On 9/4/06, b.n. <[EMAIL PROTECTED]> wrote:
I don't get it. A given arch system should be self consistent. An ~arch system in theory should be too, but being testing I understand it can not. The purpose of being on arch should be "having a self consistent system within itself" not "having a mostly working bunch of packages". The few, seldom-used packages could be the *critical* packages for a given user (I think to some scientific packages, for example... not many use them, but they can be the very reason to have Linux for someone) Shouldn't all stable packages being tested with a given compiler before that compiler becomes stable?
In an ideal world, yes. But it isn't an ideal world, and the expectation that nothing in the "stable" tree will ever break is just not something that can be satisfied [1]. Also, the gcc and release enginering teams have stated quite emphatically that they are not going to hold up progress on their projects just because other (typically maintainer-wanted) projects are not keeping up. [2] & [3] There is a debate (argument?, flame war?) going on between devs about exactly how much notice was given in advance of gcc _moving_ to stable, but the package maintainers did have 2 months between gcc 4.1 entering ~arch and it moving to stable to fix their problems and move the fixed versions to stable. So in the end, arch users are in much the same position as ~arch, except hopefully your incidences of breakage are much more rare. And IMO, you also get the right to bitch about it...but only if you also report the problems on bugs.gentoo.org! ;-) And of course, Gentoo comes with a lifetime guarantee of complete satisfaction or your money back. :-P -Richard [1] http://www.mail-archive.com/gentoo-dev%40lists.gentoo.org/msg15036.html [2] http://www.mail-archive.com/gentoo-dev%40lists.gentoo.org/msg15043.html [3] http://www.mail-archive.com/gentoo-dev%40lists.gentoo.org/msg15044.html -- gentoo-user@gentoo.org mailing list
