On Monday 04 August 2008 16:33:21 Pavel Sanda wrote: > if you want the comments :)
Sure, that was the whole point of the message. :-) > 1. if there are 'major bugs' which need to be fixed to have 'useful > release' then there is no point of make any tarbal, i.e. to prepare tarbal > only after bugzilla is freed from such bugs. It is a matter of compromise, any release will always have bugs and its priority depends on how hard it hits you. The purpose of releasing a previous tar ball is to guarantee that there are no obvious bugs that have been missed before. > 2. people around the world are very curious :) last time the tarbal was > released and even before announcing it there was bugz request for bumping > new release under gentoo. i consider very bad release policy to have one > day lyx-x.y.z.tar.gz and the other day the same named tarbal with different > contents. > except the problems with hash-checks, which is not our problem after > all, this is calling for lyx-bugzilla entries reported for wrong versions - > exactly the last-time fixes, which could make problem for us. I can release it the tar ball as lyx-1.6.0rc2-svnxxxxx that should avoid this problem. > my opinion is to make any tarbal public (even in devel list) only in case > important bugs were solved and change the tarball only for very critical > things; no way for some docs updates or last time bugfixing. in fact i > would rather see rc2 instead of slighlty changed rc1, but i dont want to > annoy you endlessly :) The problem is not me. :-) The release procedure although it takes some time it is not difficult, the problem is the notion of release candidate. I would be very suspicious about anyone releasing rc20, clearly there is some kind of problem having so many release candidates. :-) > pavel -- José Abílio
