Hy!

I've found a problem with Waiter's integration in Request. It doesn't care if the waiterTarget value set when it has an element to update.

So I changed the order of this.options.update || this.options.waiterTarget to the opposite, because in my opinion, the explicit option (waiterTarget) has to have more priority than defaulting to the updated element!

Here is a patch from what I exactly did :)
http://www.iwstudio.hu/~u-foka/WaiterTargetFix.patch

Difference in capital letters to the onError method between Asset.image Asset.images and documentation.

I have found that the Scroller class was not working properly.

scroll: function(){
  // ...
  pos = this.element != this.docBody ? this.element.getOffsets() : {x: 0, y:0},
  // ...
}

When instantiating class like this : new Scroller(window); ; the previous method calls getOffsets of window and it's not defined.

There is more than one way to fix this so I will let you decide.
Either you can implement getOffsets in window and document like this :

[Document, Window].invoke('implement', {
  // ...
  getOffsets: function(){
    return {x: 0, y: 0};
  },
  // ...
}

or you can fix the Scroller class the way you want.

It is your call, you probably know better than me.

Move me to 1000.

Thanks.

During the refactor of element.position for 1.3.2.1 the removal of null or undefined options was dropped.

This breaks ClientCide's StickyWin, so it could be that this is an issue that I should have dealt with there but since mootools is designed to be backwards compatible I posted the issue here first.

https://github.com/JLBShecky/mootools-more/commit/65a739a27d9c36dc5c5e9849f3688909b884dece

It should be '%d.%m.%Y', not a '%d/%m/%Y';

This line in both methods prevents animating if the current x,y are the same as the target x,y:

if (to.x != scroll.x || to.y != scroll.y) this.start(to.x, to.y);

However, as start() is never called, this also prevents the onComplete callback or .chain functions from firing.
(This is different to how toElement works, which will always fire the callback, even if the coordinates remain the same.)

As a user of these methods, I would expect the onComplete callback to always fire, even if no scrolling actually happens.

String sorting is not working properly for some languages, fe:
http://jsfiddle.net/jYsks/1/

In polish 'ń' goes after 'n', so it should be:

"koń", "kos", "kot".

Instead it is

"kos", "kot", "koń".

Best solution would be acceptance of additional function for parser object, then You could sort by any function user need (localized string sort, some special document number format sort, etc.).
https://github.com/testeria/mootools-more/blob/patch-1/Source/Interface/HtmlTable.Sort.js

The date validator doesn't work correctly in FF4 and IE9 (no others tested except Chrome, which works)

See:
http://jsfiddle.net/LgJsN/1809/

Entering 31/31/2009 results in a validated field, except that the value of the input field changes into "31/07/2011", which is obviously not the same as the (invalid) date input of "31/31/2009".

Also, when entering a valid date (dateFormat:'%d/%m/%Y') such as "31/02/1999" messes things up, as it changes into "02/07/2001"

'[email protected]' passes as valid email even as it is not and 'foo#[email protected]' does not validate even tough it is valid. I suggest the following change:

from
/^[A-Z0-9.%+-]+@[A-Z0-9.-]+.[A-Z]{2,4}$/i
to (some char escaping could go here):
/^[a-z0-9!#$%&'*+-/=?^`{|}~]{1,64}@([a-z0-9-]{1,63}.)+[a-z]{2,4}$/i

reference from http://en.wikipedia.org/wiki/Email_address#Syntax

There should be another method that "makes sure" there is no double dot (..) in the email.

Hi...

i've found that in the it-IT Locale there are some values missing:

week(s)Ago/Until, month(s)Ago/Until and year(s)Ago/Until

i don't know if also other languages have value missing...

cya

The email validator accept

toto@gmailcom

http://mootools.net/demos/?demo=Enhanced-Form

it's normal ?

Issue with using Sortables, if right clicking on an item in the list the list stops functioning maybe due to the check for event.rightClick in the Drag class.

can be fixed with doing a check for event.rightClick in the start function of Sortables class too

All the best

The last line of Tips.fireForParent(event, element) looks like this:

else this.fireForParent(parentNode, event);

Given the argument signature of the method, this seems like the arguments are given in the wrong order, and that the event should be given before parentNode. This leads to an error in event propagation to parent element, preventing tooltips on parent elements from triggering on mouse events on their children.

It will be great to make custom clone tag (now it is DIV) changeable by options.
In some cases it is useful to sort table rows. In that cases DIV element breaks table structure and it is needed to use TR instead
Thanks.

