We will use  Robert Pepper's  tutorial on Basic Perl programming for Perl and Lincoln Stein's CGI module for using Perl in Common Gateway Interface (CGI) programming.  The following is the set of related links.

• Nix Silver's tutorial
• Perl Tutorial by Robert Pepper http://perl.sioc.org/win32perltut.html
• Perl manpage (syntax, builtin function) and modules web pages on www.perl.com
• Perl Regular Expression Summary
• FAQ on Perl
• Lincoln Stein's CGI.pm (a Perl5 CGI Library) module
• References: Oreily publisher has the whole series of perl books.
• Learning Perl 3rd edition is a good introductory book for beginners using perl UNIX systems.
• CGI Programming with Perl, 2nd Edition. demonstrate current techniques available with the CGI.pm module. Topics include incorporating JavaScript for form validation, working with databases, creating simple search engines, maintaining state between multiple sessions, generating graphics dynamically.
• Learning Perl on Win32 systems is a good introductory book for beginners using perl on Win32 systems (95,98?,NT).
• We will use unix machines in our software engineering lab, such as wetterhorn, redcloud, blanca, crestone, to carry out the Perl and CGI related exercises. You can remote login to carry out these exercises.
• To spread out the load, use the following table to pick the computer for your exercises. If the designated computer is down, you are allowed to use other computers and viva.uccs.edu. But be aware for debugging of CGI scripts, the error msg will be kept on the /etc/httpd/logs/error_log of the machnie that run the server side CGI script. See Debugging CGI Script Section for more information.
  

  Student ID % 6	Designated Computer for Web Programming Exercises
  0	sanluis
  1	blanca
  2	shavano
  3	crestone
  4	wetterhorn
  5	redcloud

  
  Any one with 995-99-xxxx or those students from CYU will use viva.uccs.edu for their exercises.

Setup your cgi-bin directory on CS Unix Machines

• Create a cgi-bin subdirectory in your public_html directory.
• Testing the setup:
• copy form.html to your own public_html
  cp ~cs301/public_html/form.html ~<login>/public_html
  Here <login> is your CS Unix machine login.
• modify the action attribute in the form.html,  Replace
  <form action="http://cs.uccs.edu/cgi-bin/cs301/fullecho.cgi" method=post> with
  <form action="http://blanca.uccs.edu/~<login>/cgi-bin/fullecho.cgi" method=post>
  
  Here blanca is your assigned web server and <login> is your CS Unix machine login.
• form tag is used to create form for collecting user input. See http://www.w3.org/TR/1998/REC-html40-19980424/interact/forms.html for more info.
• If you are assigned with CS Unix machines, run the following command "cp -r ~cs301/public_html/cgi-bin/* ~<login>/public_html/cgi-bin"
• If you are asssigned with viva.uccse.edu,
  cd public_html/cgi-bin
  scp cs522p1@windom:~cs301/public_html/cgi-bin/* .
• If your designated machine is sanluis, type "sanluis.uccs.edu/~<login>/cgi-bin/sessionvar.cgi" as URL to the browser. 
  You should see a list of environment variables passed to the CGI program, sessionvar.cgi.  That indicates your cgi-bin directory is set up right.
• Note that we did not turn on cgiwrap on cs.uccs.edu yet. The url http://cs.uccs.edu/~cs301/cgi-bin/sessionvar.cgi will return "access forbidden" or the actual source code depends on the access right of your sessionvar.cgi
• Type "sanluis.uccs.edu/~<login>/form.html" in Address box and see if the form data get echoed back.
• If you got Internal server error, you need to check the access rights of the cgi-bin and files just created.
  • The cgi-bin directory the cgi scripts needs to have read and execute permissions for the owner. Make sure it has -drwx------ or --rwx------ setting.
  • We are using a software package called cgiwrap so that users on our CS Unix machines can create their server side cgi scripts in their own cgi-bin directory.
• Besides remote login to the CS Unix machine, you can practice programming in perl and CGI scripting on your home PC by downloading and setting up xampp software from apachefriends.org, which include Apache, Perl, PHP, and MySQL.  You can then scp or sftp the CGI programs to the CS Unix machines for testing. Note that the first line of Perl CGI script needs to be changed since it indicates the location of interpreter on the web server.
• Click here for the info on downloading and setup of Apache, Perl, PHP, MySQL on windows.
• FAQ on ntperl: Perl for WIN32 PC.
• But be aware there are system related differences between UNIX and Win32 (95, 98 or NT). For example,
• Path name difference: Use "c:\\temp\\test.pl" to refer the file name c:\temp\test.pl since \ is a specal escape character in perl.
• Remember we use \ in \n for representing newline.
• You can use "c:/temp/test.pl" unix-style forward slash for path name. Now you know perl for WIN32 is 2nd class citizen in perl world. "c:\temp\test.pl" will not be a legal file name, since it is a string with c: followed by tab character, followed by emp, then tab, finally est.pl.
• Input  operation difference: For indicating the end of input string,
• on PC, use control-Z to represent no more input for the program on PC. If you try to use control-D to indicate no more input on PC, the program will keep waiting.
• on UNIX, use control-D to represent no more input for the program on UNIX.  If you use control-Z on UNIX, it will put your program on hold until you type fg to bring it back or bg to start background processing.
• System command/utility difference: The available command and utility routines are different.  Try to use those provide by the Perl library for portability.

