anutron / clientcide

'domready' happend after window.onload always in ie6/ie7/ie8 !!!!!!!!!!!!!!!!!!!!!!!!!!!!! i don't know if you test it~shit~~~ ~~!!!! I am not the first one know it ,there are many peaple told you ,shit team~~ 不知道到底有没测试过就发布了~~这么严重的bug还敢发布，这不是害人吗

Several links in the documentation's navigation are broken.

• http://clientcide.com/docs/UI/StickyWin.UI
• http://clientcide.com/docs/UI/StickyWin.UI.Pointy

The correct links are:

• http://clientcide.com/docs/UI/StickyWin.ui
• http://clientcide.com/docs/UI/StickyWin.ui.pointy

First ist that on the StickyWin destroy() handler, the mask is missing

class StickyWin

destroy: function(){

    if (this.windowManager) this.windowManager.remove(this);
    if (this.win) this.win.destroy();
    if (this.options.useIframeShim && this.shim) this.shim.destroy();
    if (document.id('modalOverlay')) document.id('modalOverlay').destroy();     

// add this

            if (this.mask) this.mask.destroy();

    this.fireEvent('destroy');
}

I have two dialog: login and register. Each has a link to the
other. I never want them "stacked" or open at the same time: when I
open one, if another is open, I want it closed upon opening another.
In clientcide 2.1.0, I was able to get the desired effect by setting
id: to the same value. This no longer works and I've found 2.2.0 has
the allowMultiple option. So I set it to false. This mostly works
but has a bug in that as each window opens, the mask get applied over
the prior, so as I switch back and forth between login and register
multiple times, my background gets darker and darker. When I finally
close the dialog, I have to click the background once for each time a
dialog was opened to incrementally remove the masks.

Hi, I implemented the one-liner btwn :
new Tips.Pinty($('')); but the Tips Pointy feature doesn't work in any browser..
IE debugger says it's expecting a ')' in clientcide's function which contains:
var SWA = function(extend) {

 return Class.refactor(extend, {

which was true, and so I added that ')' but then the debugger points to another bug in mootools's code. So I am wondering if someone else also has the same experience, or maybe someone can tell me if I implemented Tips Pointy wrong.

Aaron pulled the latest code from the clientcide site today and had to fix the following issue, you have a trailing , on buttonsSelector: ".button", this causes an error on IE7.

var CNETcarousel = new Class({

Extends: SimpleCarousel,
options:{
    slidesSelector: ".slide",
    buttonsSelector: ".button",
},

There was also an error in this block of code which caused problems across browsers I needed to add in an additional closing bracket.

(function(){

var SWA = function(extend) {
    return Class.refactor(extend, {
        options: {
            XHRoptions: {
                method: 'get'
            },
            wrapWithStickyWinDefaultHTML: false, 
            stickyWinHTMLOptions:{}
        },
        initialize: function(options) {
            this.setOptions(options);
            this.options.RequestOptions = this.options.XHRoptions;
            this.wrapWithUi = this.options.wrapWithStickyWinDefaultHTML;
            this.uiOptions = this.options.stickyWinHTMLOptions;
            this.previous(options);
        }
    })
};
try {   SWA(StickyWin.Ajax); } catch(e){}
try {   SWA(StickyWin.Fx.Ajax); } catch(e){}
try {   SWA(StickyWin.Modal.Ajax); } catch(e){}
try {   SWA(StickyWin.Fx.Modal.Ajax); } catch(e){}

})();

Fantastic project guys and anything else I spot you can be sure I'll feedback!

Dave

One symptom of this bug was an infinite loop of resize calls (Modalizer.resize()) in IE7--initiated by the 'resize' event listener on the Modalizer. The workaround I found was to modify the StickyWin.destroy() function to first call this.modalHide() before issuing the destroy method. I'm not sure if this will work in general, but for my application it did.

The Autocompleter.JSONP uses Request.JSONP to request data. Each request on that class sends a callback key parameter in order to handle the response data. The value associated with that key is 'Request.JSONP.request_map.request_'+ index

Index is the number of times Request.JSONP was called on that page.
Since that index's always growing, the request url changes all the time and the caching process gets broken.

A simple scenario would be like this:

The user types "auto" on the autocomplete field. A request is made to:
url + "?callback=Request.JSONP.request_map.request_0&query=auto"

Now the user continues typing, "autocomple". As expected, a new request is made to:
url + "?callback=Request.JSONP.request_map.request_1&query=autocomple"

And now the user backspaces until the field reads "auto" again. At that time, a new request is made to:
url + "?callback=Request.JSONP.request_map.request_2&query=auto"

The result works as expected but, since the url changed (index was zero and now is two) the browser can't possibly cache these results.

Considering autocomplete generates a reasonable ammount of traffic, it woud be appropriate to find a way to benefit from caching.

As a proof of concept, I added:

Request.JSONP.counter--;

to the Autocompleter.JSONP.queryResponse method. This way the index is always zero and the requests get properly cached. Of course this isnt' reasonable as a solution, but it proves the point: Autocompleter.JSONP could have better caching.

