urlize parses email addresses with apostrophes incorrectly.

[Mark Gregson]

In 1.11.21 I've observed that the urlize template tag messes up email addresses containing apostrophes in the local part, specifically, it excludes any characters before the apostrophe from the link text and href attribute.

A simple example using django.utils.htmls.urlize:

>>> from django.utils.html import urlize
>>> urlize("o'neill@example.com")
u'o\'<a href="mailto:neill@example.com">neill@example.com</a>'

I would expect this to return:

u'<a href="mailto:o\'neill@example.com">o\'neill@example.com</a>'

[Mark Gregson]

Given the range of characters allowed in the local part of an email address is quite extensive and that a quoted local part is also allowed (RFC 3696 has a nice summary ​https://tools.ietf.org/html/rfc3696#section-3), I realise that there probably need to be some constraints on what range of email addresses urlize can parse. Potentially then this issue should be resolved by additional documentation clarifying those constraints.

[Mariusz Felisiak]

Thanks for the report. We excluded quotes in the #24471. It seems that only single quotes from the list of allowed special characters are affected. I agree that we can clarify this in the documentation, e.g. Email addresses cannot contain the single quotes (``'``).

[Pystar]

I will work on this.

[Hasan Ramezani]

PR​https://github.com/django/django/pull/11526

[Mariusz Felisiak <felisiak.mariusz@…>]

In c2f381e:

Fixed #30589 -- Clarified that urlize should be applied only to email addresses without single quotes.

[Mariusz Felisiak <felisiak.mariusz@…>]

In b9d1bb69:

[2.2.x] Fixed #30589 -- Clarified that urlize should be applied only to email addresses without single quotes.

Backport of c2f381ef17058e5cfea58ae507983d2e459a2888 from master