hbo has asked for the wisdom of the Perl Monks concerning the following question:
I have a moderately complex Perl/CGI application that I'm helping to maintain in production. I'm interested in automating complete code coverage testing of this app. I started out writing one test case, using LWP::UserAgent and HTML::Parser. That was sort of fun to do, and the results were OK, but I also got the sinking feeling that complete code coverage was going to take a whole lot of tedious and repetitive programming, or else a lot of very intense work trying to generalize the test cases so they could even be derived from the code itself. possibly.
A third possibility is that there may be something out there in Perl Monk land that does Exactly What I Want. But I'd settle for a nice framework, tuned toward testing CGI apps, that I could use to escape some of the drudgery of generating code coverage of my moderately complicated app.
The Test:: modules are too low level for what I want. CGI::Test is too. What I want at the high level is something to organize the data flow through a CGI-style application. Lower down, It would be cool to find something that analyzed the return values of a GET or POST to determine what the outputs were, and what the next inputs might be. This is exactly the process I followed to automate my single test case, dumping the hidden and input fields from successive posts to figure out what to do next.
Is there anything out there that fits that description?