Tuesday, May 11, 2010

How to put CDATA into script tag in XSLT

You have an XSL template for the website you are working on and you would like to embed some JavaScript in the markup. You care so you would like to keep the XHTML output valid. Easy enough - all it takes is wrapping the actual JavaScript code with CDATA. To make it safe you would also add JS comments around CDATA and move on. But is it really that easy with XSL? Let's have a look.

XSL template for JavaScript

Here is the working solution which allows for safe embedding JavaScript on XHTML pages created with XSL templates.

Example 0
<?xml version="1.0" encoding="UTF-8"?>
<xsl:transform version="1.0"
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
    <xsl:output
        method="xml"
        encoding="utf-8"
        doctype-public="-//W3C//DTD XHTML 1.0 Strict//EN"
        doctype-system="http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"
        indent="yes"
        cdata-section-elements=""
        omit-xml-declaration="yes" />
 
    <xsl:template name="javascript">
        <xsl:param name="code"/>
            <xsl:text disable-output-escaping="yes">
            &lt;script type="text/javascript"&gt;
            /* &lt;![CDATA[ */ </xsl:text>
            <xsl:value-of select="$code" disable-output-escaping="yes"/>
            <xsl:text disable-output-escaping="yes">
            /* ]]&gt; */
            &lt;/script&gt;
            </xsl:text>
    </xsl:template>
        
    <xsl:template match="/">
        <html>
            <body>
                <xsl:call-template name="javascript">
                    <xsl:with-param name="code">
                        <![CDATA[
                        if (1 > 2) {}
                        ]]>
                    </xsl:with-param>
                </xsl:call-template>
            </body>
        </html>
    </xsl:template>
</xsl:transform>
The above example will output:
  1. <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
  2. <html xmlns="http://www.w3.org/1999/xhtml">
  3. <body>
  4. <script type="text/javascript">
  5. /* <![CDATA[ */
  6. if (1 > 2) {}
  7.  
  8. /* ]]> */
  9. </script>
  10. </body>
  11. </html>
  12.  

Can it be simpler?

Let's start off with the most intuitive approach - add CDATA inside script tag as you would in XHTML.

Example 1
  1. <?xml version="1.0" encoding="UTF-8"?>
  2. <xsl:transform version="1.0"
  3. xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  4. <xsl:output
  5. method="xml"
  6. encoding="utf-8"
  7. doctype-public="-//W3C//DTD XHTML 1.0 Strict//EN"
  8. doctype-system="http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"
  9. indent="yes"
  10. cdata-section-elements=""
  11. omit-xml-declaration="yes" />
  12.  
  13. <xsl:template match="/">
  14. <html>
  15. <body>
  16. <script type="text/javascript">
  17. /* <![CDATA[ */
  18. if (1 < 2) {}
  19. /* ]]> */
  20. </script>
  21. </body>
  22. </html>
  23. </xsl:template>
  24. </xsl:transform>
The above example will output:
  1. <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
  2. <html xmlns="http://www.w3.org/1999/xhtml">
  3. <body>
  4. <script type="text/javascript">
  5. /* */
  6. if (1 &lt; 2) {}
  7. /* */
  8. </script>
  9. </body>
  10. </html>
  11.  

CDATA got stripped off and the < sign got replaced with &lt;. Of course! CDATA as an XML specific construct is interpreted in XSL as well. And that helps - try to remove CDATA from the following example and you will see an error. Good old < is the source of trouble here - without CDATA around, it makes XML invalid.

The interpreted CDATA should stay then but another CDATA is required in the output. xls:output has an option which allows to wrap content of certain tags with CDATA. However if you set it up to wrap up content of script tags with CDATA (the rest of the code like in example 1) you will not get what you need. Let's have a look.

Example 2
  1. cdata-section-elements="script"
The above example will output:
  1. <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
  2. <html xmlns="http://www.w3.org/1999/xhtml">
  3. <body>
  4. <script type="text/javascript">
  5. <![CDATA[
  6. /* */
  7. if (1 < 2) {}
  8. /* */
  9. ]]>
  10. </script>
  11. </body>
  12. </html>
  13.  