Browser.Language = (function(){
    var lang = navigator.language ? navigator.language : navigator.userLanguage;
    if(!lang.match(/\w+-\w+/))
        lang = lang + '-' + lang.toUpperCase();
    return lang;
})();

MooTools.lang.setLanguage(Browser.Language);

When using strict mode "use strict" Chrome 11 raises a JS error on Lang.js, line 89, langData is not defined. Should be defined as local variable.
https://github.com/mootools/mootools-more/blob/1.2x/Source/Core/Lang.js

Hi,

I positioned a div with draggable elements within a container of style overflow:auto.
The div with the draggable elements are in the overflow area (need to scroll down to see it)

In this case, the limitations (options.container) are wrongly calculated since the scroll offset
is ignored.

I fixed this issue with the following lines of code at the end of Drag.Move.calculateLimit:

if (element.getOffsetParent()) {
  top += element.getOffsetParent().getScroll().y;
  bottom += element.getOffsetParent().getScroll().y;            
  left += element.getOffsetParent().getScroll().x;
  right += element.getOffsetParent().getScroll().x;         
}
return {
   x: [left, right],
   y: [top, bottom]
};

Hi MooTools pals.

Once time ago I wrote a XML2Js converter, and some of you helped me (at Google Group) with a function I was requiring that time, which I believe should be part of MooTools More Object.Extras.

It is a setToPath function, which basically create object paths from strings and fill it with a value.

Next, it's code:

function setToPath (source, path, val) {
        var parts = key.split('.'),
            source2 = source; // so we can return the object
        for (var i = 0, l = parts.length; i < l; i++) {
            // So when the value does not exist (and is an own property) or is not an object
            if (i < (l - 1) && (!source.hasOwnProperty(parts[i]) || !Type.isObject(source[i]))){
                source[parts[i]] = {};
            }

            if (i == l - 1) source[parts[i]] = val;
            else source = source[parts[i]];
        }
        // Return the modified object
        return source2;
    }

.toggle(force) should exist! And maybe it will soon!

I posted this first in core by accident, should be in more: mootools/mootools-core#2065

i tried to create a div with an oversize content and apply the scroller class to it, but i will not work.

i used the example from here http://demos111.mootools.net/Scroller which is the same for 1.3
i use firefox, but its alos not working in chrome.

Hi, discussed this a bit on #mootools, but it's good to get some more feedback about whether this is expected behaviour or not, so here goes:

The "issue" occurs when creating a Drag instance and attaching a limit to it. If you then try to drag the element outside of it's limit, the element itsself will halt nicely, however, onDrag will keep firing as long as you hold the mouse button and move the mouse around. For me, expected behaviour would be that onDrag only fires if the actual element moves.

Discuss! :)

A formRequest element's method is described on on the page https://github.com/mootools/mootools-more/blob/master/Docs/Forms/Form.Request.md .

According to the source code, I think it should be formUpdate https://github.com/mootools/mootools-more/blob/master/Source/Forms/Form.Request.js#L191

The URI.js script has file encoding as ANSI, rather than "UTF-8".
Results in incorrect encoding saved when building scripts with build.rb.

Bad characters in author name should be fix, too.

Example:

var index = 0;

var myAccordion = new Fx.Accordion(togglers, elements, {
    onActive: function(toggler, element){
        alert(index===this.previous);
    }
});

myAccordion.display(index);  // return true (index = 0 and this.previous should be -1)
index++;
myAccordion.display(index); // return true (index = 1 and this.previous should be 0)

Suggestion #1:

Set "this.previous" before the last line of "display" method

display: function(index, useFx){
   ..
   ...
   this.previous = index;
   return useFx ? this.start(obj) : this.set(obj);
}

Suggestion #2:

Add "this.current" index value

initialize: function(){
  ..
  ...
  this.current = this.previous = -1;
  ..
},

display: function(index, useFx){
   ..
   ...
   this.current = index;
   this.selfHidden = false;
   ..
   ...
   this.previous = index;
   return useFx ? this.start(obj) : this.set(obj);
},

this was lighthouse ticket 537

The problem:
the function Element.getComputedSize doesn't return correct values of width and totalWidth.

Browsers affected:
Internet Explorer 7/8.

When this bug happen:
when the element has a relative width (ex. % or em).

