jQuery Forum

I have used .clone() to date always in the context of ui-draggable helpers etc, ie always on anonymous DOM elements.  I today for the first time had occasion to duplicate a properly deep DOM structure, full of id'ed elements.  I was surprised (not sure why) to find that .clone() faithfully clones every id.  Therefore, after I inserted the clone into the DOM, the DOM had two ids of everything.  Is this a problem?  I retrieved the children I needed to target with $("no-longer-quite-so-unique-id",clone).

Thanks.

I have progressed to the user authentication of a web app I'm building.  Implemented the mechanics of the login process in HTTP originally.  The idea is for the login process to create a PHPSESSID which the SQL PHP would then recover.  Worked all fine with HTTP.  I then proceeded to enable SSL to hide the login dialog and session id.  The bulk of the app should continue to download stuff using HTTP.  So I have a sequence like HTTPS redirects to HTTP makes $.ajax calls to HTTPS.   The latter bit crashes big time.  Looking at Firebug, the Net panel reports that Firefox fired an OPTIONS header at my PHP script, which that script did not handle well.  The script tried to recover the PHPSESSID from the OPTIONS header, failed and returned my "No active session" error.  I understand that the problem results from HTTP://www.mydomain.com and HTTPS://www.mydomain.com being seen as different domains.  (I'd actually appreciate a heads-up as to why a mere protocol change is considered toxic by people smarter than myself.  It seems to me that it's all still the same domain...)  And that Firefox is trying to establish with an Access-Control-Request whether my server would like to oblige.  It seems prima facie that there may be something that I am required to do on my server.  However, in researching this problem I stumbled upon JSONP.  I have wondered before what that JSONP stuff in $.ajax was all about but since I managed to get by without it thus far, I never investigated.  Reading some forum posts on JSONP, it seems that it might be a way around my problem. I have a few questions:

(1)

Why is it ok to download JS from 3rd parties, but not make $.ajax calls?  Naively, I had thought that the former presents a considerably greater security risk than the latter.  (What's the worst a $.ajax call could pull down?  Probably JS actually...)

(2)

What do I have to do with my PHPSESSID cookie?  I think that I should flag it "secure=true" such that it will only be sent along HTTPS connections.  How will this affect my JSONP request?  I am guessing that the script tag is loaded via HTTP?  So will the SQL PHP script receive the cookie?

(3)

My server currently returns JSON data like so {}.  JSONP is something like so callback= function() { return {}; }.  I understand that I can configure the $.ajax call to do the padding for me.  Should I?  Or is it more efficient/better to rewrite my PHP script?

Thank you for giving this post your attention (I apologise for the length).

Thanks.

Thanks.

dialog beneath (depending on mouse placement, of course).  I can put a stop to this by attaching .bind("mousedown",function() {return false;}); to each canvas, but feel a bit queasy about this.  Anyone's come across this before and perhaps found a more subtle (CSS?) way to stop this?

Thanks.

Thanks.

    $.ajax({

         ...

         async: false,

         success: function(data) {

              // do some sluggish post-processing

              alert(2);

         }

     });

     alert(1);

And on my multi-processor, multi-threaded box I get the alerts in order 1, 2.  I frankly expected 2, 1.

It therefore seems to me that a synchronous $.ajax call returns as soon as the data has been received, and not (as I think is very reasonable to expect) when the success() or error() function has returned.  This throws up problems with any expensive data processing (eg JSON parsing) but even with the most basic code like

Sorry to post this really dumb question.  I built myself a UI widget like so:

$.widget( "myWidget", {

    _init: function() {

        this.element.bind( "resize", function() {

            ???????._myPrivateFunc();

        } );

    },

    _myPrivateFun: function() {

        // do something

    }

});

and instantiate like so

var div= $("<div>").myWidget();

and later on fire a "resize" event like so

div.trigger("resize");

Good news is: the code finds its way into my resize handler.  Bad news is:  I have no idea how to find from the DOM element "this" (inside that handler) back to the widget instantiation such that the call to _myPrivateFunc() would actually resolve.

Could anyone point me in the right direction?  Thank you kindly.

PS:  I should mention that the firing code has no idea that myWidget is attached.  That is, it would not know how to look for the widget context, nor would it know how to call _trigger.

Hi,

A quick question re the ui object passed by Resizable into the "resize" event handler.  Every time I use Resizable, I find myself having to manually resize one or more children.  To that end, I need to understand by how much the event driver has actually resized.  My immediate thought was therefore to calculate:

dx= ui.size.width - ui.originalSize.width;

dy= ui.size.height - ui.originalSize.height;

Obviously, what I understand/want [dx, dy] to be is the change in dimensions over the previous step (so that I can mirror the incremental adjustment that has taken place).  What happens instead is that originalSize never changes and my [dx, dy]s work out to be cumulative, relative to the (fairly arbitrary and somewhat irrelevant (coz long forgotten)) original size of the Resizable widget.

I have obviously solved this dilemma by carrying around my own state data for "originalSize".  It's tedious but I live.  What I am curious to find out is whether ANYONE here found any use for ui.originalSize?  In other terms, would a request to change the definition/behaviour of ui.originalSize give rise to any compatibility issues with existing code?  Or have I perhaps missed a trick and there's another snippet of information lying about that would allow me to calculate the incremental [dx, dy] on the fly, without the need to carry my own state?

Thanks.