Ok, now there is CDATA in the output but it is not safely commented out. And it does not seem possible to get it that way. And how about creating CDATA explicitly as a text? Maybe with disabled output escaping that would work? Let's see.

Example 3
  1. <?xml version="1.0" encoding="UTF-8"?>
  2. <xsl:transform version="1.0"
  3. xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  4. <xsl:output
  5. method="xml"
  6. encoding="utf-8"
  7. doctype-public="-//W3C//DTD XHTML 1.0 Strict//EN"
  8. doctype-system="http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"
  9. indent="yes"
  10. cdata-section-elements=""
  11. omit-xml-declaration="yes" />
  12.  
  13. <xsl:template match="/">
  14. <html>
  15. <body>
  16. <script type="text/javascript">
  17. <xsl:text disable-output-escaping="yes">
  18. /* &lt;![CDATA[ */
  19. </xsl:text>
  20. <![CDATA[
  21. if (1 < 2) {}
  22. ]]>
  23. <xsl:text disable-output-escaping="yes">
  24. /* ]]&gt; */
  25. </xsl:text>
  26. </script>
  27. </body>
  28. </html>
  29. </xsl:template>
  30. </xsl:transform>
The above example will output:
  1. <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
  2. <html xmlns="http://www.w3.org/1999/xhtml">
  3. <body>
  4. <script type="text/javascript">
  5. /* <![CDATA[ */
  6.  
  7. if (1 &lt; 2) {}
  8.  
  9. /* ]]> */
  10. </script>
  11. </body>
  12. </html>
  13.  

That's definitely some progress, CDATA is in place. Now the &lt; in the JavaScript code should become < again. Why not to put the whole JavaScript code into another xsl:text tag with disable escaping? Well, that's why:

Example 4
  1. <script type="text/javascript">
  2. <xsl:text disable-output-escaping="yes">
  3. /* &lt;![CDATA[ */
  4. </xsl:text>
  5. <xsl:text disable-output-escaping="yes">
  6. <![CDATA[
  7. if (1 < 2) {}
  8. ]]>
  9. </xsl:text>
  10. <xsl:text disable-output-escaping="yes">
  11. /* ]]&gt; */
  12. </xsl:text>
  13. </script>
The above example will output:
  1. <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
  2. <html xmlns="http://www.w3.org/1999/xhtml">
  3. <body>
  4. <script type="text/javascript">
  5. /* <![CDATA[ */
  6.  
  7.  
  8. if (1 &lt; 2) {}
  9.  
  10.  
  11. /* ]]> */
  12. </script>
  13. </body>
  14. </html>
  15.  

Is it a dead end or maybe xsl:text element is not the best choice after all? It has to be something with disable-output-escaping attribute. Another such an element is xsl:value-of. Assuming that the JavaScript code exists only in the template and not in the processed XML structure, value-of needs an xsl:variable to read value from.

Example 5
  1. <xsl:variable name="s1">
  2. <![CDATA[
  3. if (1 < 2) {}
  4. ]]>
  5. </xsl:variable>
  6. <script type="text/javascript">
  7. <xsl:text disable-output-escaping="yes">
  8. /* &lt;![CDATA[ */
  9. </xsl:text>
  10. <xsl:value-of select="$s1" disable-output-escaping="yes"/>
  11. <xsl:text disable-output-escaping="yes">
  12. /* ]]&gt; */
  13. </xsl:text>
  14. </script>
The above example will output:
  1. <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
  2. <html xmlns="http://www.w3.org/1999/xhtml">
  3. <body>
  4. <script type="text/javascript">
  5. /* <![CDATA[ */
  6.  
  7. if (1 < 2) {}
  8.  
  9. /* ]]> */
  10. </script>
  11. </body>
  12. </html>
  13.  