Example:
http://jsfiddle.net/NKjoep/VvrAs/

Here also a JSSpec example:
http://dl.dropbox.com/u/1840263/getComputedSize/test1.html

So in these example, the Element.getSize is working well.
Why not to rely on it? Element.getComputedSize shouldnt get the actual real size
and then add the optional parameters (such as padding, borders...)?

on the email error message there's a typo:

Por favor, introduce una direccione de email valida. Por ejemplo, "[email protected]

It should be:

Por favor, introduce una direccion de email valida. Por ejemplo, "[email protected]

Also I've noticed that there's no acute characters where they should be:

direccion

should be:

direcci&oacute ;n (there's a space there because this editor parses the acute code)

There are a lot of other ones

validate-no-space should be changed to validate-nospace or rather validate-nospace updated in the more to be validate-no-space.

Cheers

Hi,

Please have a look a this and see if it good enough to include in Mootools more, it started as a suggestion to extend String.substitue to become a class to deal with templates :)

http://allochi.wordpress.com/2011/08/19/mootools-html-templates/

If Mootools team likes it, I can spend more time on it (which I will anyway) and send you the code, otherwise I can add it to forge, in anyway it's part of my HTML5 editor project that I'm working on now, and I would be happy if it see its way to in Mootools-More.

Thanks for Mootools!

I have a certain Keyboard usage case that is not supported by the current implementation of Keyboard. I've found a workaround, but I want to see if this feature is feasible for a future release of MooTools More.

I have a UI control (X) that has its own keyboard. One of the events (X) handles is 'esc', which tells (X) to close.

Inside of (X) is another control (Y), which is held by but not coupled to (X). (Y) knows nothing of (X) or (X)'s Keyboard. (Y) does not care about the 'esc' key. However, even when (Y) has focus, the user is going to expect 'esc' to close (X).

I have implemented a KeyboardStack, such that when (Y) pushes its Keyboard onto the stack, (X)'s Keyboard (already on the stack) becomes the manager of (Y)'s Keyboard. This allows me to simply dependency-inject a KeyboardStack into (X) and (Y), without either control being aware of anything beyond the KeyboardStack. Very minimal, generic coupling. However, while this makes it easy to push and pop a Keyboard hierarchy, it doesn't give me the ability to send key events up a chain of command.

It makes no sense for me to couple (Y) directly to (X) just to tell (X) that (Y) received an 'esc' event. What I really want is for all key events NOT handled by (Y)'s Keyboard to merely propagate up to (X)'s Keyboard.

Current propagation model: Keyboard.manager > deepest Keyboard > window

The behavior I desire: deepest Keyboard > shallower Keyboard > Keyboard.manager > window.

I need to have an event to run up the Keyboard chain until one of the Keyboard instances stops that event.

To be less vague, I have a dialog window, and inside that dialog window is an auto-complete element. The auto-complete element needs to watch for 'up', 'down', and 'enter'. However, 'esc' should always close the dialog window, even if the auto-complete element has focus. Forcing the user to blur that element (which deactivates its Keyboard instance) before 'esc' will close the window is confusing to the user. Furthermore, adding logic to the auto-complete element that teaches it how to look for a parent dialog window and close that window on 'esc' makes no sense.

Is there any existing plan in More's development path that would enable some Keyboard instances use this "Propagate unhandled events upward" model, while making sure Keyboard instances still use the current model by default? Does anyone else like this feature idea?

Without that feature, my current workaround is:

myKeyboard._manager.fireEvent('keyup:keys(esc)');

This has obvious limitations and is really hacky.

SmokeyXL brought this up during the office hour:

http://jsfiddle.net/pxLBm/

Note that as long as snap is true, you can't ever get to the final value, if snap is false, you can.

Changing the steps from 5 to 6 seems like it should fix it, but then things get even stranger.

Cannot be "just a string" as the docs lead you to believe. Must be an Array.

There is a small error in this doc page http://mootools.net/docs/more/Forms/Form.Validator#Validators:validate-currency-dollar

It seems that the example about validate-currency-dollar breaks the parser that generates the doc page.

setting errorPrefix: 'Fejl: ', inside MooTools.lang.set('da-DK', 'FormValidator', {
does not work.

Instead of showing the language specific text, it still shows 'Error: '.

All the other language specific messages are correctly shown.

