Despite all the problems I had with my old Linux hard-disk (see this entry, this entry and this entry) I was able to get some hacking done. Among else, I was able to work a bit on Test::Run, which is my improved fork of Test::Harness.
At first I found out that Ovid added some methods to the interface of the TAPx::Parser iterator. So I dumped Test::Run::Iterator (essentially the same as Test::Harness::Iterator) and used the internal TAPx::Parser iterator instead. This worked fine, but after raising this issue in the perl-qa mailing list, Ovid suggested a better way, and so I converted to it.
Making use of TAPx::Parser allowed me to remove Test::Run::Iterator and Test::Run::Point, which were no longer needed. Afterwards, I spent a lot of time refactoring Test::Run::Straps and to a lesser extent Test::Run itself. I converted string errors, to classes. Added some missing functionality that I removed by mistake, and changed the $self->mymethod(%args) calling convention to the $self->mymethod({%args}) calling convention, which is more recommended (by "Perl Best Practices") and other sources.
I also refactored the code of some of the tests. One of them (t/output.t) had a lot of duplicate code, which I abstracted in a convenient function.
Now I got a bit tired of endless refactoring and am instead considering extracting all the STDOUT/STDERR output specific functionality into separate plugins. That and possibly relying on the TAPx::Parser statistics collection functionality for keeping track of some of the statistics. (Assuming it is in fact doable).
I can't promise I'll work on it because the Israeli Welcome-to-Linux series is going to take place soon, and I'm one of the organisers. There's also the website of Hamakor, the Israeli FOSS NPO which I'd like to revamp. In fact, I've been procrastinating working on it by working on Test::Run instead.