I am using clientcide-trunk-2.1.0-compatible.compressed.js. When implementing FormValidator.Inline, the ignoreHidden property is not ignoring hidden fields, and therefore my form never validates.

If you add anchors to your lightbox with hrefs that contain spaces, they don't display correctly in FF3.

Element.setStyle('background-image', 'url(' + foo ')');

doesn't seem to work if foo contains spaces. I changed a line

        nextEffect: function(){
                switch (this.step++){
                case 1:
-           this.image.setStyle('backgroundImage', 'url('+this.images[this.activeImage][0]+')');
+           this.image.setStyle('backgroundImage', 'url('+escape(this.images[this.activeImage][0])+')');

And now it's working for me, I've only tested it in a few browsers but it works in at least FF3 and IE7.

Browsers: All

Issue: Close button should be in upper right of StickyWin instead of upper right of screen.

Reason: Close button is position:absolute, but the parent has no positioning.

Demo: http://www.clientcide.com/docs/UI/StickyWin.UI (pick a demo, look for the close button).

Notes: When used with modal this problem does not show, as the outer parent is positioned.

Fix: Pull the corrected code:
git://github.com/SamGoody/clientcide.git (1048db)

If I use this format in input value, 22/09/2009 (22 September) becomes 02/10/2010 (2 October).
Example:

HTML:
JS: new DatePicker($('calen'),{

        'format': '%d/%m/%Y'
    });

This was a pig to find and fix.

I have three tabs, first has a input box on it. Switch between the tabs and the text box disappears. Its not a hasLayout issue as its an input box which should have it anyway. After much hackery I found that if you set smoothSize to false it works a treat.

Will try to knock up a test case for it if I get time.

hey Aaron!

just a quick question about the MooScroller (cnet 1.2)...
I can't seem to make the example work somehow...
I just copied all the code from the example.

I use cnet 1.2 and the newest moo-core (1.2.3) and moo-more (1.2.3.1) but whatever I try, Firebug always says >>
"document.id(content) is null"
in this MooScroller-line >>
this.content = document.id(content).setStyle('overflow', 'hidden');

I included the "$E" and "MooScroller", but I also tried the complete cnet lib.

Thanks for your help & thumbs up on your great work on moo!

yougene

Internet Explorer needs a ; after window.execScript(text) in $exec function on Clientcide 2.1.0 compressed.

I've had problems with positioning the autocompleter. The relative-option doesn't seem to work as expected. This line of code causes this:

if (this.options.relative || this.element.getOffsetParent() != document.body) {

I think this should be, so the option won't be overridden:

if (this.options.relative && this.element.getOffsetParent() != document.body) {

With the latest Mootools core and more (1.2.3 and 1.2.3.1 respectively) along with clientcide 2.1.0, if you create a simplecarousel without any buttons (don't pass any existing elements as buttons) then IE throws an error message when trying to use document.id() to get the button to change its state to active.

I found that Autocompleter's scrolling for overflow is bugged. It obviously isn't on the demo page, but in the trunk available at the download page it is. This is a small patch I made from the fix that helped me:

Context:

    if (this.overflown) {
        var coords = this.selected.getCoordinates(this.choices), margin = this.options.overflowMargin,
            top = this.choices.scrollTop, height = this.choices.offsetHeight, bottom = top + height;

Old:

        if (coords.top - margin < top && top) this.choices.scrollTop = Math.max(coords.top - margin, 0);
        else if (coords.bottom + margin > bottom) this.choices.scrollTop = Math.min(coords.bottom - height + margin, bottom);

New:

        if (coords.top - margin < 0 && top) 
            this.choices.scrollTop = Math.max(coords.top - margin + top, 0);
        else if (coords.bottom + margin + top > bottom) 
            this.choices.scrollTop = Math.min(coords.bottom - height + margin + top, bottom);

Context:

    }
    if (this.selectMode) this.setSelection();
},

Found this one in our code today when a client booted up Chrome to show off the site. Lots of egg on face. Anyway, seems Chrome throws a wobbler on any dbug statements.

Sample code to run in Chrome:

<head>
    <script type="text/javascript" src="mootools-1.2.3-core-nc.js"></script>
    <script type="text/javascript" src="mootools-1.2.3.1-more.js"></script>
    <script type="text/javascript" src="clientcide.js"></script>
    <script type="text/javascript">
        dbug.debug( "Hello I'm working" );   // Chrome will goose right here

        var obj = {
            name: "test",
            data: []
        }

        dbug.debug( obj );

        dbug.debug( "contents of obj are: " + obj );
    </script>
</head>
<body>

</body>

Not had time to look into it much yet as the debugger in Chrome is horrid but here for reference so its not forgotten. For now I've just commented out all our dbug calls which is a bummer as thats the whole point of using dbug instead of console :(

There is no effect the first time (and only the first time) you click next/prev on the SimpleSlideShow.Carousel() in IE. The same goes for SimpleImageSlideShow.Carousel().

The issue is found here: http://www.clientcide.com/wiki/cnet-libraries/08-layout/04-simpleslideshow
I'm running IE8 on Windows 7.