Quick Summary of Perl

Setup: Editing and Running Perl

• Perl version 5.8.0 comes with Fedora Core 1 on our CS Unix machines. This perl interpreter is in /usr/bin/perl.
• On our CS Unix machines, you can type "perl -V" which will give us detailed version information.
• All perl programa start with a line similar to the following line

#!/usr/bin/perl
It tells the Unix shell to use /usr/bin/perl to interpret/execute the rest of the Perl source code.
If your perl are in different location, say /usr/local/bin/perl, you need to modify the above path name to
#!/usr/local/bin/perl
• On windows, if you install the XAMPP software package using default settings, then perl interpreter is located c:\apachefriends\xampp\perl\bin\perl.exe. The first line of your perl scripts should be
  #!c:\apachefriends\xampp\perl\bin\perl.exe
• Here the simplest Hello World! Perl program:

#!/usr/bin/perl
print "Hello World!\n";
• print is a built-in output function. It takes the following expression and output the corresponding string to the output device.
• All perl programs or perl scripts typically have extension .pl or .cgi (specifically for CGI scripts).
  • Use .pl for the file extension of a generic perl program. Typically used for standalone applications, not necessarily web application.
  • Use .cgi for the file extension of a CGI program (in this case written in Perl Language and understand CGI interface, often using CGI module.)
  • In our CS Unix machines, we have configured the web server to treat files with .cgi as web CGI application and will be interpreted by the corresponding interpreters. Files with .pl will be treated as a normal text file delivered without being interpreted!
• Use your favorite text editor, vi or pico, to create a file, hello.pl, with the above two lines.
• Login to your designated CS Unix machine
  sanluis.uccs.edu>cd ~cs301/public_html/cgi-bin
  sanluis.uccs.edu> hello.pl
  <html>
  Hello World!
  </html>
  sanluis.uccs.edu> hello.cgi
  content-type: text/html

  <html>
  Hello World!
  </html>
  sanluis.uccs.edu> wrongHello.cgi
  <html>
  Hello World!
  </html>
  

• You may get
  > hello.pl
  hello.pl: Permission denied.
  This is due to the file created by text editor will consider a text file and have 644 as access rights.  In order for it to be executed, you need to change the access right to 755 (use chmod 755 hello.pl).
• Enter "http://sanluis.uccs.edu/~cs301/cgi-bin/hello.cgi in the address box of your browser. It will display "hello world!"
• Enter "http://sanluis.uccs.edu/~cs301/cgi-bin/hello.pl in the address box of your browser. It will display
  #!/usr/bin/perl print "\n"; print "Hello World!\n"; print "\n";
  all in one line on IE broser, separate lines on Netscape 7.2! Hello.pl was not executed or interpreted.
• Note that the cgiwrap is turned off on cs.uccs.edu, therefore if you type http://cs.uccs.edu/~cs301/cgi-bin/hello.cgi, it will only not execute as script. It only shows the source code.
• Note that if you try to access the above wrongHello.cgi using http://sanluis.uccs.edu/~cs301/cgi-bin/wrongHello.cgi, the web server will return Server error! Error 500 msg. To find out details of the error, we login to sanluis and use "tail -f /etc/httpd/logs/error_log" to see the last few entries of the error_log generated by the Apache server in /etc/httpd/logs/ or /var/log/httpd. It shows that
  
  [Sat Sep 11 00:47:19 2004] [error] [client 63.191.104.66] malformed header from script. Bad header=<html>: cgiwrap
  
  
  Basically the web server complains that the wrongHello.cgi returns a web page without proper header such as "content-type: text/html\n\n" before the body of the web page <html>HelloWorld!. The header is required as the CGI secification or CGI protocol between the CGI process running the wrongHello.cgi and that of the web server. For more detail see the CGI web page.
  