Taken from the Formvalidator class:
errorPrefix: function(){
return FormValidator.getMsg('errorPrefix') || 'Error: ';
}

and from Formvalidator.Inline class:

this.errorPrefix = $lambda(this.options.errorPrefix)();

A Drag.Move with a container, inside a Drag.Move with a border, will not have correct limits.
I think the error is related to the way the parent element might be repositioned upon initialization, but I'm not sure at all.
An example can be found here: http://mootools.net/shell/Ddcy4/

Sometimes stripTags mixes some tags, for example <b> and <br/>

The following example shows the problem better:
http://jsfiddle.net/kallelat/Taa67/1/

Happens in MT 1.2+

Adding the validate-email validator does not work as expected, as "y@z" seems to be a valid emailaddres according to this validator.

The Mask size has a problem when the scroll bar is present.
You can scroll down and see that there is no mask. It's a very simple change on the code of mask.js on line:

    if (this.target == document.body) {
        var win = window.getSize();

for this:

    if (this.target == document.body) {
        var win = window.getScrollSize();

Destroying cloned element is chained to the revert effect but reset() is called immediately (is not chained). It causes complete event to be fired before cloned element is removed and this results in duplicity of the element when serialize() is called with onComplete event.

'offsetParent' - is null or not an object in Drag.Move.calculateLimit()

I need to use a slider without step, but it has a strange behavior when I click on the bar (it goes back to start).

How to reproduce :
new Slider('bar', 'knob', { steps: false });
Drag the knob, then click on the bar.

How to fix it :
In Slider.js :

70a71
>               this.steps = this.options.steps || this.full;
72c73
<               this.max = $chk(this.options.range[1]) ? this.options.range[1] : this.options.steps;

---
>               this.max = $chk(this.options.range[1]) ? this.options.range[1] : this.steps;
74d74
<               this.steps = this.options.steps || this.full;

Bye

such as
https://github.com/cheeaun/mooeditable/blob/master/Demos/MooEditable/MooEditable.CleanPaste.html

works in 1.3.0.1, but can't work with 1.3.1.1
Because of refactor

console.log('relay click'); never gets triggered if i rollover my green box. i think it's a bug. relay should be listening for elem.fireEvent();
here is a testcase, you obviously need a mootools-more version with relay support.

<!DOCTYPE html>
<html>
    <head>
        <meta charset="utf-8">
        <title>test</title>
    </head>
    <body>
        <div id="box" style="width: 200px; height:200px; background: green"></div>
        <a href="http://google.de" id="plus">+</span>
        <script src="http://ajax.googleapis.com/ajax/libs/mootools/1.3/mootools-yui-compressed.js"></script> 
        <script src="mootools-more.js"></script> 
        <script>
            $('box').addEvent('mouseenter', function(e){
                $('plus').fireEvent('click', [e]);
            });
            document.body.addEvent('click:relay(a)', function(e){
                console.log('relay click');
                e.preventDefault();
            });
            $$('a').addEvent('click', function(e){
                console.log('fixed click');
                e.preventDefault();
            });
        </script>
    </body>
</html>

Seems all theday and month names are all lower-case, but the docs and other lang packages are sentence-case.

If we set opacity of the placeholding element to zero in the options, it should set the opacity of all the child elements to zero as well, at least in Chrome it keeps showing the child elements of the placeholder for some reason.

1. Review all of the things
2. Propose moving some stuff into other MooTools Milk packages
3. Propose splitting off groups of related classes into new Milk packages

Be sure to take into account who tends to work on which things and don't make it unnecessarily difficult to maintain.

I think the following should be supported:

uri.setData(uri.getData()).toString() == uri.toString() 

and this is not the case now: http://jsfiddle.net/J7tMj/

cat=1&cat=2 are different GET parameters than cat[0]=1&cat[1]=2

See also #1027.

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

Test case: http://mootools.net/shell/7FpZe/1/

Hovering back and forth between the menu items will eventually make a drop down freeze. Safari will output a "RangeError: Maximum call stack size exceeded.", while Firebug will show nothing(?).

I have tried running the same code with e.g. Fx.Slide, and it doesn't create any errors.

Now it returns a string for single GET parameter and Array for multiple GET parameters with the same name. Because of this code that uses getData have to do typeOf('array') checks every time. I think this is easy to miss and the existing interface may cause many subtle bugs in the wild.

My proposal:

• always return Arrays in getData because correct code for the existing behavior have to check typeOf('array') anyway and this check will be true so returning Array is almost backwards-compatible for the correct existing code;
• add 'getDataObject' (not sure about the name) method that will always return single values.

setData may stay untouched and accept both variants.

File: mootools-more / Source / Fx / Fx.Reveal.js
Line: 53

Expected:
this.element.setStyle('display', this.options.display);

Actual:
this.element.setStyle('display', 'block');