CPAN Testers 2.0 end-February update and next steps

It’s never great to post an “end-February” report a week into March, but that’s how things are going lately. I’ve been busy with family and work obligations that have meant less CT2.0 hacking. I’m sorry this is coming late, but I hope I will give anyone interested a sense of where things stand.

I should note that the original deadline for finishing CT2.0 was March 1 and clearly we’re not there yet. I’ve discussed the situation with Robert and Ask at the Perl NOC, and they’ve been willing to extend the deadline for cutting off CT1.0 for a while longer. Thank you, Robert and Ask, for your understanding!

Progress in the last couple weeks:

  • I did some alpha testing of the CT2.0 Metabase hosted on Amazon. Based on that, I revised yet again (sigh) the user registration/credentials approach to minimize the hassle for old and new registrants.
  • Florian Ragwitz wrote a Catalyst app to help distribute new credentials files to legacy CT1.0 users. I haven’t deployed it yet (since I now need to regenerate all the credentials), but greatly appreciate his quick turnaround.
  • I implemented the Metabase search capabilities that Barbie will need to update the CPAN Testers statistics database. This will be based on the excellent SQL::Abstract approach to WHERE clause construction.
  • I wrote several helper modules to simplify configuration of a CPAN Testers metabase in preparation for deployment. The first of these has already been uploaded to CPAN: Net::Amazon::Config
  • I finalized the “version 0” API for the Metabase web service and revised the interface between the Catalyst app and the Metabase to reflect the latest changes to the library.

In the last several weeks, members the #catalyst and #moose IRC channels were very, very helpful and patiently answered my many stupid questions. Thank you in particular to confound, perigrin, rafl, rjbs, stevan and t0m.

Coming up:

  • Deploy all my new code onto the server for “beta” testing
  • Release all the code to CPAN that people will need to configure their clients for beta testing
  • Regenerate user profiles and deploy rafl’s app to distribute them to legacy users
  • Whip the server into production shape (e.g. proper boot scripts to auto-start the CT2.0 apps on restart)
  • Get back to work on legacy report migration

It’s at the point now where I suspect the “hard thinking” part is pretty much done and it’s a lot of grotty but straightforward tasks to go. Hopefully, beta testing won’t reveal any major issues and the end of March update will be focused on planning on orderly transition from CT1.0 to CT2.0.