. 85
( 132 .)


// set the $years array to a list of plausible years
$y = date(˜Y™);
$years = array(
$y => $y++
, $y => $y++
, $y => $y++
, $y => $y++
, $y => $y++
, $y => $y++
, $y => $y++
, $y => $y++

// use January of next year as a default expiration date
if (empty($order->cc_exp_mon))
$order->cc_exp_mon = 1;
if (empty($order->cc_exp_yr))
$order->cc_exp_yr = date(˜Y™)+1;

print labeled_row(
, select_field(array(
˜name™ => ˜cc_exp_mon™
, ˜values™ => $months
, ˜match™ => $order->cc_exp_mon
. select_field(array(
˜name™ => ˜cc_exp_yr™
, ˜values™ => $years
, ˜match™ => $order->cc_exp_yr
504 Part IV: Not So Simple Applications

print end_table();

// display the order button
print paragraph(
, submit_field(˜Test Submit™)

print end_form();

print end_page();

$_SESSION[˜order™] = serialize($order);

This chapter explained a lot of PHP concepts, using a shopping cart as a vehicle (no
pun intended).
One of the most important concepts we discussed was persistence ” the ability to
store information related to a particular user (such as shopping-cart items) across
multiple visits. PHP relies upon its connectivity to a database ” a connection medi-
ated by the PEAR classes, if you™re smart ” to store information.
Another key concept is state maintenance. Because HTTP is an inherently state-
less protocol, you have to do a bit of work to correlate one HTTP request from
someone with the next request from the same person, all the while distinguishing
those requests from the hundreds of others that might be showing up at about the
same time. PHP provides some useful state-management features. You saw, for
example, that there™s nothing to the process of registering a session identifier and
examining it later to identify a particular use.
These capabilities are key in our shopping cart application. For one thing, we
used persistence to store the contents of each shopping cart in a database, in such
a way that they were kept separate from all others. Furthermore, we used session
management to track users through our site ” across many request/response trans-
actions ” as they browsed our wares and added and removed items from their carts.
Chapter 15

XML Parsing

— Learning how to work with XML documents

— Examining an XML document retrieved from a URL

WHAT HTML the eXtensible Markup Language, or XML, is to
data. Whereas HTML is about presenting your information to the world ” handling
typefaces, sizes, colors, layout, and so on ” XML concerns itself purely with struc-
turing and identifying that information. Given the number of new and different
Web platforms that seem to pop up each week ” desktop computers, laptops, cell
phones, televisions, wristwatches, car stereos ” this separation of content from pre-
sentation is the great holy goal of Web programmers. Both languages are wildly
successful because at heart they™re both very simple, yet allow for a huge range of
applications. They even look alike ” which they should, considering they™re both
based on the Standard Generalized Markup Language (SGML).
In an XML document containing meteorological-observation data, for example,
distinct tags can identify certain numbers as wind-velocity values and other numbers
as wind-direction values. The question of how to represent these values visually ” if
they are to be displayed visually at all, rather than just read into a database or other
processing environment ” is a separate issue.
The other great thing about XML is that it™s an excellent format for transmitting
information not just data, but the kinds of queries and responses you might nor-
mally associate with a regular programming language (like PHP itself). This is
because it™s lightweight ” it™s just ASCII (or Unicode), after all ” and transparent.
You can look at an XML document for the first time and stand a good chance of
understanding it right away. Yet it™s also suitable for interpretation by machines,
which are notoriously dim when it comes to understanding.
In this chapter we™ll explore XML and the capabilities of PHP when it comes to
processing it. We™ll read a document in from a URL and reformat its contents for
use in Netsloth, our content-management application.

506 Part IV: Not So Simple Applications

The Web site Slashdot.org is used in this chapter merely for example pur-
poses. Most sites like Slashdot.org have terms and conditions governing the
use of the content they post, so be sure to pursue the proper permission
before you publicly post any content taken from another Web site through
your own Web application. For information about using headlines from
Slashdot specifically, you should take a look at http://slashdot.org/

Scope and Goals of Application
Not long ago, we had this great application to show you here. It had XML parsers,
and event handlers, and function callbacks, and all kinds of flashy bits. The idea
was to grab the headlines from a Web site like Slashdot (www.slashdot.org),
which makes their content available in about every format known to modern
humanity. One of these formats is XML, and you can see it yourself at http://
slashdot.org/slashdot.xml. We would read it, parse it, and spit it back out as
HTML, to be included on our tiny yet distinctive example of a Web content site,
Netsloth (which you might remember from Chapter 11).
But then the folks building PHP decided that they would change their underly-
ing XML support, building it around the Gnome XML library libxml2, and intro-
ducing this new extension called Simplexml, and, this is how you would build our
first example now. Completely.

