DYNAMIC WEBSITE GENERATOR ---------------------------------------------------------------------- by Darren Duncan CGI-WebsiteGenerator is a collection of Perl 5 modules that were developed originally to implement (generate/compile) an unlimited number of unrelated dynamic web-sites at once. The web sites can contain any combination of static pages or self-referencing forms, perform redirection and track their usage. However, each type of component can be used independantly as well, so you can have standalone guest books or survey forms or usage counters. The intent here was to maximize code reuse such that each instance of a web site needs very little of its own Perl code to differentiate it from others. As such, each "site" consists mostly of a set of hierarchically organized data files, such as the htmlized text for each page, settings/configuration files that say things like what questions are asked in forms such as guestbooks, and log files for the likes of guest books or usage count data. Modules are designed such that their behaviour is customized by giving them different data or "preferences", rather than by changing the values of hardcoded constants. Everything is designed with an extensible architecture so third parties can add features to it. My ultimate goal is that website owners can update their sites using their web browsers over a user-friendly interface. To see living examples of these modules in use, check out my web sites at "http://www.DarrenDuncan.net" and "http://www.chronology.net". KEEPING UP TO DATE These modules are constantly under development. You should be able to find the newest versions at my website, "http://www.DarrenDuncan.net", on the page called "Perl Libraries I Made" (name subject to change). They can also be found on CPAN under the author name of "DUNCAND". I am inclined to update the copies on my web site more often, but those intermediate versions are more likely to have been tested less, and the modules may be updated day by day without increasing the version numbers. However, the copies on CPAN are guaranteed to have unique version numbers when the module has changed. REQUIREMENTS All of my modules require Perl 5.004 or newer, even though only some of them need it, in the interest of consistency. These third-party distributions are required by some of mine. If they aren't included in your Perl distribution then you will need to obtain them from CPAN. The versions below were the newest as of April of 2000. Some of these are optional, so take them as your needs require: CGI-FormGenerator-0.9.tar.gz: - by Darren Duncan - these modules are always used: - CGI::WebUserInput 0.9 - CGI::WebUserOutput 0.9 - CGI::HashOfArrays 1.01 - these modules are optionally used: - HTML::TagMaker 1.0 (for anything except redirection headers and usage) - HTML::FormMaker 1.0 (for making input forms and form reports) - CGI::SequentialFile 1.0 (for saving guest book messages in files; form definitions may optionally be stored in this type of file as well) libnet-1.0703.tar.gz: - by Graham Barr - this module is optionally used: - Net::SMTP 2.15 (for sending e-mails of form or usage data) EventCountFile-1.0.tar.gz: - by Darren Duncan - this module is optionally used: - CGI::EventCountFile 1.01 (for saving usage data in files) Older versions of modules that I didn't make may work as well, but I haven't tried using them so do it at your own risk. INSTALLATION I haven't gotten around to doing makefiles, so you will have to install these the old fashioned way, by copying. Any files that you find in "lib" can go in their corresponding locations in the "lib" folder that came with your Perl 5 distribution. Or alternately, put them anywhere you want, but you will need to have that location added to your include path by your main program using something like this: use lib '/users/me/www_files/lib'; Any existing POD is embedded in the modules themselves, so you will have to look there to see it, or run a POD extractor on it yourself, or look at the copies on CPAN that are dutifully extracted and turned to HTML. MODULE LIST Here is a quick list of all the modules in the distribution, and what each is for. CGI::WPM::Globals - Perl module that is used by all subclasses of CGI::WPM::Base for managing global program settings, file system and web site hierarchy contexts, providing environment details, gathering and managing user input, collecting and sending user output, and providing utilities like sending e-mail. CGI::WPM::Base - Perl module that defines the API for subclasses, which are miniature applications called "web page makers", and provides them with a hierarchical environment that handles details for obtaining program settings, resolving file system or web site contexts, obtaining user input, and sending new web pages to the user. CGI::WPM::MultiPage - Perl module that is a subclass of CGI::WPM::Base and resolves navigation for one level in the web site page hierarchy from a parent node to its children, encapsulates and returns its childrens' returned web page components, and can make a navigation bar to child pages. CGI::WPM::Static - Perl module that is a subclass of CGI::WPM::Base and displays a static HTML page. CGI::WPM::MailForm - Perl module that is a subclass of CGI::WPM::Base