Posts tagged with “tips”

June 04, 09

Useful OS X terminal shortcuts

Create a file named .inputrc in your home directory and paste the following code:

# this makes the "delete" key work rather than 
# just entering a ~
"\e[3~": delete-char

# these allow you to use ctrl+left/right arrow keys
# to jump the cursor over words
"\e[5C": forward-word
"\e[5D": backward-word

# these allow you to start typing a command and 
# use the up/down arrow to auto complete from 
# commands in your history
"\e[B": history-search-forward
"\e[A": history-search-backward

# this lets you hit tab to auto-complete a file or
# directory name ignoring case
set completion-ignore-case On

# Two escapes clear command line
"\e\e": "\C-a\C-k"

It will add a few useful keyboard commands.

01:20 PM | | 1 Comment | Tags: ,
June 03, 09

Assassinating jumpy htmlText hyperlinks

UPDATE: Read this great tutorial on styling the HTML text field

I read quite recently that people are still swearing Flash jumpy hyperlinks in htmlText with Anti-aliased for readability option turned on. What are they talking about? – here's the example.

Roll over and out the links in the left column and carefully observe what happens to the text. Pretty bad stuff.

Fortunately, there is a perfect workaround discovered by Jonnie Hallman aka Destroy Today, but not everybody knows about it yet. So, (to save the world), I wrote a simple TextField wrapper class that automatically deals with the issue, additionally it also fixes another annoying bug – selectable code>TextField scrolls one line on text select – when you click and drag down the mouse while selecting the text, the content scrolls vertically! (and it shouldn't as the text height is the same as the box height, so there's nothing to scroll)

The fixed TextField is presented in the right column – again, try rollover the links.

How does the fix work then?
Basically, after setting the htmlText property, store the height of the TextField, then set autoSize to NONE. That will lock it, so no no more jumping is happening. It is all done at one go by overriding htmlText setter method.

override public function set htmlText(value:String):void
{
    autoSize = TextFieldAutoSize.LEFT;
    super.htmlText = value;
    recordedHeight = height;
    autoSize = TextFieldAutoSize.NONE;
    // adding extra height will prevent "vertical scroll on text select" bug
    // the extra height value should be just bigger than the default leading
    height = recordedHeight + getTextFormat().leading + 1;
}

Additionally, we're adding few pixels extra the TextField default leading value plus 1 to the height there (discovered by Luke Sturgeon), it will prevent "vertical scroll on text select" bug described above.

How should I apply this fix to my broken TextField?
By creating a FixedTextField instance instead, exactly the same way as you would instantiate an ordinary TextField:

var textTF:FixedTextField = new FixedTextField();

Alternatively, you could clone an existing TextField by referencing to it:

var textFieldFixed:FixedTextField = new FixedTextField(someOtherTextFieldToClonePropertiesFrom);

In fact, you can even attach the TextField stored in the Flash IDE library (obviously wrapped up in a MovieClip symbol) – I wrote a separate blog post on that matter.

Download zipped demo including the FixedTextField class.

01:30 PM | | 22 Comments | Tags: , , , , ,

A TextField called Dolly

I love Flash IDE. Especially the way you can stylise the TextField – decide what glyphs to embed and what fonts to include, set line height (leading), kerning (letterSpacing), colours, etc. I really hate creating new TextFields from scratch, entirely in code.

Using Library to store a TextField
Unfortunately, we can't store a TextField directly as a Flash IDE library symbol, so we have to wrap it up in a MovieClip and reference to it as follows:

var textFieldNormal:TextField = new TextMC().textTF;

where TextMC is the library MovieClip symbol containing a TextField named textTF.

How to duplicate (clone) a TextField?
You can't do it natively in AS3. But you can do some magic instead – clone all (excluding read-only) properties of another TextField.

var description:XML = describeType(referenceTextField);

for each (var item:XML in description.accessor)  
{
    // clone passed textfield properties that are not read only  
    if (item.@access != "readonly") targetTextField[item.@name] = referenceTextField[item.@name];  
}

targetTextField.defaultTextFormat = referenceTextField.getTextFormat();  

TextFormat of the referenceTextField won't be cloned with the loop above as it's not accessible via getter/setter/public methods. You have to clone it manually and this is what the last line does. It's very handy when you want to change the default kerning (letterSpacing) in Flash IDE and you want it to be preserved in your cloned text field.

