mailto hostnames

[stevenvachon]

[email protected] treats domain.com as a path and not a hostname. Why is this so?

[annevk]

Because it doesn't use the generic host syntax (start with //).

[stevenvachon]

Because of this, invalid hostnames are parsed: user@:*&^:. I would think that the hostname part would be either an IPv4, IPv6 or IDNA.

[annevk]

For schemes that do not follow the generic syntax for hosts you'll have to do post-processing. We had this discussion before, no?

[annevk]

#280 was the previous discussion I was thinking of.

[stevenvachon]

Post-processing seems annoying when this spec already explains how to handle hostnames.

[domenic]

Closing as per previous discussion mailto: parsing not special.

Recently a post-processing step was specified for how to extract useful information from data: URLs, in the Fetch Standard. It's done elsewhere because it's a higher-level concern; this standard is only about URL parsing, but it's the concern of software (e.g. browsers that fetch) to deal with whether the parsed URL is a useful format.

Hopefully that can serve as a model if you're interested in writing code or even a spec for how to parse mailto: URLs.