  The above hello.pl is perl script can be run standalone. It is not a CGI perl script since it did not follow the CGI protocol. In order to make that as a CGI perl script, we need to add the following two lines after #!/usr/bin/perl
  
  use CGI qw(:standard);
  print header();
  
  or
  
  print "content-type: text/html\n\n";
  
  Try http://sanluis.uccs.edu/~cs301/cgi-bin/hello.cgi or http://sanluis.uccs.edu/~cs301/cgi-bin/hello2.cgi, src, which implement the above fixes.
  
  
  

Literals

• Numeric literals are specified in any of the customary floating point or integer formats:

    12345
    12345.67
    .23E-10
    0xffff              # hex
    0377                # octal with 255 decimal value.  Each octal digit is 3 bits.
    4_294_967_296       # underline for legibility
• String literals are usually delimited by either single or double quotes. They work much like shell quotes.
• double-quoted string literals are subject to backslash and variable substitution; single-quoted strings are not. For example:

$name = "Perl Scripts";
print '$name';  # print exactly  the 5 characters between '.
print "$name";   # print 12 characters, 'Perl Scripts'  $name is substituted with that.

Variables

• The types of the variables in Perl can be easily recognized by the first character of the name/identifier.
• $ for scalar variable (single value, can contain integer value, real number, or string. Perl is not a strong typed language.)  For example,

$counter, $classname, $semester
• @ for array variable (0-based numberic indexed array with multiple values). For example,

@cookies = ("camel delites",  "peanet butter patties", "shortbread", "thin mints", "peanut butter sandwich");
• indvidual element can be accessed as $cookies[0] ("camel delites"), $cookies[2] ("shortbread").
• The index of the last element in the array can be obtained through $#cookies.
• The length of the array can be obtained by assigning @cookies to a scalar variable.

$length=@cookies;
if (@cookies == 0) { print "cookies array is empty!\n" }
• % for associate array varaible (string indexed array with multiple values). For example,

%nameValuePairs = ("name" => "Edward Chow", "email" =>"chow@cs.uccs.edu", "phone" => "262-3110");
• indvidual element can be accessed as $nameValuePairs{"name"}, $nameValuePairs{"email"}.
• Note that instead of [], {} is used for wrapping around the string index.
• A string index for an associate array is called key, e.g., in the above nameValuePair, name, email, and phone are keys.
• The set of keys in the associate array can be obtained through "keys %nameValuePairs"
• The scalar variables in Perl have the name starting with '$'.
• A typical identifier for scalar variable consists of '$',  followed by letter or '_', followed by letter or '_'  or digits.
• Special Variables:
• $<digits> : These are variables contain the matching results of regular expression execution. e.g., $1 is the first text pattern matched, $3 is the 3rd text pattern matched.  Therefore we typically avoid using these names as identifiers.
• $ARGV[n]: This is nth parameters in the command line (or called the nth elements in the command line argument vector) . n is a number.
• $_:  a variable used to contain the result of the last operation.
• Variable interpolation: This refers to a power feature in construction string literal.  The double quote string can contain variables and their values will be  replaced for the final content of the double quote string. For example,

print  "The first cookie  is $cookies[0].\n";
The string "The first cookie is camel delites" will be printed out followed by end of line.

Operators

• Operator precedence rule: the order of the operators in an expression are evaluated according to their relative

     precedence level (from highest to lowest)
• Perl operators have the following associativity and precedence, listed from highest precedence to lowest.
• Note that all operators borrowed from C keep the same precedence relationship with each other, even where C's precedence is slightly screwy.
• With very few exceptions, these all operate on scalar values only, not array values.

