Just a thought ... Provided that you agree that sending emails out makes sense, would it make more sense to send one for fixing the manifest.xml and one for the release candidate instead of together. I could start sending emails for the manifest.xml right now.
Sylvain On Mon, Mar 23, 2015 at 2:05 PM, Sylvain Joyeux <[email protected]> wrote: > Great point, Jakob ! > > Maintainers are identified by the <maintainer></maintainer> field in > manifest.xml. I honestly don't know if that particular bit was > discussed, it is just what seemed logical (at least for me, as usual > YMMV :P), and it is what ROS is using. > > The email(s) I was planning to send deal with the unassigned > maintainers by telling the recipients that they should decide who is > the package's maintainer and update the manifest.xml. We're adults, > people should not arbitrarily assign themselves if they feel they're > not a package's main maintainer. > > The recipients, in case there is no declared maintainer, are: > 1. taken from the <author></author> tags if there was no maintainers tag > 2. and fallback to the 50 last commits in the git commit history if > there was no author tag > > Mails are sent by groups of people, meaning that in the case of 2., > all affected people are going to receive a COMMON email, which allows > them to discuss the assignation of a maintainer more easily. > > In addition, I thought about adding a rock_maintainer tag for the > packages that are not developed within the Rock infrastructure, as the > "go-to" person for that package's integration. > > What I don't like about the spreadsheet thing is that (1) I really > don't want to auto-edit manifest.xml (it loses formatting and > comments) and (2) nobody will want to be tasked with editing > manifest.xml one by one. So ... I thought that distributing the work > is not error-prone but would at least be less painful. > > Sylvain > > On Fri, Mar 20, 2015 at 1:51 PM, Jakob Schwendner > <[email protected]> wrote: >> Hey, >> >> appologies if this was already discussed, but how do we identify who is a >> package maintainer? >> The original author? What about external packages, that are just >> "maintained"? Should we add a maintainer element in the manifest to make it >> superclear? >> Also, sylvain, it appeared you had created a script to extract the >> maintainers. Maybe you could send around the result, and we can start >> assigning maintainership somehow (google docs spreadsheet?) >> >> Cheers, >> >> Jakob >> >> >> _______________________________________________ >> Rock-dev mailing list >> [email protected] >> http://www.dfki.de/mailman/cgi-bin/listinfo/rock-dev _______________________________________________ Rock-dev mailing list [email protected] http://www.dfki.de/mailman/cgi-bin/listinfo/rock-dev