Hi Aaron
Not sure if this is a bug or not but I think it is atleast an issue.
I decided to have a play around with the PopupDetails but found that I was constantly getting an error.
opt.stickyWinToUse is not a constructor
http://www.buyaphoto.net/test2/js/clientcide-trunk-2.1.0.js
Line 10142
I even tried copying and pasting the example from here http://www.clientcide.com/docs/UI/PopupDetails#PopupDetail
Still the same error.
Regards
Dean

unlike the Fupate.Prompt class, the Fupdate.AjaxPrompt doesn't support the "caption" option. the StickyWin caption is hard coded in the ajaxPrompter function. It should look likes this maybe:

            makeStickyWin: function(formUrl){
                if (this.swin) return this.swin;
                this.swin = new this.options.stickyWinToUse($merge({
                    showNow: false,
                    requestOptions: this.options.requestOptions,
                    onHide: function(){
                        this.win.empty();
                    },
                    url: formUrl,
                    handleResponse: function(response) {
                        var responseScript = "";
                        this.swin.Request.response.text.stripScripts(function(script){  responseScript += script; });
                        var content = this.options.useUi?StickyWin.ui(this.options.caption, response, this.options.stickyWinUiOptions):response;
                        this.swin.setContent(content);
                        if (this.options.requestOptions.evalScripts) $exec(responseScript);
                        this.element = this.swin.win.getElement('form');
                        this.initAfterUpdate();
                        this.swin.show();
                    }.bind(this)
                }, this.options.stickyWinOptions));
                return this.swin;
            },

Hi
Just take opera, go here :
http://www.clientcide.com/wiki/cnet-libraries/11-3rdparty/02-autocompleter.local
Try to type something.
I got this error :

Error:
name: TypeError
message: RegExp.prototype.test was called with the wrong number of arguments
stacktrace: n/a; see 'opera:config#UserPrefs|Exceptions Have Stacktrace'

Well i have look around for this bug, and it seems that the author has fix it, but we don't have the last version on clientcide, so you should just update it :)
http://digitarald.de/project/autocompleter/

What steps will reproduce the problem?
1. We can see the other slides, they are not passed on display:none

1. The bug is also here on your demo on
   http://www.clientcide.com/wiki/cnet-libraries/08-layout/04-simpleslideshow

2. It happen on IE7, works fine on firefox

What is the expected output? What do you see instead?

I tried to add some lines on the simpleslideshow.carroussel class near
those lines but can't fix it :(

this.element = new Element('div').wraps(this.container).setStyles({

        width: this.container.getSize().x,
        overflow: 'hidden',
        float: 'left'
    });
    this.container.setStyles({
        width: this.options.sliderWidth,
        position: 'relative',
        overflow: 'hidden'
    });

Thx for answer :)

js compression (with javascriptpacker) fails:
in: StickyWin.js
in: initialize: function(options){...
on line: if (this.options.destroyOnClose) this.addEvent('close', this.destroy)

missing a semicolon at the end of the line

I must be doing something wrong, but I have tried the demo and the code in the docs and neither one work. I have spent all day looking at the code (including the source from the demo page) and have not had any success. I totally love Moo Tools and I want to love the Clientcide library too, but please please please help me get there.

Sincerely
Gra

Inside FormValidator.Inline:

BAD:

validate: function(field, force){
...
if (!fx && window.Fx && Fx.Scroll){
    fx = new Fx.Scroll(par, {
        transition: 'quad:out',
        offset: {
            y: -20
        }
    });
    par.store('fvScroller', fx);
}
...

SHOULD BE:

validate: function(field, force){
...
if (!fx && window.Fx && Fx.Scroll){
  fx = new Fx.Scroll(par, this.options.scrollFxOptions);
  par.store('fvScroller', fx);
}
...

The "compression" version of 2.0.2 has an error. The "no compression" works well.

I'm not sure if its a code problem, but after updating both Clientcide and Firefox to latest it does not appear. Ive checked the demo site and it works there.

Also if I switch tabs, and go back it then appears, which would seem a FF issue, but as I sayd before, the demo code works.

Test site: http://www.arrarray.com/contacto.html

http://code.google.com/p/cnetjavascript/issues/detail?id=113

see: http://code.google.com/p/cnetjavascript/issues/detail?id=111

see: http://code.google.com/p/cnetjavascript/issues/detail?id=109

see: http://code.google.com/p/cnetjavascript/issues/detail?id=107

see: http://code.google.com/p/cnetjavascript/issues/detail?id=106

see: http://code.google.com/p/cnetjavascript/issues/detail?id=105

see: http://code.google.com/p/cnetjavascript/issues/detail?id=101

see: http://code.google.com/p/cnetjavascript/issues/detail?id=98

see: http://code.google.com/p/cnetjavascript/issues/detail?id=97

see: http://code.google.com/p/cnetjavascript/issues/detail?id=92

see: http://code.google.com/p/cnetjavascript/issues/detail?id=73

see: http://code.google.com/p/cnetjavascript/issues/detail?id=67