On 09/15/2011 07:40 PM, Carl Trieloff wrote:
On 09/15/2011 12:17 PM, Itamar Heim wrote:
- need to decide if patch must be verified before submitting, or verification could be done as part of patch review (for example, gerrit supports verification as part of flow. Some projects may prefer this approach). - (using gerrit terminology) anyone can +1 a patch. only maintainer can +2 the patch. a +1 would help a maintainer do the +2. - going back to my earlier point on committer != maintainer: in a large project, we'd still like committer to have right to commit to entire git to make things simple, but may have maintainers (or call them sub maintainers) per module in the project. A committer would expect (trust would suffice I hope) the relevant maintainer[1] to give the +2 before committing the patch[2]
Itamar,
Do you want to explicitly add the +2 for projects that use gerrit? I can do that.
I don't think we should use gerrit terms, gerrit is only the tool we use to apply the guidelines we are phrasing now. I think we do need to add the trust model Itamar mentioned in his mail. Livnat
Carl.