URL with underscore is not detected [reopen]

[lxgreen]

Sorry for inconvenience, just checked with my client. They have this subdomain name: https://rfinton2_18.gr8.com/, and it is not detected by linkify

[puzrin]

https://stackoverflow.com/questions/2180465/can-domain-name-subdomains-have-an-underscore-in-it

That's invalid hostname. I'm for ignore such case and do nothing.

[puzrin]

Seems no need to fix - no activity.

[eramdam]

@puzrin hey, it seems that the SO thread you linked has had an update since then

https://stackoverflow.com/a/2183140/6398000

Most answers given here are false. It is perfectly legal to have an underscore in a domain name.

I'm running into this exact issue as well. Is there any hope to see this fixed? (Happy to open a PR if you have some pointers of where the change should be made) Ended up opening a PR https://github.com/markdown-it/linkify-it/pull/86

[puzrin]

The answer you refered says, underscores are ok for domain names, but not ok for hostnames. And URL-s uses hostnames (=> should be without underscores). What is wrong?

[eramdam]

@puzrin the issue is that, like the few people who opened this issue and the other similar one, I come accross URLs that are functional and have an underscore in their "domain" part (before the path).

Examples like the one in this comment + the one I encountered (that I cannot share because it's private to the customer of the company I work for).

[puzrin]

That example was artificial, made by user himself. That's not correct. I need something more close to real world.

[eramdam]

The URL I had was https://company_ops.frontapp.com/path (real company name removed for privacy). The underscore is in the sub-domain and not the domain itself, so maybe a cleaner solution would be to allow underscores in subdomains but not in domains?

[puzrin]

To be honest, it's suspicious, when users ask to fix situations, caused by themself. I'd suggest to fix your domain for sure instead. Or you could monkeypatch regexp exports before linkifier use - this will work and not requires upstream changes.

[eramdam]

To be honest, it's suspicious, when users ask to fix situations, caused by themself. I'd suggest to fix your domain for sure instead.

What is there to fix? The URL leads to a working page, I don't see how it's controversial to support it, especially when other linkifiers out there (like Github's one) support that case.

Or you could monkeypatch regexp exports before linkifier use - this will work and not requires upstream changes.

So I tried doing that (from a code that uses markdown-it, mind you) and I haven't had any success because (I assume) the regexes are compiled at runtime and I can't re-compile them from the outside of the module.

[puzrin]

What is there to fix? The URL leads to a working page, so it's valid, I don't see how it's controversial to support it, especially when other linkifiers out there (like Github's one) support that case.

I have a different opinion about this. Any change should not intruduce additional risks and should be useful for all. IMO you wish to push your custom request to mainstream - nobody else need it.

So I tried doing that (from a code that uses markdown-it, mind you) and I haven't had any success because (I assume) the regexes are compiled at runtime and I can't re-compile them from the outside of the module.

I don't remember details. But initially instance was expected to be patch-able. If something missed - this can be fixed.

[eramdam]

nobody else need it.

Given that this is the second issue about this, I would argue otherwise but I'll stop here because we're not going anywhere.

So I tried doing that (from a code that uses markdown-it, mind you) and I haven't had any success because (I assume) the regexes are compiled at runtime and I can't re-compile them from the outside of the module.

I don't remember details. But initially instance was expected to be patch-able. If something missed - this can be fixed.

Here's an example of what I've been doing https://runkit.com/eramdam/5e94eaefe02882001a0d2cd7, applying my (albeit dirty) working fix from https://github.com/markdown-it/linkify-it/pull/86 and it doesn't seem to work. So I must be missing a compile step somewhere to make it so markdownIt.linkify takes the changes into account.

[puzrin]

1. See https://github.com/markdown-it/linkify-it/blob/master/index.js#L158 you can override .onCompile() and patch there loca copies of regexps
2. If use in markdown-it - reassign new instace of likifier, or patch it before md instance create)

[eramdam]

1. See https://github.com/markdown-it/linkify-it/blob/master/index.js#L158 you can override `.onCompile()` and patch there loca copies of regexps

2. If use in markdown-it - reassign new instace of likifier, or patch it before md instance create)

All of this makes sense but... It doesn't seem to work 😕 Or maybe I'm not doing this correctly 😓

Here's a gist with what I've been doing https://gist.github.com/eramdam/4792b9eb7e46c0abc08c7e18826ba5ad#file-test-monkeypatch-js

[magicode]

example https://a_a_a_a_a.wix.com/

@puzrin chrome, edge and github markdown support underscore in hostname
markdown-it is no supported yet