I wrote a FixedTextField class that does that (plus fixes a couple of other TextField problems) – download zipped demo including the class.

01:00 PM | | 0 Comments | Tags: , , , , ,
May 12, 09

Finding the missing child

How many times you've encountered that annoying error ArgumentError: Error #2025: The supplied DisplayObject must be a child of the caller message?

It often happens when you're trying to remove a non-existing child which you would swear you had put there. How to find what depth is resides at, quickly? A simple "trace display list" package-level function closure traceDL(displayListElement, depth | "filter string") to the rescue!

package
{
    import flash.display.DisplayObjectContainer;
    import flash.display.DisplayObject;

    public function traceDL(container:DisplayObjectContainer, options:* = undefined, indentString:String = "", depth:int = 0, childAt:int = 0):void
    {
        if (typeof options == "undefined") options = Number.POSITIVE_INFINITY;

        if (depth > options) return;

        const INDENT:String = "   ";
        var i:int = container.numChildren;

        while (i--)
        {
            var child:DisplayObject = container.getChildAt(i);
            var output:String = indentString + (childAt++) + ": " + child.name + " ➔ " + child;

            // debug alpha/visible properties
            output += "\t\talpha: " + child.alpha.toFixed(2) + "/" + child.visible;

            // debug x and y position
            output += ", @: (" + child.x + ", " + child.y + ")";

            // debug transform properties
            output += ", w: " + child.width + "px (" + child.scaleX.toFixed(2) + ")";
            output += ", h: " + child.height + "px (" + child.scaleY.toFixed(2) + ")"; 
            output += ", r: " + child.rotation.toFixed(1) + "°";

            if (typeof options == "number") trace(output);
                else if (typeof options == "string" && output.match(new RegExp(options, "gi")).length != 0)
                {
                    trace(output, "in", container.name, "➔", container);
                }

            if (child is DisplayObjectContainer) traceDL(DisplayObjectContainer(child), options, indentString + INDENT, depth + 1);
        }
    }
}

So, where is my child?
With traceDL(someDisplayObject) you can reveal the whole display list of any DisplayObject. It will traverse its display list and show you all the instance names and types of the children (with depths they reside on), so it's easy to find out which depth your object hides. Some additional parameters like alpha, visibility, x and y coordinates, width and height (with scaling), and rotation will be displayed as well to help debugging (if it's too much, you can comment some of them out in the class file).

It's also possible to cut the traversing depth by specifying a number of levels to show: traceDL(someDisplayObject, 0) – will only show the first children, traceDL(someDisplayObject, 1) – will show you first children of someDisplayObject and also first children of its children.

Too many children shows up, it's like needle in the haystack!
No worries, just use traceDL(someDisplayObject, "filter string") where filter string is anything you think your child may contain, either it's an object's name, class (like MovieClip) or even some parameters (x: 100). Just anything that you think will match your child characteristics (that is what's outputted when you won't pass a search string).

Remember that if you hadn't explicitly set a name for a DisplayObject (either in code or Flash IDE), Flash player by default will call it instance and will assign a sequential number to it. Giving your object a name will always make it easier to identify.

I want to add a display object at a certain depth to a Library Symbol, but don't know the depth...
Nothing more perfect! With traceDL() it's childishly easy to trace any Library Symbol structure and determine what depth to use with addChildAt to dynamically add a DisplayObject to the Symbol.

Everything has slowed down!
The traceDL()function is recursive (unless you limit the depth it will dig till the last display node is scanned), and will use up a lot of your processor time, so take extra care and always remember to disable it in production code.

