On Thu, May 12, 2005 at 09:23:48PM -0700, Michael G Schwern wrote: > On Fri, May 13, 2005 at 01:27:41PM +1000, [EMAIL PROTECTED] wrote: > > OK, but if we remove that, the Stmt goes to ~70% wich is still > > shockingly low for such an important module. It is also very distressing > > that the Sub column isnt at 100% - why you would go to the effort of > > writing test cases and not at least call every method/function is beyond me. > > This is, after all, the whole point of Phalanx: Find important, undertested > modules and improve their coverage. I hope you're not just now realizing > that some of the most important and popular modules are also the most > undertested? > > As for why no 100% coverage... test cases are often written in response to > bugs and not in an attempt to cover all functionality. You can debate the > merits of either approach, but that's the way some folks do it. > > > -- > Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern > Reality is that which, when you stop believing in it, doesn't go away. > -- Phillip K. Dick
The other portion not taken into consideration by Devel::Cover is the XS code in the module. While the Pod coverage statistics cover this, IIRC, Devel::Cover is only looking at the Perl portion of the code. Depending on how much of this code is covered and the proportion of XS to Perl code, these statistics could go up or down. My gut instinct, however, tells me not much will change. Covering the XS portion of the code with gcov is possible, and Devel::Cover will create all kinds of nice webpages and statistics for you too. Paul Johnson may have this written up somewhere, but, if not, I should really write something up about this since I've used it to determine Perl's test coverage. Steve Peters [EMAIL PROTECTED]