And that worked! It is not very flexible solution, though. Defining variable for every JavaScript in the template seems to be acceptable only if necessary. Luckily xsl:value-of can also read value of a parameter passed to an xsl:template. That's how I got to the template you can find at the top of the page.

Thank you for reading. I hope you found it useful.

Update - 21.05.2010

At the beginning the template looked like this:

  1. <xsl:template name="javascript">
  2. <xsl:param name="code"/>
  3. <script type="text/javascript">
  4. <xsl:text disable-output-escaping="yes">
  5. /* &lt;![CDATA[ */ </xsl:text>
  6. <xsl:value-of select="$code" disable-output-escaping="yes"/>
  7. <xsl:text disable-output-escaping="yes">
  8. /* ]]&gt; */
  9. </xsl:text>
  10. </script>
  11. </xsl:template>

However it turned out that on some PHP installations XSLT adds CDATA inside <script> tag even if it's not specified in the <xsl:output>. Therefore <script> tag has to be generated as a text rather than being part of interpreted XML.

Jacek

Friday, April 2, 2010

Zend PHP 5 Certification Exam - review

This time it's about the Zend PHP5 exam itself, not about PHP. I have taken the exam today and singing loudly in my car on the way back home I thought: the guys who are about to take it might want to know how it goes. And this is the story about what happens when you decide to go into the lion's cave armed with your brain and that only.

Well, it's not that bad. People I met in the exam centre were very friendly. I was taking my exam in Feltham, in the test center run by TSM Consultants UK Ltd. Poor folks were working even though it was bank holiday. They said it was all right, though.

They are serving as a test spot for some insane number of vendors, one of them being Zend. It is all very professional. When I booked the exam, I was asked to turn up about 15 minutes before the exam starts. I was also told to have two forms of identification on me, at least one with a picture of my lovely face. I complied, of course, and I am glad I did. It is not just talks - they take it really seriously. No id? Then they wave you goodbye and invite you some other time when you can be bothered to read the instruction before coming.

I was being checked in by a nice lady who asked me to read the terms and conditions. I learned I was not allowed to take anything with me to the examination room - mobile, jacket, rack sack even my beloved coffee had to be left behind. They gave me a locker for my valuables, though, so I put there my coffee. I am joking, my jailbroken iPhone of course. And I had to use the toilet beforehand because if you leave the room once the exam has started there is no coming back.

When I was ready we entered the exam room. It was poket-size with about 5 computers standing around the walls. There was also proper Big Brother equipment. That's right. You are being watched, listened in and recorded during the exam. Who cares after all? If you know you are not going to make it you can always smash a few monitors and hope for becoming famous on YouTube after they release you from custody.

Speaking of the monitors - they were nice biggies, so I had no trouble reading the questions. If the chairs were a bit more comfy I would not complain, though. That's because the room was equipped with budget low-end office chairs. On the other hand the exam lasts for 90 minutes only, so being brave fellow I gritted my teeth and took my seat. The countdown timer started after I accepted the terms and condition and when the first question appeared on the screen.

After a few questions I realised the exam was not very different from the Zend PHP 5 Certification Mock Exam. It is similar type of questions and if you succeeded with a few of these mock exams you should be fine. The whole exam consists of 70 questions served in random order. For each question you may add a comment or mark it for later review. That last feature turned out quite handy when I gave all answers and wanted to come back to those which I was not entirely sure about. Review screen showed me links to all the questions and indicated the marked ones. Finally I decided I was ready and after answering zillion times 'yes' to a question going like: 'are you sure you want to finished the exam?' I saw big, beautiful and smiling at me from the screen:

CONGRATULATIONS!
You have passed Zend PHP 5 Certification Exam.

Wednesday, February 17, 2010

PHP: References To Array Elements Are Risky

References to array elements can bite! And it is not only the case with referencing in foreach loop. It seems that creating a reference to an array element replaces that element itself with a reference. If you then copy such an array and change the elements inside copy you can overwrite original value!

