On Fri, Feb 06, 2004 at 09:38:24PM -0500, sungo wrote: > > We of the mysterious irc have been pondering a 1.0 release of poe for a > while. Its good for marketing if nothing else. There's even a wiki page > (http://poe.perl.org/?V1.0-todos) documenting our thoughts on > requirements. > > Well, we have some exception support. Scheduled deprecations have mostly > been carried out. Signal reform is done. Threads are in sky's capable > hands and poed hasn't seen a patch for about a year now.
Signal reform is not quite done. We're still dispatching _signal, albeit with great reluctance. http://poe.perl.org/?POE_RFCs/Signal_Reforms While poed seems dead, the need for better documentation is still real. > So, i say its time for a 1.0 release of POE. the api hasn't > substantially changed in a good long while and there are no real user > impacting overhauls scheduled. To me, a 1.0 release implies a certain measure of spit and polish that the documentation (and perhaps the installer) lack. Perhaps addressing these things is worth a 1/10 or greater jump in the version number. This conflicts with the forking of 0.xx with 1.xx. We don't leave much room for the 0 series if we chew it up with large revision jumps. > This gives us a chance to do something else we've been bantering about > for a while. Drop support for any perl versions earlier than 5.6.1. > Before I discuss reasons for dropping that support, I want to detail > version numbers. 1.0+ would carry on, not worrying about supporting > earlier perls. 0.x would be branched in cvs such that 0.x releases can > continue if some intrepid soul wishes to become the maintainer of POE > for pre5.6.1 versions of POE. Maybe the backward-looking version can be 1.00 and the forward-looking release 2.00. Then we can chew up the 0 series with reasonable large updates and still leave the pre-5.6.1 maintainer significant room to maneuver. > There is a lot of code that can be improved inside of POE if we only > support perl 5.6.1 and above. All POE::Preprocessor usage can go away. > Unicode support can be firmed up. Test::More can be assumed. There will > be no more need for two makefiles and diverging build processes. There > are other things that i'm forgetting but that's the gist. We can also use newer, more efficient Perl idioms. I'm sure we can squeeze a few CPU percent out of the overhaul. > So, there are my thoughts. I would like comments, thoughts, > suggestions, etc etc. Silence is assumed assent. So speak up if you have > objections, concerns, etc. What will the neighbors think?! -- Rocco Caputo - [EMAIL PROTECTED] - http://poe.perl.org/
