Alexander Wirt wrote: > I'm strongly against that. I want packages properly tested and in your own > interest you should wait for testing migration of your packages. (of course > there can be exception, but not in general). > > I want well tested packages in bpo (which means tested by users, not > uploaders).
Pardon me if this is a stupid question, but, it seems to me that relying on a package propigating to testing doesn't ensure that a build of it for backports is particularly well tested. The two are two entirely different builds targeted at two entirely different systems. I've run unstable and testing on servers for ages. Now that I have some recent months of experience using backports, and a lesser amount of experience providing a couple, my level of trust that a given backport will be well-tested and free of integration problems is somewhere between my trust in testing and my (lack of) trust in unstable. Is anything being done to push that assurance level closer to that of testing, or ideally, closer to that of stable? For example, are there any plans to only allow a package into bpo proper after the *backport* has received testing-like aging in a bpo-unstable? -- see shy jo
signature.asc
Description: Digital signature