 

associativity	operators	Example
left	terms and list operators (leftward)	print "List of ", $class,  " students\n";  @array=(1, 3, sort 4, 2);
left	->  (infix dereference operator)	print $query->header('text/html');
nonassoc	++    --	$counter++;  print ++($foo = 'Az');      # prints 'Ba'!
right	** (Exponential)	-2**4 is -(2**4), not (-2)**4.
right	! (logical not) ~ (bitwise negate, 1's complement)  \  (create reference variable) unary + and -	if (!$found) { print "not found.\n"; exit 1;} 0666 & ~ 027 is 0640. here it is octal code. $scalarref = \$foo; $arrayref  = \@ARGV; $hashref   = \%ENV; $minus10 = -10;
left	=~   !~	left operand is scalar variable with data, right operand is search pattern, substitution, or translation. $fullname =~ s/,.*//;
left	* / %  x  (repetive operator)	print '-' x 80;             # print row of dashes @ones = (5) x @ones;    # set all elements to 5
left	+ -  .  (concatenate operator)	$filename=$class . $semester . "Photo.html";
left	<< >> (shift operator)	$i= 5 << 2; # equivalent to 5*4
nonassoc	named unary operators, e.g. -e  File exists	print "file exsit" if -e $filename;
nonassoc	< > <= >= (numeric comparsion) lt gt le ge  (stringwise comparison)	$foo=291; $bar=30; if ($foo < $bar) {print "$foo is less than $bar (numeric)\n"; } if ($foo lt $bar) {print "$foo is less than $bar (string)\n";}
nonassoc	== != <=>  eq ne cmp	<=> return -1, 0, 1 for < , =, > cmp return -1, 0, 1 for string comparsion
left	& (bitwise AND)
left	| (bitwise OR) ^ (bitwise XOR)
left	&&	short circuit logical AND
left	||	short circuit logical OR
nonassoc	..  ...	for (101 .. 200) { print; } # print $_ 100 times @alphabet = ('A' .. 'Z');
right	? : (c compact if then else)	($n == 1) ? "_": "s";
right	= +=  -=  *= etc.
left	,  => (digraph, for documenting arguments that come in pairs)	print start_html(-title=>'Update confirmed',                 -bground=>'#ff9977'),
nonassoc	list operators (rightward)	open HANDLE, "filename"         or die "Can't open: $!\n";
right	not	low preference logical not
left	and	low precedence logical AND
left	or xor	low precedence logical XOR

Selection Structures

• if (condition1) {

} elsif (condition2) {
} else {
}
• Unlike C or Java, Perl requires both branches to be wrapped around with { }.

Iterative Structures

• for loop

@names=("chocolate",  "mint", "peanut", "butter",  "mm");
for ($x=0; $x <= $#names; $x++) {
        print "$names[$x]\n";
}
• foreach <scalar looping variable> (<array variable>) {

}
@cookies = sort qw(chocolate mint peanut butter mm);
foreach $type (@cookies) {
  print "cookie type=$type, noOfBoxes=$cookieJug{$type}<BR>";
}
• do until, do while

do {
        $line = <STDIN>;
        ...
    } until $line  eq ".\n";
• while BLOCK , while BLOCK continue BLOCK

for ($i = 1; $i < 10; $i++) {
        ...
}

is the same as this:

$i = 1;
while ($i < 10) {
        ...
} continue {
        $i++;
}

Associate Array and DBM

• initialize associate array.   Use ( followed by key=>value pairs).

@cookies = sort qw(chocolate mint peanut butter mm);
%cookieJug = (chocolate => 10,
              mint => 20,
              peanut => 30,
              butter => 40,
              mm => 50);
foreach $type (@cookies) {
   print "cookie type=$type, noOfBoxes=$cookieJug{$type}\n";
}
• update the associate array element, updateAssocArray.pl

%cookieJug = (chocolate => 10,
              mint => 20,
              peanut => 30,
              butter => 40,
              mm => 50);
print "Enter the cookie shipping just arrived:";
chop($type=<STDIN>);
print "Enter the number of boxes that arrived:";
chop($boxno=<STDIN>);
$boxInStore = $cookieJug{$type};
print "The no. of $type boxes in store before=$boxInStore\n";
$boxInStore += $boxno;
$cookieJug{$type} = $boxInStore;
print "The no. of $type boxes in store after=$boxInStore\n";
• The content of associate array disappears after the Perl script execution. You can save it to a file with file format arrangement, or you can use DBM file that map automatically to an associate array without file conversion.
• Tie associate array with DBM file for data persistence.
• Initializes the associate array bound with DBM file, dbminit.pl

%cookieJugInit = (chocolate => 10,
              mint => 20,
              peanut => 30,
              butter => 40,
              mm => 50);
dbmopen %cookieJug, "cookieData", 0666;  #octal code 0666 for write access
foreach $type (keys %cookieJugInit) {  # keys function return an array of all keys in an assoc array
  $cookieJug{$type} = $cookieJugInit{$type};
  print "$type $cookieJug{$type}\n";
}
dbmclose %cookieJug;
 
