On Sat, Oct 19, 2002 at 09:37:30PM -0400, Ed Tomlinson wrote: > On October 19, 2002 08:05 pm, Ian Clarke wrote: > > On Sat, Oct 19, 2002 at 07:34:43PM -0400, Ed Tomlinson wrote: > > > The linux kernel used to go directly from pre to release - this caused > > > problems more often than not. Now after the last pre, there are a few > > > release candidates to clean up any lingering nasty problems. Suggest > > > that freenet can benefit from the linux experience. > > > > I would love to know when they made the switch - I suspect it was once > > It happened during the build up to 2.4.0... > > > the Kernel had reached a much greater level of reliability and stability > > than Freenet will for a long time. > > They are two different beasts - there is no need for the same levels with > freenet > > > The problem with having over-strict criteria for a release, is that the > > release won't happen - it will perpetually be two weeks away. I believe > > that right now, Freenet is suitable for a 0.5 release. Is it perfect? > > No. Are there aspects of it that can be improved? Certainly. That is > > why 0.5 won't be our last ever release. > > We are not far apart. In my original post the only difference is that I > suggested 7 days instead of 1 for testing. During that time we get to > fix nasty bugs only. I did not say to restart from zero after each change. This is not what everyone else seems to be saying. And what do you think we've been doing? All the feature work in the pre series has been trivial, and introduced because the Real Bugs weren't showing themselves at the time. > > > It hurts us to set unrealistic expectations for a release - since in the > > build-up to the release, significant enhancements are discouraged. > > Sure. On the other hand a smooth release process help and rushing makes > this hard. > > > Some have suggested branching, but we tried that before, and it turned > > into an absolute mess. > > Again the linux model could help. Once 0.5 is out for a month or two do > not fork. Instead make it more usable. Optimise the code and fix small > bugs. Very few new features should be added. Once you are happy with 0.5 > then fork. > > > On top of all of this, I have yet to see an argument against the fact > > that the current build is more stable than our current stable release. > > That, in-itself, is a good argument for releasing ASAP. Well, the startup time on a multi-gig store with a full routing table is probably _a very long time_ currently... > > Agreed. I just think an extra 6 days will lead to a better tested release, > which will benefit freenet - more people will use it. > > Ed >
-- Matthew Toseland toad at amphibian.dyndns.org amphibian at users.sourceforge.net Freenet/Coldstore open source hacker. Employed full time by Freenet Project Inc. from 11/9/02 to 11/11/02. http://freenetproject.org/ -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20021020/cbf8a055/attachment.pgp>