All of the presented code was tested on Mac (PHP 5.2.11), Linux (PHP 5.2.6-1+lenny4) and Windows XP (PHP 5.3.0) using PHP cross platform testing lab on Mac based on VirtualBox.

Changing copy affects original array

Sounds impossible? I agree. I couldn't believe it myself. Nevertheless here is the proof:
Example 1
$a = array('one', 'two', 'three', 'four'); 
$a2 = &$a[2]; 
 
$b = $a; 
$b[1] = 'two again'; 
$b[2] = 'reference bites'; 
 
var_dump($a, $b);
The above example will output:
array(4) {
  [0]=>
  string(3) "one"
  [1]=>
  string(3) "two"
  [2]=>
  &string(15) "reference bites"
  [3]=>
  string(4) "four"
}
array(4) {
  [0]=>
  string(3) "one"
  [1]=>
  string(9) "two again"
  [2]=>
  &string(15) "reference bites"
  [3]=>
  string(4) "four"
}

Changing copy of a copy affects original array too

If you look at the dump carefully you will see that $a[2] and $b[2] are displayed as references here. That would mean the element has been replaced with a reference. And if you copy a reference you just get what? Same reference, right? So any copy of $a would contain that reference. Going forward any copy of a copy of $a would contain that same reference. Let's check it:
Example 2
$a = array('one', 'two', 'three', 'four'); 
$a2 = &$a[2]; 
 
$b = $a; 
$c = $b; 
$c[2] = 'references bites more'; 
$b[1] = 'two again'; 
$a[0] = 'one more'; 
var_dump($a, $b, $c);
The above example will output:
array(4) {
  [0]=>
  string(8) "one more"
  [1]=>
  string(3) "two"
  [2]=>
  &string(21) "references bites more"
  [3]=>
  string(4) "four"
}
array(4) {
  [0]=>
  string(3) "one"
  [1]=>
  string(9) "two again"
  [2]=>
  &string(21) "references bites more"
  [3]=>
  string(4) "four"
}
array(4) {
  [0]=>
  string(3) "one"
  [1]=>
  string(3) "two"
  [2]=>
  &string(21) "references bites more"
  [3]=>
  string(4) "four"
}

Tricky "foreach" with reference explained

As you can see, the only element affected is the one referenced. That can lead to serious potential problems. Working on a copy of an array with scalar elements seems to be not as safe as one may thought. However, that explains one of the biggest pitfalls of iterating arrays in PHP: foreach with reference. The following is rather common knowledge:
Example 3
$a = array('one', 'two', 'three', 'four'); 
foreach ($a as &$v) {} 
foreach ($a as $v) {} 
var_dump($a);
The above example will output:
array(4) {
  [0]=>
  string(3) "one"
  [1]=>
  string(3) "two"
  [2]=>
  string(5) "three"
  [3]=>
  &string(5) "three"
}
But what was the explanation for that, again? It is quite clear that $v keeps reference to $a[3] after the foreach loop is finished. So what values does $v (effectively $a[3]) get within the next foreach loop? Let's see:
Example 4
$a = array('one', 'two', 'three', 'four'); 
foreach ($a as &$v) {} 
foreach ($a as $v) { 
 var_dump($a[3]); 
}
The above example will output:
string(3) "one"
string(3) "two"
string(5) "three"
string(5) "three"
Well, it's quite obvious. Or is it? As the foreach manual page states: "unless the array is referenced, foreach operates on a copy of the specified array and not the array itself". I would assume that "array is referenced" means array elements are referenced like this: foreach ($a as &$v) {} Apparently that is not the case with the second loop and it should work on a copy. Let's have a look what exactly would happen if the second foreach loop really worked on a solid copy:
Example 5
$a = array('one', 'two', 'three', 'four'); 
$b = $a; 
foreach ($a as &$v) {} 
foreach ($b as $v) {} 
var_dump($a);
The above example will output:
array(4) {
  [0]=>
  string(3) "one"
  [1]=>
  string(3) "two"
  [2]=>
  string(5) "three"
  [3]=>
  &string(4) "four"
}
I imagine you did expect this result. Beware references to array elements and thanks for reading. I hope you found this article useful. Please don't think twice before you leave your comment.

