Querypath vs jQuery difference

[jakejackson1]

From querypath created by Jelle-S: technosophos/querypath#113

First of all, I love this library!

On the homepage of Querypath, the library is described as "jQuery for the server". And after using it a couple of times, that's exactly what it is!

But -and I don't think this is a real "issue"- I found a quite large difference with jQuery and I was wondering what the reason for this difference is (and I'm guessing it's performance and memory usage, but I just wanted to check to be sure).

Ok, so in jQuery you can do the following:

var $body = $('body');
// $body represents the body DOM element.
$body.find('div.changeme').text('I am changed');
// $body still represents the body DOM element.
var $some_element = $body
    .find('#footer')
    .find('#copyright')
    .find('.some-class')
    .find('.some-element');
// $body still represents the body DOM element.

If you do the same in Querypath, the $body variable would change constantly:

$body = qp('myhtmlpage.html')->find('body');
// $body represents the body DOM element.
$body->find('div.changeme')->text('I am changed'); 
// $body now represents div.changeme and we would need
// to call $body->end() for it to represent the body DOM
// element again.
$some_element = $body
    ->find('#footer')
    ->find('#copyright')
    ->find('.some-class')
    ->find('.some-element');
// $body now represents .some-element and we can only
// use $body->end() once, twice would result in an empty
// QueryPath object, so we can only go as far back as .some-class.

The alternative would be to clone $this on each call and set the matched elements on the cloned object and then returning the clone in stead of $this. I can imagine how that could be a memory issue, but I'd just like your thoughts on this.

[jakejackson1]

Over the course of QueryPath's history, I've changed this a few times based
on feedback.

In QueryPath 2.x, find() (and most traversing methods) was destructive. To
get around that, I also provided the branch() method, which worked like
jQuery's find() function (e.g. is non-destructive).

In QueryPath 3.x, find() is non-destructive (works like jQuery) and
findInPlace() is destructive. Like you suggest, find() basically clones the
base object and then queries it.

The rationale for QP2's model was, as you guessed, memory management and
speed. But when I rewrote the internals on QP3, those weren't so much of an
issue. So we put it to a pole, and most people voted to move back to the
jQuery model.

With all of that said, it's been so long since I've spent any real time
reading the jQuery documentation that I'm worried I might be diverging
unintentionally. If you do catch things, please let me know!

In fact, if you're interested, you may be able to weigh in on issue #112.
I'm trying to figure out how to handle selectors passed to children(). I
think I broke QueryPath 3 when I changed the querying engine.

technosophos/querypath#112

Thanks!

Matt