URLs containing `?' are not fully comprehended

URLs containing `?' are not fully comprehended

[Karl Kleinpaste]

The following is a legitimate URL.

mailto:him@over.there?subject=testing

It should induce a message buffer with Subject already filled in.

Technically, "?body=stuff" should work, too, but outside actual HTML
<a href> enclosures, I suspect no heuristic will work well because of
the likelihood of needing to catch spaces in body content, whereas
one-word subjects are pretty common.

Anyhow, "?anything" should still be considered part of the URL when
buttonizing and highlighting.

Re: URLs containing `?' are not fully comprehended

[Arne Jørgensen]

Karl Kleinpaste <karl@charcoal.com> writes:

> The following is a legitimate URL.
> 
> mailto:him@over.there?subject=testing

Are you sure? I thought it was just a Netscape and InternetExplorer
'feature' ?

In a http-request ? (and &) can be part of the url though.

        /arne
-- 
stud. scient. Arne Jørgensen
Børglumvej 2, v. 351, DK-8240 Risskov, Denmark
phone: +45 89 37 93 51, cellular: +45 21 65 01 13
email: arne@daimi.au.dk, http://www.daimi.au.dk/~arne/

Re: URLs containing `?' are not fully comprehended

[Bjørn Mork]

arne+usenet@daimi.au.dk (Arne Jørgensen) writes:
> Karl Kleinpaste <karl@charcoal.com> writes:
> 
> > The following is a legitimate URL.
> > 
> > mailto:him@over.there?subject=testing
> 
> Are you sure? I thought it was just a Netscape and InternetExplorer
> 'feature' ?

>From RFC 2368:

  Following the syntax conventions of RFC 1738 [RFC1738], a "mailto"
  URL has the form:
     mailtoURL  =  "mailto:" [ to ] [ headers ]
     to         =  #mailbox
     headers    =  "?" header *( "&" header )
     header     =  hname "=" hvalue
     hname      =  *urlc
     hvalue     =  *urlc


Bjørn

Re: URLs containing `?' are not fully comprehended

[Simon Josefsson]

arne+usenet@daimi.au.dk (Arne Jørgensen) writes:

> > The following is a legitimate URL.
> > 
> > mailto:him@over.there?subject=testing
> 
> Are you sure? I thought it was just a Netscape and InternetExplorer
> 'feature' ?

No, the syntax is described in rfc2368:

 mailtoURL  =  "mailto:" [ to ] [ headers ]
 to         =  #mailbox
 headers    =  "?" header *( "&" header )
 header     =  hname "=" hvalue
 hname      =  *urlc
 hvalue     =  *urlc

And #mailbox is of course one or more rfc822-"mailbox", so the problem
of recognizing mailto URLs is not simple.

Re: URLs containing `?' are not fully comprehended

[Kai Großjohann]

Simon Josefsson <jas@pdc.kth.se> writes:

> And #mailbox is of course one or more rfc822-"mailbox", so the problem
> of recognizing mailto URLs is not simple.

With percent and bang paths and stuff?  Whee!

kai
-- 
Beware of flying birch trees.

Re: URLs containing `?' are not fully comprehended

[Toby Speight]

Kai> Kai Großjohann <URL:mailto:Kai.Grossjohann@CS.Uni-Dortmund.DE>

0> In article <200005162100.XAA29747@marcy.cs.uni-dortmund.de>, Kai
0> wrote:

Kai> Simon Josefsson <jas@pdc.kth.se> writes:

>> And #mailbox is of course one or more rfc822-"mailbox", so the
>> problem of recognizing mailto URLs is not simple.

Kai> With percent and bang paths and stuff?

You have to write your percent symbol as "%25", of course.

Re: URLs containing `?' are not fully comprehended

[Andreas Fuchs]

Today, Toby Speight <streapadair@gmx.net> wrote:
>>> And #mailbox is of course one or more rfc822-"mailbox", so the
>>> problem of recognizing mailto URLs is not simple.
Kai> With percent and bang paths and stuff?
> You have to write your percent symbol as "%25", of course.

And what was the C trigraph for '?' again?

kind regards,
-- 
Andreas Stefan Fuchs                             in Real Life aka
asf@acm.org, asfuchs@gmx.at, asf@ycom.at         in NNTP and SMTP,
antifuchs                                        in IRCNet and
Relf Herbstfresser, Male 1/2 Elf Priest          in AD&D

Re: URLs containing `?' are not fully comprehended

[Per Abrahamsen]

Karl Kleinpaste <karl@charcoal.com> writes:

> Anyhow, "?anything" should still be considered part of the URL when
> buttonizing and highlighting.

