CPAN Testers

You might recall we made an announcement a few months ago, regarding a deadline. The deadline, 1st March 2010, was the date we had to switch from SMTP to HTTP for report submissions. Since December there has been a lot of work behind the scenes, and some public changes too. Much of the publicly visible work has been done that can be done, while the remaining backend pieces are completed. David Golden has been posting regarding his progress on the Metabase, which will stand as a core component of CPAN Testers in the future.

David has made a lot of progress in the last few months, taking the Metabase, which was essentially still in the design phase, through to development and is very near deployment. So near in fact that within the next week he hopes to have a few high profile testers using it. While the CPAN Testers Statistics database won't be able to read the reports just yet, they will not be lost. Any reports submitted to the new HTTP gateway will all be mass consumed when all the kinks have finally been ironed out and the report parser is eventually pointed at the Metabase.

So what of the deadline? Well, although we had high hopes to achieve it, we missed it. As CPAN Testers is a completely volunteer effort our time to write and test the new code, has been governed by our free time. However, Robert and Ask have been extremely supportive of us, and having seen the results so far, are willing to keep the SMTP gateway open for a little while longer. That being said, it does require that all current CPAN Testers keep their smokebots throttled so we are not overwhelming the server. With the decrease in submissions over the last few months, Robert and Ask have been able to better manage the server, and we would like to keep it that way. With the headway that we are making, it hopefully won't be too long before we do move over to the HTTP gateway and the Metabase.

With so much concentration on the move to CT2.0, we unfortunately have very little else to report as yet. Although there is the QA Hackathon in Vienna to look forward to. Hopefully by then we'll be able to review the state of play with the CPAN Testers and make any refinements needed, as well as look at any feature enhancements that people might find useful.

One aside to the CT 2.0 work, has been for me to take a look at some of the distributions listed on the CPAN Testers Statistics site, Failure Counts and Failure Percentage tables. Taking some of the distributions in hand, I've been able to submit patches for a few, which included a significant overhaul for one (its currently with the author, who was delighted to have me work on the distribution). If you or your Perl Monger group are interested in learning about writing tests, I can only suggest you take a look at some of the distributions in those tables and see whether you can help fix them. In the vast majority of cases the authors will be very grateful to you for helping them, and you'll also be helping to contribute to the Perl community. I'm looking at adding some more tables to list those distributions submitted in the last 6 months, and possibly in the last month, to hopefully highlight quicker distributions that may appreciate a bit of help.

Last month we had a total of 149 tester addresses submitting reports. The mappings this month included 23 total addresses mapped, of which 14 were for newly identified testers. Again another low mapping month, due to work being done on CPAN Testers 2.0.

Keep watching for updates to CT2.0. Its coming very soon ....

Cross-posted from the CPAN Testers Blog.