The Back End

Perl

why Perl?

Perl's lineage and Larry Wall

interpreted vs. compiled

modules

getting started in Perl

the shebang line and which

naming your programs

add_user.cgi
batch_pages.pl

executable mode

$ batch_pages.pl
bash: .//batch_pages.pl: Permission denied

$ chmod u+x batch_pages.pl

tabs and line breaks

print "\tword1\tword2\tword3\tword4";

Whether this was printed to the screen or to a text file, the result would be tabbed:

       word1       word2       word3       word4

print "word1\nword2\nword3\nword4";

word1
word2
word3
word4

procedural syntax with ;

As you may have noticed in our print statement above, the line ended with a semicolon. Perl is a procedural language, executing line by line, and the lines are defined by semicolons. They are analagous to a period at the end of a sentence.

What can be said in one sentence...

print "abcdefg";

print "a";
print "b";
print "c";
print "d";
print "e";
print "f";
print "g";

abcdefg

comments

# print tab delimited words
print "\tword1\tword2\tword3\tword4";

print "\tword1\tword2\tword3\tword4"; # print tab delimited words

program flow with exit

At any point in your program, you can choose to quit, give up, throw your arms in the air scream "stop!". Luckily the most you need to do is use "exit". Perl programs will normally run until the last line in your program file is reached, unless they encounter "exit":

print "a";
print "b";
print "c";
exit;
print "d";

if ($x == $y) {exit;}
else {print "values are not equal";}

variables

scalar variables

$x = 4;

$x = "4";

arrays

An array is a list of values and is denoted by the @ symbol. Making lists of variables is a logical way to solve certain problems in programming. Each value within an array is known as an element. Let's declare a three-element array:

@dogs = ("Fido", "Rover", "Buddy");

Each element in the array is actually a scalar variable, and is numbered by Perl starting with 0. To print "Fido" and only "Fido" we would access the first element in the array like this:

print "$dogs[0] is a bad dog.";

And we get:

Fido is a bad dog.

print "$dogs[1] and $dogs[2] are good dogs.";

Rover and Buddy are good dogs.

print "@dogs";

FidoRoverBuddy

using library files with require

require ("../coordinator.lib.pl");

#!/usr/bin/perl

# set the parameter values for the MySQL connection
$databaseName = "clients";
$databaseUser = "marcus";
$databasePw = "superfly23";
########################
# url
$mindmined="http://www.mindmined.com";
# site root
$root="/home2/mindmined";
$www="$root/www";

error handling with || die

open FILE, "users.txt";

If users.txt exists right in that same directory, and we are given authorization (read permission) to open it, all will be well. But if the file is missing or protected, this statement will fail, but we'll receive no explicit statement to that effect. We'd have to figure out where the program failed by some other means.

To get a relevant error message, add the "or" operator (||) after the basic statement and force the program to quit there and then with "die". We then feed "die" what we'd like to be printed in the vent of failure:

open FILE, "users.txt" || die "Couldn't open file: $!";

the -w switch

#!/usr/bin/perl -w

using CGI

use CGI;

instantiating an instance of CGI

$cgiobject = new CGI;

trapping errors in the browser

use CGI::Carp('fatalsToBrowser');

grabbing parameters sent from an HTML form

$first_name=$cgiobject->param("first_name");
$last_name=$cgiobject->param("last_name");
$bio=$cgiobject->param("bio");

output to the browser

When using the "print" command in Perl, by default we are printing to the STDOUT filehandle, which usually means it is printed onscreen at the command line. In a CGI program however, we will most likely want to print something to the browser in response to the user's form submission. For this we need to provide a "header" which tells the browser "here comes some HTML". Two line breaks (\n\n) signify the end of the header. Everything printed after that point is directed to the browser and interpeted as a webpage.

print "Content-type:text/html\n\n";

print "<html>";
print "<title>Success</title>";
print "<body><h2>Successful form submission...</h2>";
print "first name: $first_name<br>";
print "last_name: $last_name<br>";
print "bio: $bio<br>";
print "</body>";
print "</html>";

browser redirect

print "Location: http://www.somesite.com/successpage.html\n\n";

writing to and testing from the command line

$ add_user.cgi last_name="Springsteen" first_name="Bruce" bio="Born in the USA"

i = 0;
foreach $dog (@dogs) {
     i++;
     print STDOUT "$i\n";
}

file handling commands

opening a file for reading

Anytime we are dealing with a file, it is your responsibility as programmer to give the file a special name by which Perl will know and refer to it. This is know as the filehandle. Filehandles are always expressed in capital letters and must be unique from one another, lest you utterly confuse your program.

In order to open a file for reading, use the open keyword followed by your chosen filehandle, then the path to the file itself. Don't forget to follow it up with an error trapping statement in case something goes wrong:

open USERDATA, "/home/marcus/users.txt" or die "Couldn't open file: $!";

opening a file for writing

open(USERPAGE, "> /home/marcus/users.html");

appending to a file

open(USERPAGE, ">> /home/marcus/users.html");

printing to a filehandle

print USERPAGE "<html><body>user1<br>user2<br>user3<br></body></html>";

writing HTML from within Perl

print USERPAGE "<img src=\"flowers.gif\" border=\"0\">";

print USERPAGE "<html>\n<body>\nuser1<br>\nuser2<br>\nuser3<br>\n</body>\n</html>";

closing a file

close(USERPAGE);

manipulating strings

using templates with 1:1 replaces

Conceptually, you can break down nearly every webpage into two things: the layout and the content. The layout is an element that you establish in the design phase, perhaps pasting in your text and graphic content at that time as well. But what if we want to leverage a successful layout for more than one page on our website, or change the content on the page periodically to keep the site fresh? Sounds like time to turn the layout into a template.

A template is just an HTML file with all the content removed, leaving only the text and images (and perhaps links for navigation) which we want to see every time. Let's take a very simple example webpage to start, with the what we consider the content bolded:

<html>
<body>
<table width="50%" border="1">
<tr><td width="50%">Name</td><td width="50%">Email</td></tr>
<tr><td width="50%">Marcus Del Greco</td><td width="50%">marcus@mindmined.com</td></tr>
</table>
<a href="home.html">back home</a>
</body>
</html>

<html>
<body>
<table width="50%" border="1">
<tr><td width="50%">Name</td><td width="50%">Email</td></tr>
<tr><td width="50%"><full_name></td><td width="50%"><email></td></tr>
</table>
<a href="home.html">back home</a>
</body>
</html>

back home

Now comes the exciting part. We need to substitute the real data for our placeholders. After loading our template file into a scalar variable called $file, we can match the pattern of the placeholders and replace them with the real data:

$file =~ s/<full_name>/$full_name/;
$file =~ s/<artist_bio>/$email/;

open(NEWFILE, "> user1.html");
print NEWFILE "$file";
close(NEWFILE);

using templates with 1:1 replaces

using templates with xml knockoff markup

pattern matching and sed

substrings