On 31.10.2017 00:45, Bruce Horrocks wrote:
> Good answers to the query to extract an email address. Maybe it's time to
> badger the Posix people to add [:email:] and [:url:] to the list of known ranges?
>
> And possibly the sub-parts, e.g. [:email_person:]@[:email_server:] (but they
> only match on the appropriate part of whatever string [:email:] would match
> (otherwise 'email_person' is pretty much [:alpha:] but without knowing how
> many characters to match)).
>
> Thoughts?
This would open a can of worms. Currently there are simple lexical character
types defined, with your proposal there's semantical information introduced;
and there are countless semantical items out there!
Given that most (all?) such or similar technical items like email addresses
can be represented by regexps it would probably be more sensible to define a
library of regexps for such purposes.
Janis