Sunday, February 7, 2010

PHP: Cross Platform Testing Lab on Mac How To

Taking PHP development seriously, it is not very uncommon practice to do cross-platform tests of own code. With virtual machines it is dead easy to test your PHP code on three popular systems (Mac OS X, Linux and Windows) using only one physical machine. To make it even more pleasant all systems can share PHP code from one location hence there is no need to copy every time you want to test. And here comes the cream: you can get fully-blown eye-candy virtualizer for free.

VirtualBox is good and is free!

Yes, that is correct. VirtualBox is free and made by SUN. It is good piece of software, very reliable and easy to use. You can get VirtualBox for Windows, Mac or Linux.

VirtualBox performance on Mac Mini

I used VirtualBox to create Vrtual Cross-Platform PHP Lab on my Mac Mini run by Snow Leopard (10.6.2). My Mac was armed with 1.83GHz Intel Core 2 Duo CPU and 2BG of RAM. I have created one virtual machine for Windows XP SP2 and one for Debian 5.03 "lenny". Each virtual machine was limited to 256MB of memory and 2GB of hard drive space.

Running both virtual boxes with AMPs on them along with MAMP on the hosting Mac did not visibly affect overall host system responsiveness. I could have Chrome, Firefox, Aptana, Preview, iTunes, Skype and Activity Monitor running concurrently and smoothly. Only Spaces (9 desktops) were a bit more choppy than usual. Each of the virtual boxes was also responsive.

Both guest systems were visible and accessible over the local network. They could easily obtain IP addresses from DHCP server existing within the network. All the network communication was also reliable except for situations when one of the boxes (either real or virtual one) generated heavy traffic on the network interface. Downloading 100MB from Internet at 200kB/s was enough to cause connection disruption.

Set up Mac OS X host system

If you are PHP developer working on Mac, probably you already have half of the work done. Nevertheless assuming you are starting from scratch the following steps need to be accomplished:
  1. Download MAMP (Mac OS X AMP) and install it.
  2. Download VirtualBox for Mac and install it.
Both packages come with nice Macish installers and therefore the whole process is really straightforward.

Create WAMP virtual machine on your Mac

Now it's time to create your Windows virtual machine and set it up. I used Windows XP because it is relatively lightweight and fast compared to alternatives. VirtualBox is not fussy though and allows you to prepare virtual machine for many different operating systems including quite a few versions of Windows. Here is a step by step guide which tells you how to create a virtual machine, install Windows (XP) on it and configure WAMP to access PHP projects stored in the host filesystem:
Create new virtual machine for Windows.
Simply go to Machine -> New. Then just follow the guide, it takes about 2 minutes. There are no difficult questions and if you want to know more - the help is really good. I have created Virtual Machine called "virtual-winxp" with 256MB of RAM and 2GB of new virtual hard drive being dynamically expanding storage. The network adapter I have chosen was bridged adapter. The latter is important if you want to have access to LAN and Internet from your the guest OS and be able to access the guest OS from your Mac.
Install your preferred version of Windows.
My Win XP installation process was real time. When installer announced 30 minutes left - it was 30 indeed minutes. Network name of my Win XP was again "virtual-winxp".
Download WAMP and install it
Nothing I could think of to add here.
Install VirtualBox Guest Additions
Go to Devices -> CD/DVD Devices and choose VBoxGuestAdditions.iso. You can do so even when the guest OS is running. Virtual CD will appear in your drive and you will know what to do. VirtualBox Guest Additions are essential if you want to share your PHP code with the host OS.
Configure shared folder
That would allow you to share your PHP code with the host OS. Go to Devices -> Shared Folders... and click "Add shared folder" icon - the one with "+" sign. Then choose your projects folder from Mac, say /Applications/MAMP/htdocs/projects and name it, say "projects". Now that folder will be accessible from Windows as pseudo network drive \\vboxsrv\projects. It is probably a good idea to make it permanent if you don't want to repeat this step every time you start your guest OS.
Configure your WAMP to use projects folder
Let's assume you want to access your PHP projects through your WAMP as http://virtual-winxp/projects. Go to your Windows, click on the WAMP icon in system tray, choose Apache -> Alias directories -> Add an alias. When the command window appears, type in the alias name "projects".
On the second screen enter the real directory name being "//vboxsrv/projects". Forward slashes are not a mistake, Apache for Windows likes it like that. Your WAMP should restart and when you go to http://localhost/projects you should see your Mac projects directory.
Make your WAMP visible for your Mac
If you prefer to test results of your PHP on your Mac (who wouldn't) rather than on the guest OS, it might be a good idea, to make your WAMP visible in the local network. All you need to do is click on your WAMP icon in system tray, choose "Put Online" and wait till the Apache restarts.

