Add parsing of other things (add requests here)

[r1chardj0n3s]

It'd be neat if it could parse:

• URLs, producing the same result as urlparse.urlparse()
• email addresses, producing a (realname, email address) pair like email.utils.parseaddr()
• IPv4 and IPv6 addresses (producing .. what?)

...?

[jenisys]

The basic feature that is currently missing (IMHO), is that a regular expression for self-defined type ( extra_types) can be provided.

parse pattern: "host: {host:IPv4Address}"

This works already today, but only basic char matching is used.
But what about:

parse pattern: "host: {host:IPv4AddressOrHostname}"

A different datatype would be returned for these 2 choices, the regexp parser could preselect
which choice occured via "(?P<ip_address>...)?|?P...)?" pattern.

It would be nice if the regular expression would already do the checking/matching.
This would help in ambiguous matching areas, like behave.

Example:

#!python
class ParseType4MyNonEmpytList(parse.ParseType):
    """
    Extra Type can also provide an instance of "parse.ParseType" (NEW, used as tag interface).
    If an instance of parse.ParseType is detected, its "pattern" attribute is used 
    instead of the default one.
    """
    pattern = "\w(,\s*\w)*"
    def __call__(self, text):
        seq = ",".split(text)
        return seq

The approach that is sketched above could also be used to define the built-in parse types.
By looking at the IPv4AddressOrHostname example above, additional generic ParseTypes could be provided,
for example by parsing a number of ParseTypes as choice for one parameter.

[jenisys]

I provided a slightly simpler solution as pull-request.

[r1chardj0n3s]

I agree, this is a very good approach, and though I have a couple of quibbles with your patch I do welcome it!

Also I should have IP addresses in the core ;-)

[r1chardj0n3s]

I'm going to close this general bucket as it's been open for well long enough now :)