• Read associate array bound with DBM file, dbmread.pl

dbmopen %cookieJug, "cookieData", 0444;  # 0444 for read access
print "cookie      noOfboxes\n";
print "=" x 25 , "\n";
foreach $type (keys %cookieJug) {
  $paddedType = pad($type, 14);                  # use subroutine call to pad string  size to 14
  print "$paddedType $cookieJug{$type}\n";
}
dbmclose %cookieJug;

sub pad {
   my ($type, $maxChars) = @_;
#   print "parameter type=$type, maxChars=$maxChars\n";
   $length = length $type;
#   print "no of characters in $type=$length\n";
   for ($i=0; $i<$maxChars-$length; $i++) {
      $type .= " ";  # if use += it will be error
   }
   return $type;
}

• Order processing: Substract value from associate array bound with DBM file, dbmUpdate.pl

print "Enter the order cookie type:";
chop($type=<STDIN>);
print "Enter the order cookie amount:";
chop($amount=<STDIN>);
# check if inventory has enough
dbmopen %cookieJug, "cookieData", 0666;
$inventory = $cookieJug{$type};
if ($inventory < $amount) {
   print "Not in stock! We will order for you. Take about a week.\n";
} else {
   print "In stock! Ship within 24 hours.\n";
   $cookieJug{$type} -= $amount;
}
dbmclose %cookieJug;
• Note that the dbm file created by you during the debugging by default only your login with write access.  To allow web server running as nobody to write, you need to change the access right of the dbm file to 0666.
• Here is an example of using dbm for creating password list and perform password checking.

It consists of two scripts: password.pl and updatePassword.pl
• password.cgi will be accessed as http://sanluis.uccs.edu/~cs301/cgi-bin/password.cgi
  #!/usr/bin/perl
  use CGI qw(:standard);

  print header();
  print start_html(-title=>'Update Password',
    -bgcolor=>'#EDDF99'),
   h1('Update Administrator User Password'),
   hr;
  # retrieve the current password list
  dbmopen %passwordList, "order/password", 0666;
  print start_form(-method=>'POST',
    -action=>'updatePassword.cgi');

  print p("Here is the current list of login and password\n");
  print "<TABLE border=1 BGCOLOR=\"#EDDFC9\">\n";
  print "<TR><TH>Login</TH><TH>Password</TH></TR>\n";
  foreach $login (keys %passwordList) {
    $password = $passwordList{$login};
    print "<TR><TD>$login</TD><TD>$password</TD></TR>\n";
  }
  dbmclose %passwordList;
  print "</TABLE>\n";
  print p("Enter your administrator login and password for updating the password.\n");
  print "<TABLE border=1 BGCOLOR='#93DDDB'>\n";
  print "<TR><TD>Login:</TD><TD><INPUT type=text name=\"adminLogin\"></TD></TR>\n";
  print "<TR><TD>Password:</TD><TD><INPUT type=password name=\"adminPassword\"></TD></TR>\n";
  print "</TABLE>\n";
  print p("Enter the new user login and password.\n");
  print "<TABLE border=1 BGCOLOR='#ff99DB'>\n";
  print "<TR><TD>Login:</TD><TD><INPUT type=text name=\"login\"></TD></TR>\n";
  print "<TR><TD>Password:</TD><TD><INPUT type=password name=\"password\"></TD></TR>\n";
  print "</TABLE>\n";
  print submit(-value=>"Update"), reset, end_form;
  print end_html;
   

• updatePassword.cgi will be called when the submit button is pressed by the user when access the web page created by the password.cgi.

#!/usr/bin/perl
use CGI qw(:standard);

print header();
print start_html(-title=>'Update Password',
  -bgcolor=>'#DD9999'),
 h1('Update Password');
# get the current password
dbmopen %passwordList, "order/password", 0666;
$adminLogin = param('adminLogin');
$adminPassword = param('adminPassword');
if ($adminLogin eq 'root') {
   if ($adminPassword ne 'cs301') {
      print "Illegal login or password.", end_html;
      exit();
   }
} elsif ($passwordList{$adminLogin} ne $adminPassword) {
      print "Illegal login or password.", end_html;
      exit();
}
# legal admin
print "Correct Admin Password!<BR>\nCreating User Login and Password...<BR>\n";
$login = param('login');
$password = param('password');
print "login=$login, password=$password<BR>\n";
$passwordList{$login} = $password;
dbmclose %passwordList;
print "New Login and Password Created!\n", end_html;
 

• If you are concerned about the security, you can use https instead of http. Also it is usally not a good practice to hard coded the password in the script. Since if there is a glitch in server configuration, the intruder may see the source of the script:-)
• Exercise:
  Write the Perl script, dbmadd.cgi,  that allows an authorized person to update inventory when cookies shipping arrives.  Copy the password.cgi and updatePassword.cgi to your cgi-bin directory and create a testing login=cookie and password=monster as an authorized account for us to test your dbmadd.cgi.