Create LAMP virtual machine on your Mac

Preparing Linux virtual box was even easier because there was no clicking involved! Ok, in fact I was installing Debian without GUI, so I needed to amend a couple of files manually. No rocket science though, just some config files.
Create new virtual machine for your Linux
Similar procedure as for virtual machine for Windows, only different name ("virtual-debian" in my case). Settings exactly the same: 256MB RAM, 2GB virtual HDD expanding storage and bridged network adapter. Please remember, that the bridged network adapter is important if you want to access LAN and Internet from your guest OS and also access the guest OS from your Mac.
Install your preferred distribution of Linux
I have chosen Debian because it has very good package management and I like it. The hostname I set for it was again "virtual-debian".
Install nano
As mentioned before, you might need to change a few config files. If you are not familiar with vim it is probably a good idea to install more friendly console editor. When logged in as root run:
#aptitude install nano
Install LAMP for Debian
Just follow the link, it's all in there.
Prepare for installing VirtualBox Guest Additions
These are installed as a kernel modules which are being compiled for you by the installer. That means that you need to install kernel headers beforehand. First of all find out what kernel you are running:
#uname -r
I got result like this:
2.6.26-2-amd64
Once you know the version it's time to install headers. Simply repeat the version at the end of linux-headers package:
#aptitude install linux-headers-2.6.26-2-amd64
The installer might ask you a few questions, for example about installing additional packages. It's safe to answer "yes" to all of them.
Install VirtualBox Guest Additions
Again, go to Devices -> CD/DVD Devices and choose VBoxGuestAdditions.iso. You can do so even when the guest OS is running. Virtual CD will become ready for you to mount and access it:
#mount /media/cdrom
Then you need to copy installer to your hard drive:
#cd /media/cdrom
#cp VBoxLinuxAdditions-amd64.run /~

and run it:
#cd
#./VBoxLinuxAdditions-amd64.run

Configure shared folder
Same as with Windows virtual box. That would allow you to share your PHP code with the host OS. Go to Devices -> Shared Folders... and click "Add shared folder" icon - the one with "+" sign. Then choose your projects folder from Mac, say /Applications/MAMP/htdocs/projects and name it, say "projects". Now that folder will be available to mount from your Linux (Debian) as... projects (read next point for explanation). It is probably a good idea to make it permanent if you don't want to repeat this step every time you start your guest OS.
Configure your Debian to mount shared folder on startup
Now everything is set up and ready we only need to mount projects folder from your Mac on your Linux virtual box. Assuming you want to access your projects folder as http://virtual-debian/projects and you don't want to mess around with Apache config too much (at least at this point) simple create the mount point and mount:
#cd /var/www/
#mkdir projects
#mount -t vboxsf projects projects

Now inside projects directory you should see contents of your Mac projects directory. To make this mount occur every time you start your Linux, you need to add a line to /etc/fstab config file. Open the file:
#nano /etc/fstab
and add the following line:
projects    /var/www/projects    vboxsf    ro    0    0
Save the file and you are sorted. From now on, whenever you boot Linux the folder will be mounted for you and ready to use. And don't worry about the mount error - it's because default Linux mount happens before the vboxadd service is loaded.

