On Sun, Aug 23, 2015 at 8:18 PM, Andrea Pescetti <pesce...@apache.org> wrote:
<snipped some complex procedural discussion> > It is not mandatory, but very useful (and I would > make a recommendation out of it) that when voting on a release one doesn't > simply cast a +1 as such. > > I mean, of course a -1 must always be explained, but a +1 should be > explained too, like this: > "+1 Built source on Windows, checked README files, checked ALv2 headers" > "+1 Did only a cursory review but I trust you guys on the code" > and so on. > > Remember, the PPMC is assumed (whether this is written somewhere or not) to > give a +1 based on (mainly) technical reasons; the IPMC will take this for > granted and (broadly speaking) mainly look for compliance issues. If from > the set of PPMC votes the Release Manager can understand, for example, that > no testing at all was done on Linux, he may decide to extend the VOTE until > Linux gets proper coverage; if the PPMC members do not supply this > information, we can't know what was tested and what not. > > So, Jan's question was not for me, but in terms of the "proper technical > review" it would help to see VOTE e-mails more informative than a simple +1, > so that one can be sure that all areas are covered. > > [Feel free to quote/forward this message in public] > > Regards, > Andrea. This makes me think that perhaps having an official check list to ensure that nothing gets forgotten and to make the splitting of the large task that a release is easy and focus resources more efficiently may be a very useful tool to have. What do other projects do in this regard? G -- Visit my Coding Diary: http://gabriela-gibson.blogspot.com/
