On Tue, Jan 27, 2009 at 1:24 AM, Eric Wilhelm <[email protected]>wrote:
> > Well, I'm not getting anything done talking about it, but I'm not > convinced that it will get easier to setup solely by putting it on the > cpan. Nonetheless, I can certainly add it all to my stack of things > which need to get polished and shipped. > You said the thing that was holding you up on releasing it to CPAN was feedback on what needs to be documented. And I gave you some. If the real problem is the lack of round tuits, I empathize. My list of "wish I had time to do X" is long too. That would be nice to avoid. If you want to try to replicate Ken's > setup, maybe that works. > Ken -- if you're reading this, could you please put Module::Release::KWILLIAMS in devtools/lib? > In any case, this seems a bit of a red herring. Shipping the release > takes me one command and less than 5 minutes. Knowing that M::B is > ready to ship takes quite a bit longer. Setting up your machine with > CPDK might take less time than the discussion if you tell me where you > get stuck. > (sigh) Yes -- and if someone *else* has the time to figure out if M::B is ready to ship, they don't have your one command and 5 minutes without a lot of yak shaving. Thus we risk a bottleneck on the dev release cycle which means it takes longer to get feedback on changes, releases have more accumulated changes in them when we do release so it's harder to isolate which changes cause further problems and so on. That's my point. I'm not critiquing your release tool as a useful tool -- apart from the feedback on documentation needs you asked for -- but I am arguing it would be good to be able to spread the work around if that can help speed up the release cycle. So "in any case" -- I've added some fixes for the Windows-loaded-system-virus-checker heisenbug and added some diagnostics to ensure_blib so we can figure out why the heck we're getting blib failures on a number of systems. Could you please take the 5 minutes to push out an alpha so we can get CPAN Testers to beat on it and get some idea what's behind the ensure_blib failures that we've been getting? Thank you, David
