On Tue, Mar 24, 2015 at 4:49 AM, Steffen Planthaber <[email protected]> wrote: > Hi, > > At least for the packages on github, aren't the maintainers those with > write access to the repos?
Nope. Github authorization system is really not fine-grained so we originally decided to have a per-subproject Maintainers team (re-read the page I sent a few days earlier). Meaning that the github access information is unreliable. > I think managing a maintainer list additionally to the information that > is already on github is not mandantory. Also the list WILL get outdated. > Who has write access to packages won't. Of course it will. It's very commont to not revoke access unless there's clear misbehaviour. > Actually, only someone with write access can change the manifest.xml to > update the maintainer itself. I know that having write access is not the > same as maintainer (which could leave rock development) but having write > access is a similar responsibility and will never outdate. This is where you are wrong. It seems that for some people I have made a mistake, but I still am optimistic that we can behave all adult software developers, and realize that the maintainer of a package is - in the end - the one which sets the rule for it. Ergo, if you aren't create a pull request. > I also think it would be rather beneficial to use the manifest.xml ONLY > for dependencies and put extra information in the readme.md that is > automatically displayed on github. Actually, it should be in both. The autoproj-based tooling cannot reliably parse the markdown information. Sylvain _______________________________________________ Rock-dev mailing list [email protected] http://www.dfki.de/mailman/cgi-bin/listinfo/rock-dev