 

Subroutine

• syntax:

sub <name of subroutine> {
     .....
}
• Here is an example, a subroutine that pad the string with space characters to make up certain length, the first parameter is the string, the second parameter is the length of final padded string :

sub pad {
   my ($text, $maxChars) = @_;
   $length = length $text;
   for ($i=0; $i<$maxChars-$length; $i++) {
      $text .= " ";  # if use += it will be error
   }
   return $text;
}
• Here is how the subroutine will be called:

$paddedType = pad($type, 14);
• When pad is called, $type and 14 will be put in a special array, @_.
• We can use my($p1, $p2, ..., $pn) = @_; to retrieve individual parameter.
• You can also use $parameter = shift; to get/remove the first parameter from the implicit stack variable @_.

#!/usr/bin/perl
 

$text= 291;

$paddedText = pad($text, 12);

sub pad {
   my $t = shift;
   my $n = shift;
   print "text=$t, length=$n\n";
}

I/O

• Reading from STDIN (console input)

print "Please Enter your name: ";
chop ($name=<STDIN>);  # remove end of line
print "Thanks for making me happy, $name !";
• Reading from a file called shoppingCart.txt

open(IN, "shoppingCart.txt");
while (<IN>) {  # equivalent to $_=<IN>
     print;          # equivalent to print $_;
}                      # loop thru each line of the file
• Append to a file

%cookiePrice = (chocolate => 1.0,
              mint => 2.0,
              peanut => 3.0,
              butter => 4.0,
              mm => 5.0);
print "Enter new item to be added to the shopping card:";
chop ($item=<STDIN>);
print "Enter the quantities:";
chop ($quantity=<STDIN>);
$price = $quantity * $cookiePrice{$item};
open(OUT, ">>shoppingCart.txt";   # >> mean append, > mean write (overwrite)
print OUT "$item $quantity $price\n";
close(OUT);
• Note that to print to a specifc file we  use OUT (the file handle) as the first parameter.

Use $. for the line number.
open IN, "c:\\perl\\sub.pl";

while (<IN>) {
        print "Line number $. is : $_";
}

• Command line processing

#!/usr/bin/perl
$class = $ARGV[0];
$semester = $ARGV[1];
print "command paramerters: \n1st argument is class=$class\n2nd argument is semester=$semester\n";
 
• To print multiple lines of text (use <<string delimiter):
  print <<END;
  You can change the appearance of this page by submitting
  the fill-out form below.  If you return to this page any time
  within 30 days, your preferences will be restored.
  END
      ;

  print <<EOF;
  <html><head><title>$TITLE</title></head>
  <frameset cols="50,50">
  <frame src="$script_name/query" name="query">
  <frame src="$script_name/response" name="response">
  </frameset>
  EOF
      ;

• Common Mistake: Add space after ending string delimiter. There should not be any space after EOF or END.
• Here is an example for adding new user to a dbm file, addpd.pl:

#!/usr/bin/perl
use CGI qw(:standard);

dbmopen %passwordList, "order/password", 0666;
print "current password list\n=============\n";
foreach $login (keys %passwordList) {
  $password = $passwordList{$login};
  print "login=$login  password=$password\n";
}
print "\nenter new login:";
$login = <STDIN>;
chomp $login;
print "enter password:";
$password = <STDIN>;
chomp  $password;
print "login=$login, password=$password\n";
print "=====\n Here is the password list\n";
$passwordList{'root'} = "cs301";
$passwordList{$login} = $password;
dbmclose %passwordList;

dbmopen %pdList, "order/password", 0444;
foreach $login (keys %pdList) {
  $password = $pdList{$login};
  print "login=$login  password=$password\n";
}
dbmclose %pdList;
 

Split

• A powerful  function to separate the fields in a text string based on some delimiters (: , ; , tab, or white space).  The result is an array.
• $line="aanan:PqxLPda:4211:15:Arvin Anan:/users/server/students/aanan:/bin/csh";

@fields = split(/:/, $line);    #  $fields[0]=aanna; $fields[2]=4211
print "The login of  $fields[4]  is $field[0].\n";
#  the $line is an entry in the password file, $field[1] is the encrypted password
• Here is a line of an excel spreadsheet file save in tab delimited text file format.