and implements an e-mail submission form with unlimited questions. CGI::WPM::GuestBook - Perl module that is a subclass of CGI::WPM::Base and implements a complete guest book with unlimited questions that also e-mails submissions to the website owner. CGI::WPM::SegTextDoc - Perl module that is a subclass of CGI::WPM::Base and displays a static text page, which can be in multiple segments. CGI::WPM::Redirect - Perl module that is a subclass of CGI::WPM::Base and issues an HTTP redirection header. CGI::WPM::Usage - Perl module that is a subclass of CGI::WPM::Base and tracks site usage details, as well as e-mail backups of usage counts to the site owner. HOW TO USE THE MODULES First of all, it is always a good idea to skim through the POD for any module in order to learn how to use it. However, the POD for all the CGI::WPM::* modules is incomplete right now. What they do have is Name and Dependencies, as well as a list of their methods and a list of the "preferences" they handle. What they lack is Synopsis, Descriptions, Method and Preference descriptions. These will be added when I have the time. Now, at this time I don't have an exhaustive tutorial saying how to use these modules, but I have included two sample "web sites" in the folder called "demos". You should be able to install the folder anywhere on your web server, and only need to modify a single file in each one to make that site operational, and that file is in the root of each web site folder, default.pl. The "website" folder includes a complete site with static pages, email forms, guest books, and usage tracking both for page hits, incoming, and outgoing visitors. The "redirector" folder only does usage counting, and issues a redirection page; it is intended for sites that have changed addresses so you can figure out which other sites still have to fix their links. In the future I will put up more demos, for example, a guestbook-only site. But you can easily derive such from the "website" folder if you desire. Of course, the number of distinct guest books you can have in a single site is unlimited. The file "default.pl" is the "main program" which calls everything else. This file needs to correspond to a web url on your server in order for a web browser to see it, although it can be run from the command shell. All of the other files can be outside the public html directory. What I do is keep all of the site files together, and put only a link/alias/shortcut to default.pl in my public web directory. Following is a portion of the "default.pl" file: use lib '/users/me/www_files/lib'; # this is optional require CGI::WPM::Globals; require CGI::WPM::Base; my $globals = CGI::WPM::Globals->new(); if( $globals->user_input_param( 'debugging' ) eq 'on' ) { $globals->is_debug( 1 ); $globals->persistant_user_input_param( 'debugging', 1 ); } $globals->site_root_dir( '/users/me/www_files/demos/website' ); $globals->system_path_delimiter( '/' ); $globals->user_vrp( lc( $globals->user_input_param( $globals->vrp_param_name( 'path' ) ) ) ); $globals->current_user_vrp_level( 1 ); $globals->site_title( 'Sample Website By WPM' ); $globals->site_owner_name( 'John Sample' ); $globals->site_owner_email( 'john@sample.net' ); $globals->site_owner_email_vrp( '/mailme' ); # opt web page within the site You need to modify your copy of "default.pl" such that the information above is correct for your server's layout and operating system. Not all of the function calls are required; for example, the last block usually just applies to e-mail. All else I can suggest to figure things out is examine the files in the "demo" folder side-by-side with exploring the generated web site in your browser, and see what affect changing certain things has. The demo site makes use of all the distribution modules in some way or other, but a real site doesn't have to. FEATURES When all of the modules in this set are used together, these are some of the features that programs using them can have: * Fully compliant with the HTML 4.0 standard and Cascading Style Sheets. * Single or multiple pages in each site, which can be organized hierarchically. At each hierarchy level, a naviagation bar can be automatically provided or left out. * Display static pages read from an ordinary HTML document; the source can be a complete HTML document or only that which would go between the 'body' tags. * Provide html forms for sending private e-mails to the site owner, with a visitor-defined option for the sender getting their own copy. The user input is automatically error-checked for completeness and shows the visitor what they did wrong if not. The mail is sent using SMTP directly, so it works on systems that don't have 'sendmail'. The number and field type of questions is unlimited, and the input validity checking is customizable. * Provide guest books with all the features of private mail forms and more. Guest book entries are displayed with the sender's name and time of submission, but their e-mail address is hidden. * Display plain text documents with their formatting intact (using 'pre' tags) along with relevant details like title, author, version and dates created or modified. Long documents can optionally be segmented and displayed as a series of parts, which can be navigated with