Chapter 19 concentrated on responding to browser requests and producing documents using CGI. This chapter approaches the Web from the other side: instead of responding to a browser, you pretend to be one, generating requests and processing returned documents. We make extensive use of modules to simplify this process because the intricate network protocols and document formats are tricky to get right. By letting existing modules handle the hard parts, you can concentrate on the interesting part—your own program.
The relevant modules can all be found under the following URL:
There you'll find modules for computing credit card checksums, interacting with Netscape or Apache server APIs, processing image maps, validating HTML, and manipulating MIME. The largest and most important modules for this chapter, though, are found in the libwww-perl suite of modules, referred to collectively as LWP. Table 20-1 lists just a few modules included in LWP.
The HTTP:: and LWP:: modules request documents from a server. The LWP::Simple module offers an easy way to fetch a document. However, the module can't access individual components of the HTTP response. For these, use HTTP::Request, HTTP::Response, and LWP::UserAgent. We show both sets of modules in Recipe 20.1, Recipe 20.2, and Recipe 20.10.
Once distributed with LWP, but now in distributions of their own, are the HTML:: modules. These parse HTML. They provide the basis for Recipe 20.5, Recipe 20.4, Recipe 20.6, Recipe 20.3, Recipe 20.7, and the programs htmlsub and hrefsub.
Recipe 20.12 gives a regular expression to decode fields in your web server's log files and shows how to interpret the fields. We use this regular expression and the Logfile::Apache module in Recipe 20.13 to show two ways of summarizing data in web server log files.