Setup Fixed IP Addresses for your Virtual Farm

My virtual boxes get IP addresses from DHCP server running within my network. Their IPs used to change every now and then and I must admit I found it quite annoying. To avoid that problem I have assigned IP addresses statically to both virtual boxes on my DHCP server. Now they still receive the whole information package (DNS, gateway and so on) but their IPs are same every time.

Access your virtual boxes by names

If you prefer to access your virtual boxes via their names (like virtual-winxp) rather then IPs you may want to add them to your /etc/hosts file on your Mac. It's dead easy and takes no time. Simply run your Terminal.app (it's in /Applications/Utilities directory) and switch to root:
$su -
here you enter your password, and then edit the /etc/hosts file:
#nano /etc/hosts
Here you need to add one line for each virtual box. First IP then the name you like, for example:
192.168.0.11    virtual-debian
192.168.0.12    virtual-winxp
Save it and job done.

Thanks for reading. I hope you found this guide useful, it worked for you and made you a bit happier. Please don't think twice before you leave your comment.

Wednesday, January 27, 2010

PHP: array_rand does not shuffle!

PHP function array_rand() does not shuffle results anymore and behaves like array_keys() in some circumstances.

Yes, that is correct. The function array_rand() gets random key (or keys) from given array. Funny thing is that it does not shuffle the keys anymore. So the following code:

<?php
$a 
= array (=> 'key 5''key after 5'=> 'key 3''a' => 'key a'=> 'key 1'=> 'key 0''key after 0',);
print_r($a);
print_r(array_rand($a5));
print_r(array_rand($acount($a));
print_r(array_keys($a);
?>

would produce that result:

Array
(
    [5] => key 5
    [6] => key after 5
    [3] => key 3
    [a] => key a
    [1] => key 1
    [0] => key 0
    [7] => key after 0
)
Array
(
    [0] => 3
    [1] => a
    [2] => 1
    [3] => 0
    [4] => 7
)
Array
(
    [0] => 5
    [1] => 6
    [2] => 3
    [3] => a
    [4] => 1
    [5] => 0
    [6] => 7
)
Array
(
    [0] => 5
    [1] => 6
    [2] => 3
    [3] => a
    [4] => 1
    [5] => 0
    [6] => 7
)

That means, that getting randomly array-size number of keys from the array array_rand($a, count($a)) give exactly same result as getting all the keys in order array_keys($a).

Friday, January 22, 2010

PHP: Array Pointer in Undefined State

Array pointer in PHP may become "undefined" and that is when you try to move it after the last of before the first element. But what is more surprising - copying an array remaining in that state... resets the pointer.

First let's create a messy array and see how it dumps:

$a = array (5 => 'key 5', 'key after 5', 3 => 'key 3', 'a' => 'key a', 1 => 'key 1', 0 => 'key 0', 'key after 0',);
var_dump($a);

The result (as expected):

array(7) {
  [5]=>
  string(5) "key 5"
  [6]=>
  string(11) "key after 5"
  [3]=>
  string(5) "key 3"
  ["a"]=>
  string(5) "key a"
  [1]=>
  string(5) "key 1"
  [0]=>
  string(5) "key 0"
  [7]=>
  string(11) "key after 0"
}

When Array Pointer Becomes Undefined

Let's try and go beyond the last element in the array:

// move the pointer to the last element
end($a);

// move the pointer one element further
next($a);

// ... and see where we are
var_dump(key($a), current($a));

// try to go back
prev($a);

// ... and see where we are
var_dump(key($a), current($a));

The above code will produce the following result:

NULL
bool(false)

NULL
bool(false)

Which simply means that the array pointer becomes "undefined" if moved after last element. In fact the pointer is not "after" the last element because you cannot move it back with

prev()
function. It simply remains in undefined state.

