HTML::EmbperlObject - Extends HTML::Embperl for building whole website with reusable components and objects

    <Location /foo>
        PerlSetEnv EMBPERL_OBJECT_BASE base.htm
        PerlSetEnv EMBPERL_FILESMATCH "\.htm.?|\.epl$"
        SetHandler perl-script
        PerlHandler HTML::EmbperlObject 
        Options ExecCGI
    </Location>

HTML::EmbperlObject allows you to build object-oriented (OO) websites using HTML components which implement inheritance via subdirectories. This enables elegant architectures and encourages code reuse. The use of inheritance also enables a website-wide look and feel to be specified in a single HTML file, which is then used as a template for every other page on the site. This template can include other modules which can be overridden in subdirectories; even the template itself can be overridden. In a nutshell, EmbperlObject makes the design of large websites much more intuitive, allowing object-oriented concepts to be utilised to the fullest while staying within the rapid application development model of Perl and HTML.

HTML::EmbperlObject is basicly a mod_perl handler or could be invoked offline and helps you to build a whole page out of smaller parts. Basicly it does the following:

When a request comes in, a page, which name is specified by EMBPERL_OBJECT_BASE, is searched in the same directory as the requested page. If the pages isn't found, EmbperlObject walking up the directory tree until it finds the page, or it reaches CWDocumentRoot or the directory specified by EMBPERL_OBJECT_STOPDIR.

This page is then called as frame for building the real page. Addtionaly EmbperlObject sets the search path to contain all directories it had to walk before finding that page. If EMBPERL_OBJECT_STOPDIR is set the path contains all directories up to the in EMBPERL_OBJECT_STOPDIR specified one.

This frame page can now include other pages, using the CWHTML::Embperl::Execute method. Because the search path is set by EmbperlObject the included files are searched in the directories starting at the directory of the original request walking up thru the directory which contains the base page. This means that you can have common files, like header, footer etc. in the base directory and override them as necessary in the subdirectory.

To include the original requested file, you need to call CWExecute with a CW'*' as filename. To call the the same file, but in an upper directory you can use the special shortcut CW../*.

Additionaly EmbperlObject sets up a inherence hierachie for you: The requested page inherit from the base page and the base page inherit from a class which could be specified by CWEMBPERL_OBJECT_HANDLER_CLASS, or if CWEMBPERL_OBJECT_HANDLER_CLASS is not set, from CWHTML::Embperl::Req. That allows you to define methods in base page and overwrite them as neccessary in the original requested files. For this purpose a request object, which is blessed into the package of the requested page, is given as first parameter to each page (in CW$_[0]). Because this request object is a hashref, you can also use it to store additional data, which should be available in all components. Embperl does not use this hash itself, so you are free to store whatever you want. Methods can be ordinary Perl sub's (defined with [! sub foo { ... } !] ) or Embperl sub's (defined with [$sub foo $] .... [$endsub $]) .

The runtime configuration is done by setting environment variables, in your web server's configuration file. Basicly the configuration is the same as for normal Embperl. All Embperl configuration directives also applies to EmbperlObject. There are a few addtional configuration directives listed below. Addtionaly you have to set the CWPerlHandler to CWHTML::EmbperlObject when running under mod_perl or use CWepocgi.pl instead of CWembpcgi.pl when running as CGI Script.

Perl regex which files should be ignored by EmbperlObject

Perl regex which files should be processed by EmbperlObject

Name of the base page to search for

Directory where to stop searching for the base page

Additional directories where to search for pages. Directories are separated by CW; (on Unix CW: works also). This path is always appended to the searchpath.

If the requested file is not found the file given by CWEMBPERL_OBJECT_FALLBACK is displayed instead. If CWEMBPERL_OBJECT_FALLBACK isn't set a staus 404, NOT_FOUND is returned as usual. If the fileame given in CWEMBPERL_OBJECT_FALLBACK doesn't contain a path, it is searched thru the same directories as CWEMBPERL_OBJECT_BASE.

If you specify this call the template base and the requested page inherit all methods from this class. This class must contain CWHTML::Embperl::Req in his CW@ISA array.