Erm, any demo/example usage?
Sure, there.

var rectangle:MovieClip = new MovieClip();
rectangle.name = "table";

var circle:Shape = new Shape();
circle.x = 100;
circle.y = 200;
circle.rotation = 45;
circle.scaleY = 0.5;

var square:Sprite = new Sprite();
square.name = "square";
square.rotation = 45;
square.scaleX = 0.5;
square.scaleY = 2;

var someSprite:Sprite = new Sprite();

addChild(rectangle);
rectangle.addChild(circle);
rectangle.addChild(square);
addChild(someSprite);

traceDL(this);
//0: instance4 ➔ [object Sprite]        alpha: 1.00/true, @: (0, 0), w: 0px (1.00), h: 0px (1.00), r: 0.0°
//1: table ➔ [object MovieClip]     alpha: 1.00/true, @: (0, 0), w: 0px (1.00), h: 0px (1.00), r: 0.0°
//   0: square ➔ [object Sprite]        alpha: 1.00/true, @: (0, 0), w: 0px (0.50), h: 0px (2.00), r: 45.0°
//   1: instance2 ➔ [object Shape]      alpha: 1.00/true, @: (100, 200), w: 0px (1.00), h: 0px (0.50), r: 45.0°

traceDL(this, "square");
//   0: square ➔ [object Sprite]        alpha: 1.00/true, @: (0, 0), w: 0px (0.50), h: 0px (2.00), r: 45.0° in table ➔ [object MovieClip]

traceDL(this, 0);
//0: instance4 ➔ [object Sprite]        alpha: 1.00/true, @: (0, 0), w: 0px (1.00), h: 0px (1.00), r: 0.0°
//1: table ➔ [object MovieClip]     alpha: 1.00/true, @: (0, 0), w: 0px (1.00), h: 0px (1.00), r: 0.0°

traceDL(rectangle, 1);
//0: square ➔ [object Sprite]       alpha: 1.00/true, @: (0, 0), w: 0px (0.50), h: 0px (2.00), r: 45.0°
//1: instance2 ➔ [object Shape]     alpha: 1.00/true, @: (100, 200), w: 0px (1.00), h: 0px (0.50), r: 45.0°

traceDL(this, "clip");
//1: table ➔ [object MovieClip]     alpha: 1.00/true, @: (0, 0), w: 0px (1.00), h: 0px (1.00), r: 0.0° in root1 ➔ [object MainTimeline]

traceDL(this, "shape");
//   1: instance2 ➔ [object Shape]      alpha: 1.00/true, @: (100, 200), w: 0px (1.00), h: 0px (0.50), r: 45.0° in table ➔ [object MovieClip]

If you were comfortable with AS2 "any depth" before and now feel a bit lost in AS3 depth management, traceDL() will definitely help you to switch.

You got me, I want it now!
Download the class traceDL.as

12:11 AM | | 10 Comments | Tags: , , ,
March 12, 09

Managing Flash (and Adobe CSx) file versions without changing the extension

When Flash CS4 came out we got yet another FLA file format. All sweet, it has the same legendary FLA file extension but now it can only be opened with Flash CS4 IDE. What's wrong with it? Imagine the following scenario: if you bought Flash CS4 (or decided to use trial version) and played about with your CS3 files there, you could accidentally save over them with CS4. When the trial expires, you're no longer able to access your files (well, you could revert them if you keep everything under version control but you'd still lose your changes). Or you might just have downloaded some cool source FLA files from a Flash blogger and doh! – they won't open with your CS3 although they are not using CS4 features at all (I had that problem with BIT-101's FiTC source files before having CS4).

The easiest fix was found by Senocular, let me quote Trevor:

All this takes is 3 additional key strokes for the first time you save a source file. No Fireworks engineering effort is required, it works for all versions of Fireworks, and Fireworks source files can continue to be easily recognized by other applications – and even, for that matter, humans. Using a .fw.png extension clearly indicates (to developers) that the PNG file is a Fireworks source file.

So next time you save the file, please do append 4 extra characters after the filename and before the actual file extension, e.g., if you save from Flash CS4 use filename.cs4.fla – in fact, you could apply this rule to all (almost) Adobe CS products, i.e., when saving from Photoshop, use filename.cs3.psd, etc.

I hope that this naming convention will get somehow included in Adobe CS5 (so opening the .cs4.fla file and re-saving it will still at least auto-rename it to .cs5.fla), so everybody will be able to easily distinguish between different file versions coming from the same named app.

The published SWF will follow the same convention: filename.cs4.swf and I think that makes sense as well, it may give you a clue what version of Flash Player the build is targeted for (unless you change the Publish Settings).

11:45 PM | | 0 Comments | Tags: , ,
← Previous Page 2 of 2