It contains the full name, login, SS#, grades of hw1, hw2, hw3, and letter grade, separated by tab character.
We use the split to separate them into an array.
$gradeLine="Chow, Edward\tchow\t454000000\t93\t33\t22\tA";
$array=split(/\t/, $gradeLine);
print  "$array[0]  your hw1 grade is $array[3]\n";

Regular Expression for pattern matching

• See summary web page with examples

CGI

• See CGI.htm web page
• password.cgi example
  This simple example has security weak point. It was used by an intruder to deface the web page. Instead of login and password, a javascript and a refresh meta tag was entered as data. When the data are redisplayed, the javascript pops up alert message and the meta tag redirect web page to a nasty one. The quick fix is to limit the length of the input or to detect the <script> or <meta> pattern.
  
  I have disabled the password.cgi by changing its access rights to 600. To demonstrate the security threat, change the access right to 700 and then enter <meta http-equiv="Refresh" content="20; URL=http://sanluis.uccs.edu/~cs301/nasty.html"> into the user login or password field of the web page. You can remove the hash array element by using delete(). rmentry.pl will remove key or value of a hash with <meta pattern.
  password.cgi source, updatePassword.cgi
• improved passwd.cgi example
  passwd.cgi source, updatepasswd.cgi source.
• inventory.cgi example

inventory.cgi source, update.cgi source
• setPrice.cgi example

setPrice.cgi source, updatePrice.cgi
• orderCookie.cgi example

orderCookie.cgi source, processOrder.cgi source, confirmOrder.cgi

Debugging CGI Scripts

• When your CGI scripts contain compiling errors or logic error, you will see messages from web servers indicating

Internal Server Error
 

The server encountered an internal error or misconfiguration and was unable to complete your request.

Please contact the server administrator, root@localhost and inform them of the time the error occurred, and anything you might have
done that may have caused the error.

Premature end of script headers: /users/server/students/www/cgi-bin/cs301/debugipaddr.pl

• To get more detail error message for our CS Unix apache web servers,  login in to the specific machine and
• cd /etc/httpd/logs     (it actually goes to /var/log/httpd.  You will see access_log and errro_log)
• tail -f error_log         (-f mean follow up,  new updates to log, will show up. tail shows last par fo the file)
• You may see something like the following:

[Tue Feb 22 11:26:52 2000] [error] [client 128.198.161.177] Premature end of scr
ipt headers: /users/server/students/www/cgi-bin/cs301/debugipaddr.pl
syntax error at /users/server/students/www/cgi-bin/cs301/debugipaddr.pl line 7,
near "@testdata ("
Execution of /users/server/students/www/cgi-bin/cs301/debugipaddr.pl aborted due
 to compilation errors.
• Note that the syntax error msg may not show up on the error log until next error requests, due to buffering.
• The above syntax error is exactly the same error if you run the script standalone.  You should get rid of  the syntax error by running your scripts standalone before, debugging it through the web access.

Common Programming Mistakes:

• If you miss ending double ",  you will see

Can't find string terminator '"' anywhere before EOF at debugipaddr.pl line 24.
• If at the beginning line you spell the path to perl wrong,  this is what you get

[Tue Feb 22 11:47:19 2000] [error] [client 128.198.161.177] Premature end of scr
ipt headers: /users/server/students/www/cgi-bin/cs301/debugipaddr.pl
[Tue Feb 22 11:47:26 2000] [error] (2)No such file or directory: exec of /users/
server/students/www/cgi-bin/cs301/debugipaddr.pl failed
• Forget to include "print header();" which generates content type metaheader.
  [Tue Feb 22 11:26:52 2000] [error] [client 128.198.161.177] Premature end of scr
  ipt headers: /users/server/students/www/cgi-bin/cs301/debugipaddr.pl
  

Common but Tricky End-of-Line Problem when transfers cgi scripts from Window to Unix:

-----Original Message-----
From: CS301 Web Programming Mail List [mailto:cs301-l@uccs.edu] On Behalf Of Edward Chow
Sent: Sunday, October 03, 2004 12:42 AM
To: CS301 Web Programming Mail List
Subject: [cs301-l]Re: CGI error

Heidi,

This is a tricky but common error!

