's possible to only enable completion from the <code /> html tag.

Or perhaps also stuff like code review threads on GitHub/GitLab

[voldikss]

So to understand better: your ideal use case is grabbing completion only from source code?

No matter whether from souce code. What I originally want is to grab English words for completion. The language words in your screenshot is also acceptable, but your regular expression doesn't filter out east-asian characters.

[kantord]

Technically it's possible to filter out only certain scripts. For example the Unicode property Latn will only include characters from the Latin script, what do you think?

TBH it's not a big deal for me either, as my main use case is autocompletion from pull requests and other English text, I just noticed by accident that certain characters are not supported

[voldikss]

I feel that this is useful sometime but maybe English words completion is sufficient. I tend to not include this feature, what's your thoughts?

[kantord]

yup, that makes sense