Copying Array Resets The Pointer

When you copy an array, the array pointer is copied as well and it points to the same element as in the original. But what happens when the pointer is undefined? Will the copy inherit that as well? Let's consider the code:

// copy the array
$b = $a;

// check state of the original
print ('original: key() and current()' . PHP_EOL);
var_dump(key($a), current($a));

// check state of the copy
print ('copy: key() and current()' . PHP_EOL);
var_dump(key($b), current($b));

And the result:

original: key() and current()
int(5)
string(5) "key 5"

copy: key() and current()
NULL
bool(false)

Just as anticipated, the copy has its pointer in undefined state. But what happened to the original? Apparently copying the array worked like

reset()
and moved the pointer to the first element! Needless to say, copying the array preserves the pointer's value if it's defined:

// move the pointer to the last element
end($a);

// move the pointer to one element before last
prev($a);

// see where we are
print ('key() and current()' . PHP_EOL);
var_dump(key($a), current($a));

// copy the array
$b = $a;

// check state of the original
print ('original: key() and current()' . PHP_EOL);
var_dump(key($a), current($a));

// check state of the copy
print ('copy: key() and current()' . PHP_EOL);
var_dump(key($b), current($b));

And the output:

key() and current()
int(0)
string(5) "key 0"

original: key() and current()
int(0)
string(5) "key 0"

copy: key() and current()
int(0)
string(5) "key 0"

Moving Array Pointer Before First Element

All the above applies to the situation when you move the pointer "before" the first element. But don't take my word for it - try it out.

Friday, January 15, 2010

PHP5: Bitwise Left Shift Goes in Cycle

Going through ZEND PHP Certification Study Guide I came across a bit confusing example. It says that bitwise left shift by 32 positions on integer value of 1 performed on 32-bit machine would give 0 as a result.


Reading through the guide I usually write small snippets of code to check things myself and remember them better. So I did this time. To my big surprise, the result was different from what I had read. Here is the code:

<?php
$x = 1;
echo $x << 32;
// outputs 1

I wanted to find out what was going on, so I tried:

<?php
$x = 1;
print ('PHP_INT_SIZE: ' . PHP_INT_SIZE . "\n");
$maxBits = PHP_INT_SIZE * 8;

$format = '%0' . $maxBits . "b\n";

for ($i = $maxBits - 1; $i <= $maxBits + 1; $i++ ) {
printf('x << ' . $i . ': ' . $format, $x << $i);
}

for ($i = $maxBits * 2 - 1; $i <= $maxBits * 2 + 1; $i++ ) {
printf('x << ' . $i . ': ' . $format, $x << $i);
}

And here is the result I got:

PHP_INT_SIZE: 4
x << 31: 10000000000000000000000000000000
x << 32: 00000000000000000000000000000001
x << 33: 00000000000000000000000000000010
x << 63: 10000000000000000000000000000000
x << 64: 00000000000000000000000000000001
x << 65: 00000000000000000000000000000010

That means that bitwise shifting left goes in cycle modulo architecture-defined-number-of-bits. On 32-bit machine shift left by 32 bits gives same result as no shift at all (32 % 32 = 0). Shift by 33 is same as shift by 1 (33 % 32 = 1).
But is it really true? How about shifting different value than 1? I tried to shift 2 and here is the result:

PHP_INT_SIZE: 4
x << 31: 00000000000000000000000000000000
x << 32: 00000000000000000000000000000010
x << 33: 00000000000000000000000000000100
x << 63: 00000000000000000000000000000000
x << 64: 00000000000000000000000000000010
x << 65: 00000000000000000000000000000100

I checked the results on Mac (PHP 5.2.6), Windows (PHP 5.3) and Linux (PHP 5.2.12) and they were identical. It looks like new language feature which wasn't present while the ZEND PHP Certification Study Guide has been created. Although the Bitwise Operators manual page warns: "Don't left shift in case it results to number longer than 32 bits."
It's for another discussion whether to use it or not. Anyway, I guess it's good to know.