Andre Poenitz wrote: > On Wed, Feb 05, 2003 at 10:47:48AM +0100, Lars Gullik Bjønnes wrote: >> | Is completely independent of the rest of the world, and can be >> | stabilized in five minutes by just switching back to "last known good". >> >> Is that stabilizing? I thought that counted as reverting and ditching. > > The chance that it would have been necessary is about zero. > > Having something to work on would have made me much happier and I > certainly would not have considered leaving lyx development. > >> That was really when we could expect qt frontend to be finished. > > That was after John got distracted. In August it looked like "end of > October" or so. > >> Having the release five weeks after the real code freeze does not >> sound to bad to me. > > Next time please set fixed dates, and if the stuff is not ready, scrap the > release plans and let everybody go on.
We have been talking about using cvs branches for ages but nobody has ever bothered, simply because noone else would bother to check out this branch and either (a) comment or (b) help. It strikes me that one possible solution to this mess is to create a development branch and _force_ everybody to use it. Whenever a particular feature is classed to be stable, we spend some time porting it to head and release it, reset the development branch point and continue. Since most features are driven by one or two people only, this places the onus on them to (1) get their pet feature stable and (2) 'back port' it to the head. The advantage is that they don't have to drag their heels waiting for anyone else to get their own pet project up to speed too. Of course, we don't have to release each time some new, but stable, addition is made to head, but can accumulate several changes. In fact, it sounds a bit like the way JMarc manages his stable branches. What do you think? -- Angus
