Re: Parsing URL attributes - url-mailto, url-imap, url-ldap broken?

[Chong Yidong <cyd@stupidchicken.com>]

Richard Stallman <rms@gnu.org> writes:

> Is anyone working on this?

AFAICT, Diane Murray's analysis is correct.  The 2006-10-09 change was
intended to parse http query strings better, but url-http apparently
doesn't use the result; on the other hand, url-mailto and other parts
of url that depended on the old behavior got broken.

Since Magnus Henoch has not responded, I went ahead and reverted the
patch that caused this problem.  After the release, we can re-apply
the patch, and adapt url-mailto etc to the new behavior of
url-generic-parse-url.

> Date: Thu, 12 Apr 2007 17:02:56 +0200
> From: Diane Murray <disumu@x3y2z1.net>
> To: emacs-devel@gnu.org
> Subject: Parsing URL attributes - url-mailto, url-imap, url-ldap broken?
>
> I just noticed that `url-mailto' doesn't parse URLs with query
> arguments like it used to.  If I call
> (url-retrieve-synchronously "mailto:ml+atest@x3y2z1.net?subject=testing")
> the word "testing" should be inserted in the Subject field, but it
> isn't.  In Emacs CVS of 2006-09-19 it does that.  I'm assuming after
> because of the following changes, now it doesn't:
>
>   2006-10-09  Magnus Henoch  <mange@freemail.hu>
>
>           * url-parse.el (url-generic-parse-url): Handle URLs with empty
>           path component and non-empty query component.  Untangle path,
>           query and fragment parsing code.  Add references to RFC 3986 in
>           comments.
>           (url-recreate-url-attributes): Start query string with "?", not ";".
>
> I think `url-imap' and `url-ldap' won't work correctly either, from
> what I've seen in the source code.
>
> Since the change was probably made to parse http query strings better,
> and there is no such use of url-attributes in url-http.el, I believe
> it should be safe to somehow revert to the old way, since many of the
> other url scheme files depend on that way of doing things.