jQuery Forum

I'm not an url expert either but the // part was added for this ticket
http://dev.jquery.com/ticket/1750
it is of course possible another case exists that should be fixed.
David

I'm definitely no expert (at anything!) but it seems to me that the
sole intent of that test is to catch 'off-site' script gets, by
checking that the url begins with either 'http' or '//', and is a
script get.
Ariel's change modifies the test to be : begins with 'http' or '/',
and does not have the requesting host in it (preceded by '/'), and is
a script get.
Unfortunately this would also grab all 'on-site' root-relative urls,
eg '/myroot/myscripts/script.js', which might not be desirable.
As slight adjustment *might* be simply to reinstate the current double
slash check back into Ariel's proposal (in place of the single slash
check), ie:
- begins with 'http' or '//' (indicating a full url, with/without
protocol), and
- does not contain '/'+[requesting host] (indicating 'off-site'), and
- is a script get
if ( (!s.url.indexOf("http") || !s.url.indexOf("//")) &&
s.url.indexOf("/"+location.host) < 0
&& s.dataType == "script" && s.type.toLowerCase() == "get" ) {

Thanks for the comments.
I now realize the '//' was ok, it was past midnight you know... :)
So.. I'll push it in when I get some time (today for sure). I think a
regex will be clearer.
if( /^(http:)?\/\//.test(s.url) && ..... )
Thanks
--
Ariel Flesler
http://flesler.blogspot.com

<div>Small test suite, should be tested with port and different domains... maybe port isn't neccesary, I'm not sure.</div>
<div> </div>
<div>Please test, thanks
 </div>
<div><span class="gmail_quote">On 5/13/08, <b class="gmail_sendername">Ariel Flesler</b> <<a href="mailto:aflesler@gmail.com">aflesler@gmail.com</a>> wrote:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">
Thanks for the comments.
I now realize the '//' was ok, it was past midnight you know... :)
So.. I'll push it in when I get some time (today for sure). I think a
regex will be clearer.
if( /^(http:)?\/\//.test(s.url) && ..... )
Thanks
--
Ariel Flesler
<a href="http://flesler.blogspot.com">http://flesler.blogspot.com</a>
On 13 mayo, 07:21, Wizzud <<a href="mailto:r...@wizzud.com">r...@wizzud.com</a>> wrote:
> I'm definitely no expert (at anything!) but it seems to me that the
> sole intent of that test is to catch 'off-site' script gets, by
> checking that the url begins with either 'http' or '//', and is a
> script get.
> Ariel's change modifies the test to be : begins with 'http' or '/',
> and does not have the requesting host in it (preceded by '/'), and is
> a script get.
> Unfortunately this would also grab all 'on-site' root-relative urls,
> eg '/myroot/myscripts/script.js', which might not be desirable.
>
> As slight adjustment *might* be simply to reinstate the current double
> slash check back into Ariel's proposal (in place of the single slash
> check), ie:
> - begins with 'http' or '//' (indicating a full url, with/without
> protocol), and
> - does not contain '/'+[requesting host] (indicating 'off-site'), and
> - is a script get
>
> if ( (!s.url.indexOf("http") || !s.url.indexOf("//")) &&
> s.url.indexOf("/"+location.host) < 0
>        && s.dataType == "script" && s.type.toLowerCase() == "get" ) {
>
> On May 13, 7:22 am, David Serduke <<a href="mailto:davidserd...@gmail.com">davidserd...@gmail.com</a>> wrote:
>
>
>
> > I'm not an url expert either but the // part was added for this ticket
>
> ><a href="http://dev.jquery.com/ticket/1750">http://dev.jquery.com/ticket/1750</a>
>
> > it is of course possible another case exists that should be fixed.
>
> > David
>
> > On May 12, 8:49 pm, Ariel Flesler <<a href="mailto:afles...@gmail.com">afles...@gmail.com</a>> wrote:
>
> > > I'm not an url expert, but I think this could be the fix.
>
> > ><a href="http://dev.jquery.com/ticket/2816">http://dev.jquery.com/ticket/2816</a>
>
> > > Thanks
>
> > > --
> > > Ariel Fleslerhttp://flesler.blogspot.com- Ocultar texto de la cita -
>
> - Mostrar texto de la cita -
<br clear="all">
--
Ariel Flesler
<a href="http://flesler.blogspot.com">http://flesler.blogspot.com</a>

Experimented with ports.
document.domain and location.hostname don't include the port part,
location.host does, so it was appropiate to use it.
--
Ariel Flesler
http://flesler.blogspot.com

Closed at http://dev.jquery.com/changeset/5586
Thanks Scott for all the help :)
--
Ariel Flesler
http://flesler.blogspot.com