an automatically provided navigation bar. * Automatically redirect visitors to an external url from an internal one. * Usage of the site can be counted, and for each usage event are the count of total occurances, count of today's occurances, date of first occurance, and date of last occurance. Usage details that can be counted include: which site pages were viewed, which external urls we redirect visitors to, which external urls have a link to us (that were used), and miscellaneous environment details like GET vs POST vs HEAD requests. The details are counted independently of each other. Usage information can be compiled and e-mailed to the owner once per day for backup/report purposes. Referring urls are smartly subgrouped such that search engines and normal web sites appear separately, and also the search keywords are in their own group so it is easy to see at a glance what kinds of searches led visitors to you. * Search and replace of content at site visitation time so that repeating and/or complicated data in static source files can be substituted by other static or dynamic content which changes over time, such as a hyperlink to e-mail the owner (addresses change), or directory url holding graphics. * Generated sites are usable on all visitor computing platforms, no matter what web browsers they have, even if text-only. * The site can be seamlessly moved from one web or physical address or computer platform to another, and nothing will break. I have tested it on UNIX, Linux, Windows, Macintosh successfully. All platform-specific details are confined to a single site-settings file, which is quoted above. Most site details are in other files, which are platform agnostic. The site can be invoked from a command prompt for testing if desired. * The site can be invoked using an optional query parameter that will turn debugging on. The site-program will then know that it is the owner viewing the site, and can do some things differently in accordance. Currently the only difference is that a separate set of usage logs are kept, so normal hit counters aren't racked up when testing something. * The site's support files can all be stored outside the 'public_html' directory, so attempts to view them directly will fail (no corresponding web url). Only the main shell (or link to it) must be in the 'public_html'. * Hooks are provided so that future versions of this program will let the site owner edit the site's content using its own web forms. That is, remote administration. Currently, the only way to edit a site's content is to use ftp/telnet to directly change the site support files. For security purposes, admin_mode can only be invoked by calling a separate module method, which would likely be in a separate shell/main program. * The progam modules are designed in an elegant and extensible format, so additional "plug-in" modules can be made to provide additonal functionality, such as a threaded discussion board, or database-access pages, without changing the other program code. In the future, it should be easy to add features for content searching and reporting, internal site spidering (more efficient), and more. * The modules have extensive error checking and graceful exit routines built-in so that nothing the site visitor does should make the program crash (although the Perl interpreter itself or the web server is a different story). If the program did crash, then the visitor would get a nondescript "server error" message without any idea why it happened. Likewise, many checks are in place to account for problems introduced by the site programmer themselves, such as improperly made settings, nonexistant files, or non-compiling plug-in modules. For better or worse, the program currently gives details on exactly what went wrong, so that the owner who is testing it has an easy time fixing it. * The modules are designed to be compatible with mod_perl, but I haven't tested this possibility yet; I will soon. I say this because I avoid using global variables and I pass all subroutine parameters, and I don't use "main" at all. All file-scoped variables are initialized at compile time (of that file), but some keys in my hashes may not be (they're optional), so hopefully that doesn't cause problems with the "initialize everything" rule. SUPPORT Currently I don't have any support arranged with other people, lists, newsgroups, or otherwise. Feel free to ask me if you can't figure things out on your own, or another person whom you know has used this. AUTHOR Copyright (c) 1999-2000, Darren R. Duncan. All rights reserved. These modules are free software; you can redistribute them and/or modify them under the same terms as Perl itself. However, I do request that their copyright information remains attached to the files. If you modify this module and redistribute a changed version then please attach a note listing the modifications. I am always interested in knowing how my work helps others, so if you put this module to use in any of your own code then please send me the URL. Also, if you make modifications to the module because it doesn't work the way you need, please send me a copy so that I can roll desirable changes into the main release. Address comments, suggestions, and bug reports to B. Share and Enjoy!