It isn't, but I don't understand why.  Here is the regexp:

(defcustom gnus-button-url-regexp "\\b\\(\\(www\\.\\|\\(s?https?\\|ftp\\|file\\|gopher\\|news\\|telnet\\|wais\\|mailto\\):\\)\\(//[-a-zA-Z0-9_.]+:[0-9]*\\)?\\([-a-zA-Z0-9_=!?#$@~`%&*+|\\/:;.,]\\|\\w\\)+\\([-a-zA-Z0-9_=#$@~`%&*+|\\/]\\|\\w\\)\\)"

As far as I can see, the relevant part is

        \\([-a-zA-Z0-9_=!?#$@~`%&*+|\\/:;.,]\\|\\w\\)+
                        -^-

which *does* include question mark.

Re: URLs containing `?' are not fully comprehended

[Karl Kleinpaste]

Per Abrahamsen <abraham@dina.kvl.dk> writes:
> As far as I can see, the relevant part is
>         \\([-a-zA-Z0-9_=!?#$@~`%&*+|\\/:;.,]\\|\\w\\)+

That regexp always makes my eyes hurt, but I think the segment in
error is
        [-a-zA-Z0-9_=#$@~`%&*+|\\/]
which lacks `?' (and `!', and a bunch of others...which raises the
question of why this segment isn't just like the preceding segment).

Re: URLs containing `?' are not fully comprehended

[Per Abrahamsen]

Karl Kleinpaste <karl@charcoal.com> writes:

> That regexp always makes my eyes hurt, but I think the segment in
> error is
>         [-a-zA-Z0-9_=#$@~`%&*+|\\/]
> which lacks `?' (and `!', and a bunch of others...which raises the
> question of why this segment isn't just like the preceding segment).

If you notice, there is no "*" or "+" after that section, it is there
to ensure that a following punctuation sign isn't considered part of
the URL.  This way, sentences like

        See http://www.dict.com, everything is there.

        Did you found it on http://www.google.com?

        I can't connect to http://slashdot.org.

will do "the right thing".

Now lets's see if it does ;-)

Re: URLs containing `?' are not fully comprehended

[Toby Speight]

Karl> Karl Kleinpaste <URL:mailto:karl@charcoal.com>

0> In article <vxku2fytkso.fsf@mesquite.charcoal.com>, Karl wrote:

Karl> The following is a legitimate URL.
Karl>
Karl> mailto:him@over.there?subject=testing

While we're on URL highlighting, some folks use <...> instead of
<URL:...> to delimit multi-line URLs.  Is it possible for us to
accommodate this use?  (IME, it's becoming more common)

Re: URLs containing `?' are not fully comprehended

[Steinar Bang]

>>>>> Toby Speight <streapadair@gmx.net>:

> While we're on URL highlighting, some folks use <...> instead of
> <URL:...> to delimit multi-line URLs.  Is it possible for us to
> accommodate this use?  (IME, it's becoming more common)

Well... Message-Ids also commonly reside inside <...>.

Re: URLs containing `?' are not fully comprehended

[Toby Speight]

Steinar> Steinar Bang <URL:mailto:sb@metis.no>
>>>>> Toby Speight <streapadair@gmx.net>:
>>
>> While we're on URL highlighting, some folks use <...> instead of
>> <URL:...> to delimit multi-line URLs.  Is it possible for us to
>> accommodate this use?  (IME, it's becoming more common)


0> In article <whbt25rxra.fsf@viffer.metis.no>, Steinar wrote:

Steinar> Well... Message-Ids also commonly reside inside <...>.

But they don't normally[1] begin with "http:", "ftp:", "mailto:", or
any of the other scheme names Gnus recognises.  I'm not after matching
bare hostnames in this case (as I think they're evil anyway), just that
the existing[2] matches be multi-line if they're immediately preceded
by "<".  I think we're reaching the limits of regexps with this stuff
now - but is there a sufficiently user-configurable alternative?

[1] I can handle _occasional_ false positives
[2] before the over-liberalisation that's caused so much talk here

Re: URLs containing `?' are not fully comprehended

[Russ Allbery]

Toby Speight <streapadair@gmx.net> writes:

> While we're on URL highlighting, some folks use <...> instead of
> <URL:...> to delimit multi-line URLs.  Is it possible for us to
> accommodate this use?  (IME, it's becoming more common)

The <URL:...> form is basically dead, unfortunately, because it was a
recommendation of an RFC that's now been superseded by a newer RFC that
suggests just using <...> because that's what everyone is actually doing
and because almost no one used <URL:...>.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>