// keep the errors off the page
ini_set(˜display_errors™, 0);
$url = ˜http://www.slashdot.org/slashdot.xml™;
$cachefile = “/tmp/slashdot.xml.cache”;

if (($xml = file_get_contents($url)))
file_put_contents($cachefile, $xml);
error_log(˜Unable to contact www.slashdot.org™);
if (($xml = file_get_contents($cachefile)) === FALSE)
Chapter 15: XML Parsing 507

error_log(“Unable to open cache file: $cachefile”);
print <<<EOT
Unable to obtain Slashdot.org content.
Please try again later.
$stories = simplexml_load_string($xml);
print <<<EOT
<h3>Slashdot Stories:</h3>
foreach ($stories->story as $story)
print <<<EOT
<li><a href=”{$story->url}”>{$story->title}</a>
print <<<EOT

This would make for a rather short chapter. You™ll notice particularly that the
“handle the XML” part of this code is just ten lines. So we™ve jazzed it up a bit.
We™ll want to be able to include more information about each story, including
the topic-representing images the site provides. At the same time, we want our own
page at Netsloth to keep running if Slashdot gets slashdotted and goes off the air,
while minimizing the amount that we add to their site™s traffic. Both goals involve
setting up local caches of content. Figure 15-1 shows you the new Netsloth home

The stories shown in Figure 15-1 are for example purposes only. None of the
content in this chapter represents postings that ever actually appeared on
508 Part IV: Not So Simple Applications

Figure 15-1: Netsloth21 home page with mock sample stories

Code Overview
The essential purpose of our application is to reach out across the Internet, grab an
XML document, pick it apart, and reformat the chopped-up pieces into a form that™s
acceptable for use in our Netsloth content-management suite. To accomplish these
goals, this software will need to be able to look at an XML document, distinguish
the tags from the tagged text (also known as character data), and separate them if
necessary. The piece of code that does this is generically called an XML parser.
XML parsers can be designed in a couple of different ways.

An introduction to parsers
If you™re going to work with XML, you need a parser. Parsers come in two different
general varieties:

— Tree-style parsers (also called Document Object Model (DOM) parsers),
which read through entire XML documents at once and convert the
imported data into hierarchically organized objects representing whole
documents at once. Microsoft™s MSXML parser is of this kind.
Chapter 15: XML Parsing 509

— Event-style parsers, which read through XML documents just like
tree-style parsers, but fire events as they go. These events correspond
to different elements (such as opening tags, closing tags, and character
data) encountered in the read-through. It™s therefore possible to process
different elements with code that listens for events of different kinds.
The Simple API for XML (SAX) is an event-style parser implemented
in a number of programming languages.

Both kinds of parsers do the job ” it™s possible to use either to examine an XML
document programmatically. The difference between the two is performance.
Because tree-style parsers have to read in a whole XML document and store it in
memory as an object, they tend to be more resource-intensive than event-type
parsers. Generally, tree-style parsers are a good idea if you™re going to be examin-
ing the whole tree, or widely scattered parts of it ” particularly more than once.
Event-type parsers are better for quick, once-off examination of small parts of the
Of course, you can just ignore the whole question of what style of parser suits
you, and use Simplexml instead.

Using Simplexml
The Simplexml extension is, as of this writing, a work in progress. Still, it™s hard to
imagine that it could get much easier to use. It™s a bit like one of those auto-focus,
auto-everything-else cameras: you point it at some XML and, click, you™ve got an

$xml_object = simplexml_load_file(“/path/to/my/file.xml”);
$xml_object = simplexml_load_file(“http://a.server/file.xml”);
$xml_object = simplexml_load_string($some_xml_content);

Note especially the second example. Because PHP has general support for using
URLs the same way you would use a path to a file on your server, you can go
directly from an XML document from some far-off location to a useable object in
your own code. Still, there are a few things to look out for. The nature of the prop-
erties of the simplexml_element object produced depends on the content of the
XML. For example, we can make up a simplistic XML document and store it in a

$doc = <<<EOT


. 85
( 132 .)