If you use vi to edit your file, you will find at the bottom of the display there is an indication of "dos" file.
If you ftp the catalog.cgi file from home PC, the dos file will contain additional character at the end of each line.

I use the bigend program I wrote to display the byte content of your catalog.cgi

sanluis.uccs.edu> ~cs520/bin/bigend catalog.cgi
argv=catalog.cgi
LOC 0: 23 21 2f 75 73 72 2f 62 69 6e 2f 70 65 72 6c 0d "#!/usr/bin/perl.p B"
LOC 16: 0a 0d 0a 75 73 65 20 43 47 49 20 71 77 28 3a 73 "...use CGI qw(:sp B

It shows after "perl" which is 70 65 72 6c in hexa decial, your dos file has 0d 0a (carriage return, line feed) for the end of line.
The next line is empty therefore we saw another 0d 0a follows.

If I switch to ~cs301/public_html/cgi-bin/catalog.cgi

sanluis.uccs.edu> ~cs520/bin/bigend catalog.cgi
argv=catalog.cgi
LOC 0: 23 21 2f 75 73 72 2f 62 69 6e 2f 70 65 72 6c 0a "#!/usr/bin/perl.p B"
LOC 16: 75 73 65 20 43 47 49 20 71 77 28 3a 73 74 61 6e "use CGI qw(:stanp B"

You will find there is only one 0a for representing the end of line, not 0d 0a. My catalog.cgi file is created on Unix machine.
The Window and Unix have different ways to representing the end of line.

Why this is causing the problem? Why the CGIWrap complains that it can not find the file or the Perl interpreter?

The reason is that CGIWrap or the web server will treat the catalog.cgi as a Unix file (since the web server is run on a Unix machine) and try to read the first line. It finds the first two characters is #! and understand it needs to find a shell program or a Perl interpreter. The file name of the shell or the Perl interpreter is followed #! Until 0a (line feed). Since your dos catalog.cgi has 0d (carriage return) before 0a. The cgiwrap will consider 0d as the last character of the file name, when it tries to find /usr/bin/perl(0d) with 5 character as its file name, it will not be able find any file there. The perl interpreter in /usr/bin directory only has 4 characters as its file name.

The error message generated by CGIWrap actually is pretty good. It indicates this could be caused by uploading file. But it says control-M was at the end of the first line, instead of 0d for ascii code.

How can we fix the problem?

The best way is to start vi in binary mode and remove the 0d carriage return character of your dos catalog file.

• vi -b catalog.cgi
  here -b stands for starting vim editor in binary mode.
• You should see at end of the first line after "perl" is a ^M character. The equivalent of 0d character.
• Move your curser on top of the ^M character and hit x to remove it.
• Save your file. It should solve the problem.
  

One of the old way I used is to

• Rename your catalog.cgi as catalogDOS.cgi.
• On unix machine, create a new file catalog.cgi
• Enter #!/usr/bin/perl as the first line.
• Read in the content of catalogDOS.cgi, in vi type the following editing command ":r catalogDOS.cgi"
• Delete the 2nd line in the file which is #!/usr/bin/perl with 0d at the end.
• Save the file as catalog.cgi.
  
  

The CGIwrap error message suggests change the ftp transfer mode from binary to ascii. But I found it does not work. The od in the DOS end of line was not removed by ftp.

Edward

-----Original Message-----
From: Hparsaye@aol.com [mailto:Hparsaye@aol.com]
Sent: Saturday, October 02, 2004 9:51 PM
To: chow@cs.uccs.edu
Subject: Re: CGI error

Dr Chow,
I put -- at the end of the first line to see what is going to happen and make sure perl will see my file, but with out it I get this error:

Thanks
CGIWrap encountered an error while attempting to execute this script:
Error Message: No such file or directory
Error Number: 2
This message usually indicates there is a problem with the script itself. Often this indicates either that the #! line of the script is incorrect, or the script was uploaded in binary mode instead of ascii mode. Check to make sure that the script does not have control-M's at the end of every line. That will prevent it from executing. An easy fix that takes care of this most of the time is to put '#!/.../perl --' instead of '#!/.../perl' on the first line of the script. This is typically a problem if the script was edited or uploaded from a DOS/Windows/Macintosh station to a unix based server.
If you are not the owner of this script, please forward this error and the URL that caused it to the script owner. That is often the component in the URL right after /cgiwrap/.