[whatwg] URL standard: Query string parsing; host parsing

[poccil14 at gmail.com]

I sent this, but didn't think to post it to this list.
-----------------------

From: poccil14 at gmail.com
Sent: Wednesday, March 13, 2013 5:08 PM
To: Julian Reschke
Subject: Re: [whatwg] URL standard: Query string parsing; host parsing


My jar example, just like the mailto example, is considered an "opaque" URI,
since its scheme-specific part doesn't begin with a slash.

My concern, however, is not how URLs/URIs should be resolved, but rather how
the URL standard should deal with extracting the query string from URLs and
issues with parsing the host name from URLs.  See my original message.


-----Original Message----- 
From: Julian Reschke
Sent: Wednesday, March 13, 2013 4:23 PM
To: Boris Zbarsky
Cc: whatwg at lists.whatwg.org
Subject: Re: [whatwg] URL standard: Query string parsing; host parsing

On 2013-03-13 21:14, Boris Zbarsky wrote:
> On 3/13/13 4:02 PM, Julian Reschke wrote:
>> On 2013-03-13 18:38, poccil14 at gmail.com wrote:
>>> jar:http://example.com/jar?x=1!/com/example/Foo.class
>>>
>>> is parsed in the URI standards as:
>>>
>>> scheme - jar
>>> scheme-specific part - http://example.com/jar?x=1!/com/example/Foo.class
>>
>> I have no idea what you're talking about, see
>> <http://greenbytes.de/tech/webdav/rfc3986.html#rfc.section.3>.
>>
>> This will parse into:
>>
>> scheme: jar
>> hier-part: http://example.com/jar
>> query: x=1!/com/example/Foo.class
>
> I should note that jar: URIs are ... special.
>
> For example, given a base of
>
>    jar:http://example.com/jar?x=1!/com/example/Foo.class
>
> the relative URI "Bar.class" should, as far as I know, resolve to:
>
>    jar:http://example.com/jar?x=1!/com/example/Bar.class
>
> What that means for parsing them, I cannot say...

Under RFC 3986, it would resolve to

   jar:http://example.com/Bar.class

Looks like a broken scheme to me.

Best regards, Julian