[don't cc me, please] On Sat, 2002-07-27 at 04:07, Brian May wrote: > On Fri, Jul 26, 2002 at 11:02:18AM +0200, Adrian 'Dagurashibanipal' von Bidder wrote: > > Then the software shouldn't have been released with woody at all, imho. > > So if you can predict that a future version of the software will make > the current version obsolete you shouldn't release the package in stable? This is bs. Known fundamentally *broken* software doesn't belong into stable. Known absolutely outdated software doesn't belong into stable. Software that only works if major parts of the system are modified is just hard to fit into a distribution properly. To be specific: The discussion here has given me the impression that selinux was severely broken to the point of being unusable without *heavy* manual tweaking. Such things don't belong into stable - people using it are probably comfortable enough rolling it on their own. Virus scanners etc. are outdated very, very fast. Integrating such softwre into stable is not useful (I find the meaning of stable == no upgrades except security and major breakage is useful). (At least) two ways to get around this: - don't package at all. Users shall have to look for themselves - package an installer. This installer could possibly install an additional line into sources.list to a antivirus-software-repository (which will only carry this one program and it's signatures). So stable is stable, but necessary software upgrades can happen. The admin even knows exactly what will be upgraded and does not have unpleasant surprises because suddenly somebody decided that ... should be upgraded. (As I said in my previous mail: everything imho, of course). cheers -- vbi -- secure email with gpg http://fortytwo.ch/gpg
Attachment:
signature.asc
Description: This is a digitally signed message part