Currently the parser can recognize opening parentheses and closing parentheses and exclude closing parentheses when appropriate, while we don't have the same behavior with fullwidth characters. See this example:
import { tokenize } from "linkifyjs";
const links = [
"http://foo.com/blah_blah",
"http://foo.com/blah_blah_(wikipedia)_(again)"
];
const texts = [
`${links[0]} ${links[1]}`,
`Link 1(${links[0]}) Link 2(${links[1]})`, // halfwidth parentheses
`Link 1（${links[0]}） Link 2（${links[1]}）`, // fullwidth parentheses
];
for (const text of texts) {
const tokens = tokenize(text);
tokens.filter(token => token.isLink).forEach((token) => console.log(`"${token.v}"`));
}
// texts[0]: succeed without parentheses
// "http://foo.com/blah_blah"
// "http://foo.com/blah_blah_(wikipedia)_(again)"
// texts[1]: succeed with halfwidth parentheses
// "http://foo.com/blah_blah"
// "http://foo.com/blah_blah_(wikipedia)_(again)"
// texts[2]: fail to handle fullwidth parentheses
// "http://foo.com/blah_blah）"
// "http://foo.com/blah_blah_(wikipedia)_(again)）"
My proposal is to define fullwidth characters as tokens, and add new behaviors in the parser.
The logic should be fairly simple as fullwidth brackets are semantically the same as their halfwidth counterparts.
(In our use case we care more about fullwidth parentheses （）, but in general this can apply to other fullwidth characters, e.g. 「」『』＜＞.)
Currently the parser can recognize opening parentheses and closing parentheses and exclude closing parentheses when appropriate, while we don't have the same behavior with fullwidth characters. See this example:
My proposal is to define fullwidth characters as tokens, and add new behaviors in the parser.
The logic should be fairly simple as fullwidth brackets are semantically the same as their halfwidth counterparts.
(In our use case we care more about fullwidth parentheses
（）, but in general this can apply to other fullwidth characters, e.g.「」『』＜＞.)