------art_1361_10337166.1171613440754
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

2007/2/15, mathew <meta / pobox.com>:
>
> On 2/14/07, Jonas Pfenniger <zimbatm / oree.ch> wrote:
> > 2007/2/13, mathew <meta / pobox.com>:
> > > Those are not errors. Username and password are not allowed in HTTP
> > > URIs.
> >
> > Right, but since it seems handled by ruby's URI, why not make it
> > completely right ?
>
> I feel like there's some strange new usage of the term "right" that
> I'm not privy to.


First you'll have to learn frenchglish :-p

If we're going to make the URI class accept invalid URIs so long as
> they can be parsed generically, then the patch needs some more work:
>

I must have misexpressed myself. What I was arguing is that you can't reset
the password or the user and password on an already parsed uri. It happens
that the current implementation understands userinfo on the http scheme but
that's another case.

-- 
Cheers,
  zimbatm

------art_1361_10337166.1171613440754
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

2007/2/15, mathew &lt;<a hrefailto:meta / pobox.com">meta / pobox.com</a>&gt;:<div><span classmail_quote"></span><blockquote classmail_quote" styleorder-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On 2/14/07, Jonas Pfenniger &lt;<a hrefailto:zimbatm / oree.ch">zimbatm / oree.ch</a>&gt; wrote:<br>&gt; 2007/2/13, mathew &lt;<a hrefailto:meta / pobox.com">meta / pobox.com</a>&gt;:<br>&gt; &gt; Those are not errors. Username and password are not allowed in HTTP
<br>&gt; &gt; URIs.<br>&gt;<br>&gt; Right, but since it seems handled by ruby&#39;s URI, why not make it<br>&gt; completely right ?<br><br>I feel like there&#39;s some strange new usage of the term &quot;right&quot; that<br>
I&#39;m not privy to.</blockquote><div><br>First you&#39;ll have to learn frenchglish :-p <br></div><br><blockquote classmail_quote" styleorder-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
If we&#39;re going to make the URI class accept invalid URIs so long as<br>they can be parsed generically, then the patch needs some more work:<br></blockquote><div><br>I must have misexpressed myself. What I was arguing is that you can&#39;t reset the password or the user and password on an already parsed uri. It happens that the current implementation understands userinfo on the http scheme but that&#39;s another case.
<br></div></div><br>-- <br>Cheers,<br>&nbsp;&nbsp;zimbatm

------art_1361_10337166.1171613440754--