It would be nice if autolink could be matched on more leading punctuations #235

New Issue

claunia · 2026-01-29T14:31:01Z

claunia commented

2026-01-29 14:31:01 +00:00

Originally created by @jijiechen on GitHub (Oct 27, 2018).

In current implementation of Markdig as of 0.15.2, links are not automatically created if there are more kinds of leading punctuations than the ones recognized by Markdig at here:

https://github.com/lunet-io/markdig/blob/master/src/Markdig/Extensions/AutoLinks/AutoLinkParser.cs#L34

(Only whitespace, *, *, _, ~ ( are accepted for now).

So if links could be matched and created on more kinds of leading punctuations, that'll be nice. Either list more allowed punctuations or make the IsValidPreviousCharacter method an instance method and overridable is helpful.

I know and fully understand that it's a normal manner to put whitespaces before links in English (because links are just new words in sentences.) But in other languages, like Chinese, it's also normal there are other punctuations before links.

For instance:

请访问网址查看更多：https://share.weiyun.com/5D8g7nO

Punctuations need to be matched are:

new []{ '：', '。', '，', '、', '—', '；', '！', '？', '（' };

Originally created by @jijiechen on GitHub (Oct 27, 2018). In current implementation of Markdig as of 0.15.2, links are not automatically created if there are more kinds of leading punctuations than the ones recognized by Markdig at here: https://github.com/lunet-io/markdig/blob/master/src/Markdig/Extensions/AutoLinks/AutoLinkParser.cs#L34 (Only whitespace, `*`, `*`, `_`, `~` `(` are accepted for now). So if links could be matched and created on more kinds of leading punctuations, that'll be nice. Either list more allowed punctuations or make the `IsValidPreviousCharacter` method an instance method and overridable is helpful. I know and fully understand that it's a normal manner to put whitespaces before links in English (because links are just new words in sentences.) But in other languages, like Chinese, it's also normal there are other punctuations before links. For instance: ``` 请访问网址查看更多：https://share.weiyun.com/5D8g7nO ``` Punctuations need to be matched are: ```cs new []{ '：', '。', '，', '、', '—', '；', '！', '？', '（' }; ```

claunia added the enhancement PR Welcome! labels 2026-01-29 14:31:01 +00:00

claunia closed this issue

2026-01-29 14:31:01 +00:00

claunia commented

2026-01-29 14:31:01 +00:00

@xoofx commented on GitHub (Oct 29, 2018):

Yes, markdig should follow what GitHub is actually expecting for autolinks.

You can make a PR by changing the test:

c.IsWhiteSpaceOrZero() || c == '*' || c == '_' || c == '~' || c == '(';

and using a string instead:

public string ValidPreciousCharacters {get; set; }

ValidPreciousCharacters = "*_~( \t";

ValidPreciousCharacters.Contains(c)  (check other whitespace characters)

PR welcome!

@xoofx commented on GitHub (Oct 29, 2018): Yes, markdig should follow what [GitHub is actually expecting](https://github.github.com/gfm/#autolinks-extension-) for autolinks. You can make a PR by changing the test: ``` c.IsWhiteSpaceOrZero() || c == '*' || c == '_' || c == '~' || c == '('; ``` and using a string instead: ``` public string ValidPreciousCharacters {get; set; } ValidPreciousCharacters = "*_~( \t"; ValidPreciousCharacters.Contains(c) (check other whitespace characters) ``` PR welcome!

claunia commented

2026-01-29 14:31:02 +00:00

@MihaZupan commented on GitHub (Oct 29, 2018):

Would something like #264 be okay?

(Most of it is just exposing the public field)

@MihaZupan commented on GitHub (Oct 29, 2018): Would something like #264 be okay? (Most of it is just exposing the public field)

claunia commented

2026-01-29 14:31:02 +00:00

@jijiechen commented on GitHub (Oct 29, 2018):

This is a just-pushed commit!
I think so, it should match the requirement. Many thanks!

@jijiechen commented on GitHub (Oct 29, 2018): This is a just-pushed commit! I think so, it should match the requirement. Many thanks!

claunia commented