Steve Langasek <vorlon@debian.org> writes: > On Wed, Aug 01, 2007 at 10:44:00AM +0200, Marc 'HE' Brockschmidt wrote: >> Anyway, now Rperl-lover can upload the package on his own, but as a pure >> perl robot, he is bound to fuck up. After a year, *you* will need to >> kick him to understand how SONAMEs work :) > And yet I'm speaking in favor of the proposal, yes? The only reasonable explanation is that you like to kick maintainers :-P > Getting folks to understand how SONAMEs work is such a high bar that I don't > think it should be a requirement even for DDs, let alone DMs. Well, i was desperately searching for something that you encounter on a regular basis. In reality, you are right - if we would only allow people who understand all technical details to be DDs, the project would be a lot smaller. And have different problems. But there are a lot of basic issues that lead to problems on a regular basis, some of which every maintainer should know about. > Instead, my interest is in improving our toolset so that it can do the > work for me of letting maintainers know, ASAP, when they've done > something wrong with a library. :) Besides the fact that I have always been favor of a mandatory remote stabbing device (if possible directly connected to some QA system), I'm not saying that we should try to solve our quality problems by allowing only perfect packages to be uploaded, but that we should discuss which and how many mistakes we can identify after an upload and what the average input quality should be. > (For one thing, consider that this proposal wouldn't let a DM change > library or dev package names for a transition on their own, so there > would have to be a DD involved as well in the case of a > wrongly-renamed package.) Heh, that's not a good example. The usual mistake is to *not* change the package names when breaking the interface. >> My whole argument boils down to "I don't trust DDs". I would be happy to >> vote in favour of this proposal if the list of packages each DM can >> upload is controlled by a small group of people (the DM keyring >> maintainers) and not a group of ~1000 people. > I don't trust DDs either; I trust the system, because it's essentially the > same system we've used all this time for creating the best Linux distro > around, with just a few parameters tweaked. The problems I fear stem from the fact that the DM proposal *changes* the system we are used to. The advantage of having Debian Maintainers is that they don't need to go through a sponsor every time, in other words reducing the time that needs to be spent by all involved people on a single upload. I don't believe that all of the time saved is spent on quality checks that we will miss in the future, but a part is. That *will* have an effect, the only question is how big the change actually is. Marc -- BOFH #379: We've picked COBOL as the language of choice.
Attachment:
pgpYq84RuwL5J.pgp
Description: PGP signature