Following a suggest by Adam, I had a quick look at the code this morning, and have update the CPAN Testers Statistics site already!
Check out the new Failure Counts page for the distributions with the highest number of failure reports (FAIL and UNKNOWN) submitted for them. Only the last uploaded version of a distribution is counted, so may look to update this page on a more regular basis.
I was quite surprised to see some of the distributions near the top of the list, but then again that could just highlight how popular they are and that they tend to get tested on a wider variety of distributions than others.
If anyone has any other suggestions for other stats they'd like to see, let me know.
Re:Not just failures...
Alias on 2007-03-10T18:11:54
Ratio is not a good idea, because ever piece of shit on CPAN will be at the top.
In this case, more fails implies more testing, implies more important.
I'd like to see the ones shows and the ORDER shown to stay, but maybe note total tests and percentage in other columns?
Re:Not just failures...
bart on 2007-03-11T22:28:29
I'm sorry, I don't follow. Are we talking about the same thing? Because I was talking about the number of failure reports on CPAN testers. Whether there have been 20 reports with 2 failures, or 50 with 5 failures, that looks like a comparable result to me.
Perhaps you were talking about the number of tests in a test suite.