You can use EmbperlObject also offline. You can do this by calling the function CWHTML::EmbperlObject::Execute. CWExecute takes a hashref as argument, which can contains the same parameters as the CWHTML::Embperl::Execute function. Additionaly you may specify the following parameters:

object_base

same as CW$ENV{EMBPERL_OBJECT_BASE}

object_addpath

same as CW$ENV{EMBPERL_OBJECT_ADDPATH}

object_stopdir

same as CW$ENV{EMBPERL_OBJECT_STOPDIR}

object_fallback

same as CW$ENV{EMBPERL_OBJECT_FALLBACK}

object_handler_class

same as CW$ENV{EMBPERL_OBJECT_HANDLER_CLASS}

See also the CWobject and CWisa parameters in Embperl's Execute function, on how to setup additional inherence and how to create Perl objects out of Embperl pages.

With the following setup:

 <Location /foo>
    PerlSetEnv EMBPERL_OBJECT_BASE base.htm
    PerlSetEnv EMBPERL_FILESMATCH "\.htm.?|\.epl$"
    SetHandler perl-script
    PerlHandler HTML::EmbperlObject 
    Options ExecCGI
 </Location>

Directory Layout:

 /foo/base.htm
 /foo/head.htm
 /foo/foot.htm
 /foo/page1.htm
 /foo/sub/head.htm
 /foo/sub/page2.htm

/foo/base.htm:

 <html>
 <head>
 <title>Example</title>
 </head>
 <body>
 [- Execute ('head.htm') -]
 [- Execute ('*') -]
 [- Execute ('foot.htm') -]
 </body>
 </html>

/foo/head.htm:

 <h1>head from foo</h1>

/foo/sub/head.htm:

 <h1>another head from sub</h1>

/foo/foot.htm:

 <hr> Footer <hr>

/foo/page1.htm:

 PAGE 1

/foo/sub/page2.htm:

 PAGE 2

/foo/sub/index.htm:

 Index of /foo/sub

If you now request http://host/foo/page1.htm you will get the following page

 <html>
 <head>
 <title>Example</title>
 </head>
 <body>
 <h1>head from foo</h1>
 PAGE 1
 <hr> Footer <hr>
 </body>
 </html>

If you now request http://host/foo/sub/page2.htm you will get the following page

 <html>
 <head>
 <title>Example</title>
 </head>
 <body>
 <h1>another head from sub</h1>
 PAGE 2
 <hr> Footer <hr>
 </body>
 </html>

If you now request http://host/foo/sub/ you will get the following page

 <html>
 <head>
 <title>Example</title>
 </head>
 <body>
 <h1>another head from sub</h1>
 Index of /foo/sub
 <hr> Footer <hr>
 </body>
 </html>

(Everything not given here is the same as in the example above)

/foo/base.htm:

 [!

 sub new
    {
    my $self = shift ;

    # here we attach some data to the request object
    $self -> {fontsize} = 3 ;
    }

 # Here we give a default title
 sub title { 'Title not given' } ;

 !]

 [-

 # get the request object of the current request
 $req = shift ;

 # here we call the method new
 $req -> new ;

 -]

 <html>
 <head>
 <title>[+ $req -> title +]</title>
 </head>
 <body>
 [- Execute ('head.htm') -]
 [- Execute ('*') -]
 [- Execute ('foot.htm') -]
 </body>
 </html>

/foo/head.htm:

 [# 
    here we use the fontsize
    Note that 
      $foo = $_[0] 
    is the same as writing 
      $foo = shift  
 #]

 <font size=[+ $_[0] -> {fontsize} +]>header</font>

/foo/sub/page2.htm:

 [!

 sub new
    {
    my $self = shift ;

    # here we overwrite the new method form base.htm
    $self -> {fontsize} = 5 ;
    }

 # Here we overwrite the default title
 sub title { 'Title form page 2' } ;

 !]

 PAGE 2

G. Richter (richter@dev.ecos.de)

perl(1), HTML::Embperl, mod_perl, Apache httpd