However a colleague of mine suggested that it may be due to the cost of pass-by-value as opposed to pass-by reference.

So I wanted to sort it out and this simple test is the result. After all, the first rule of profiling is “measure, measure, measure” (OK, the first three rules). The following code compares string comparisons as pass-by-value, pass-by-reference and inline, as well as an integer (constant) comparison.

The profiling code:

        function compare_value($str)
        {
            if ($str == 'MekTek')
                ;
        }

        function compare_ref(&$str)
        {
            if ($str == 'MekTek')
                ;
        }

        $string = "MekTek";
        $integer = 0;

        $now = microtime(true);
        for ($i = 0; $i < 10000000; $i++)
            compare_value($string);
        echo("value: " . (microtime(true) - $now) . "\n");

        $now = microtime(true);
        for ($i = 0; $i < 10000000; $i++)
            compare_ref($string);
        echo("ref: " . (microtime(true) - $now) . "\n");

        $now = microtime(true);
        for ($i = 0; $i < 10000000; $i++)
            if ($str == 'MekTek')
                ;
        echo("str: " . (microtime(true) - $now) . "\n");

        $now = microtime(true);
        for ($i = 0; $i < 10000000; $i++)
            if ($integer == 0)
                ;
        echo("int: " . (microtime(true) - $now) . "\n");

The result:

    value: 5.2698040008545
      ref: 5.4901170730591
      str: 5.3154430389404
      int: 1.0382018089294

Isn’t that fascinating. I love it when profiling returns obvious results. The time taken to compare a string depends on the length of the string, regardless of other factors.

The message is clear. Do not compare strings in performance-critical code. Use constants instead.

Imperial Galaxy, the first MMORG for Facebook

metarand » Creative Enclave announces Imperial Galaxy: the social networking MMOG

I’ll be sending invitations to all my friends as soon as the Beta goes live, and it will be in the Applications directory and so on.

We’re very excited about IMPERIAL GALAXY. The background of the game is based on Garth Nix’s forthcoming novel A CONFUSION OF PRINCES (due out in 2009) and we’ve been playtesting it for the last three months. The game itself has been under development in its pre-Facebook incarnation for more than three years.

So a week from now, you can join the Imperial Navy and with your friends go forth to reclaim the lost Rozaxra Domain from Sad-Eye and Deader aliens, Mrouzh Rebels, Naknuk Secessionists, pirates, rebels and other star-scum. To your ships, my friends!

This is the old version of the method.  It finds some specific views and scans their datapaths, to see if any entries match.  This version mostly worked fine, but occassionally it would fail, and finding the cause of the failure was proving difficult.

    <method name="check_ok" args="from, to">
        <![CDATA[
        for (var i = 1; true; i++)
        {
            var match = to.classroot.datapath.xpathQuery(
                'matches/match[' + i + ']/text()');
            if (match == null) break;

            for (var j = 1; true; j++)
            {
                var type = from.datapath.xpathQuery(
                    'types/type[' + j + ']/text()');
                if (type == null) break;

                if (match == type)
                    return true;
            }
        }
        return false;
        ]]>
    </method>

WHAT WORKED

I finally tracked down the problem; it occured whenever “to” was not replicated (i.e. there was only one).  So I assume that the replication manager was confusing my view navigation, as the number and type of views to navigate changed.

The solution is shown in the code below.  Basically I navigate the dataset directly, and avoid any view navigation.  Nothing can go wrong, right?

    <method name="check_ok" args="from, to">
        <![CDATA[
        var e  = to.datapath.dupePointer();
        e.selectParent(2);
        e.setXPath('matches');
        if (e.selectChild())
        {
            do
            {
                if (e.getNodeName() == 'match')
                {
                    var c  = from.datapath.dupePointer();
                    c.setXpath('types');
                    if (c.selectChild())
                    {
                        do
                        {
                            if (c.getNodeName() == 'type')
                            if (e.getNodeText() == c.getNodeText())
                                return true;
                        }
                            while (c.selectNext());
                    }
                }
            }
                while (e.selectNext());
        }
        return false;
        ]]>
    </method>

Unsurprisingly, this is very very bad.

The bug is documented here:
    bug: destroy before resource is loaded leads to full load queue and hang

The fix is noted in the bug report.  Basically the following code is added to the application.

    <script>
        LzMakeLoad.unload = function ()
        {
            this.loader.unload( this.loader.mc );
        }
        LzMakeLoad.destroy = function (recur)
        {
            this.loader.unload( this.loader.mc );
            super.destroy( recur );
        }
    </script>

The display is text. The text has an ‘id’ of ‘loading’ and ‘visibility’ set to false.

<text id="loading" x="10" y="10" visible="false"
    fontsize="12" fgcolor="white" fontstyle="bold">
  Loading
</text>

The display is controlled in an idle handler. The visible property of the display is set depending on the number of open connections in the loader.

<handler name="onidle" reference="LzIdle">
  <![CDATA[
  loading.setVisible(LzLoadQueue.openConx > 0);
  ]]>
</handler>

The display shows you when the loader, and therefore the application, is busy. You can see how well load-spreading strategies like OpenLaszlo: background loading of images are working. It is a window into a part of your application that is normally hidden.

OpenLaszlo “official” demos

Here you will find a range of interactive demos showing some of what OpenLaszlo makes possible.
www

OpenLaszlo Incubator Project

The Incubator directory contains components and other code that have been contributed to the OpenLaszlo project, but have not yet been fully integrated into the product.
src

Applejuice

Applejuice is a free web-based photo gallery. Its user interface is simple and uncluttered, to give your photos center stage.
www

Our First iPhone App

www - src

OpenLaszlo YouTube Player Demo and Source Code

www

Open Source SimFaux OpenLaszlo Code Now Available via Subversion

SimFaux is an interactive TV station simulator written in OpenLaszlo, by Don Hopkins, for the HuffingtonPost Contagious Festival.
www

Source code for Zoom Maps project (Openlaszlo)

www - src

Laszlo: An Open Source Framework for Rich Internet Applications

www

blogbox

Blogboxes provide exciting, instantly deployed functionality for your blog or Web site.
www

Glancer: Friend finder and Scheduler application

www - src