The Internet Encyclopedia (Volume 3)

P1: C-149-Stotts

Perl WL040/Bidgolio-Vol III Ch-04 August 14, 2003 11:22 Char Count= 0

NETWORKPROGRAMMING INPERL 47

use LWP::UserAgent; # imports other

modules too

$client = new LWP::UserAgent();

$acmeReps = new URI::URL('www.acme.com/

reports/index.html');

$outHead =

new HTTP::Headers(User-Agent=>'MyBot

v2.0', Accept=>'text/html');

$outMsg = new HTTP::Request(GET, $acmeReps,

$outHead);

$inMSg = $client->request($outMsg);

$inMsg->is_success? {print $inMsg->

content;} : {print $inMsg->message;}

The network connections and message formatting to

HTTP protocol requirements is handled transparently by

the LWP functions. Here, the client causes a request like

this to be sent to the Web server at http://www.acme.com:

GET/reports/index.html HTTP/1.0

User-Agent: MyBot v2.0

Accept: text/html

If the requested file is not there, the response from the

server will be an error message. If it is there, the response

will contain the HTML file for the “bot” to process in some

fashion.

Database Use

Many Web sites depend on databases to maintain per-

sistent information—customer data, statistics on site us-

age, collection of experimental results, medical records.

Perl has several ways a CGI script (or any Perl script,

Web-based or otherwise) can store and retrieve database

records.

Module “DBM”

Perl comes with a module called DBM (database module)

that contains functions implementing a built-in database

structure. DBM treats an external data store internally

as hashes, or key/value pairs. This internal database is

intended for simple, fast usage; searching requires as few

as three lines of script. The method is convenient for fast

retrieval, even from very large databases, when there is a

unique key for searching (e.g., ISBN numbers). It is not

as useful for complex data or queries:

dbmopen (%db, $database, 0400) || die "Can't

open DB"; #open read only

for ($k=$max; $k< $max+20; $k++){print

"$k $db{$k}"} #get and print data

dbmclose (%db); #close database

In this example, we know the index values are numbers

and are unique. The database looks to the Perl script like

a hash, so the associative array variable%dbis used to get

the data values from it.

One useful feature Perl provides is DBM filters. These

are small data transformation routines, written by a pro-

grammer to manage situations where the format of the

data fields in a database is not quite compatible with the

form needed by a script. Rather than put small data ma-

nipulation code chunks scattered throughout the script,

or write and call extra functions, DBM filters can be at-

tached directly to the fields of a database; data transfor-

mation then takes place automatically as field values are

moved into and out of the database. This feature makes

the code using the database easier to read and less error

prone due to less code replication.

Module “DBI”

For more advanced applications, a relational database is

often more desirable than the simple structure of DBM.

For this Perl provides the DBI module, or data base in-

terface. DBI is designed to hide the details of specific

database systems, allowing the programmer to create

scripts that are general. The interface allows expressing

SQL queries, executing them against a specific database,

and retrieving the results. The DBI module does not con-

tain code specific to any database vendor’s product, but

it does have references to numerous vendor-specific mod-

ules called DBD’s (database drivers). A DBD module will

contain the detailed code needed to communicate with a

specific brand of database system.

Figure 1 illustrates the relationship among the execut-

ing script, the DBI, the DBD, and the physical database.

When a Perl script invokes a DBI function to execute a

query, the query is routed to the appropriate DBD mod-

ule according to how the DBI database handle was opened

(as an Oracle DB, as an Access DB, etc.). The DBD module

communicates with the actual files or tables of the phys-

ical database system and produces query results. These

results are communicated back to the DBI module, which

relays them back to the user’s Perl script. This layer of

indirection gives Perl scripts a generality that makes mi-

gration from one physical DB system to another relative

painless.

Module “ODBC”

In addition to the DBI module, programmers wishing to

write Perl scripts that interface to external databases, such

as Access or Oracle, can obtain an ODBC compatibility

package as a free download from several third party dis-

tributors. This module contains functions that implement

the ODBC standard database interface. Perl scripts writ-

ten to use this standard can then work with any relational

database under them, as long as that database has its

own ODBC interface. Almost all major relational database

vendors provide an ODBC implementation for their prod-

ucts. ODBC provides the same advantages as DBI, but the

ODBC standard was designed outside Perl and is available

in many other programming languages as well.

Processes and IPC

Although the Web is the major Internet application, it is

certainly not the only one, and Perl includes facilities to

support general network applications. Although not de-

signed specifically for concurrent or multiprocess compu-

tations, Perl does support various forms of processes, in-

terprocess communication (IPC), and concurrency. (Perl

having been born of Unix, this section is heavy on Unix

process concepts such as pipes and forking.) Processes are