Le 27/04/15 23:01, Russel Winder a écrit : > On Mon, 2015-04-27 at 19:45 +0200, Emmanuel Lécharny wrote: >> […] >> To be clear (or to clarify what I mean) : >> - ctr and rct are both way to process an update *for committers*. For >> contributors, a review is necessary, of course. I may have made things >> blurry in my previous mails. > > I fear there is a very CVCS definition of "committer" implicit in the > above. No, not at all. 'Committers' in this context is very specific to 'The ASF' : ie those who *can* update the main repository.
We use Git and SVN at The ASF, but only apache 'committers' can update the sources. Non committers may just update a fork of the project. > In DVCS the roles of committer and contributor are clearly distinct. Not at Apache, for legal reasons. > > A committer is a person who enacts a reviewed contribution, and only that. > All contributions should be reviewed, even those of a committer. This is > about roles being distinct from people. A person may be a committer and a > contributor but the workflow of that person as contributor should be > identical to that of Joe or Jane Bloggs providing a contribution. Not at Apache, and it all depends on the project's policy regarding reviews. Re-read what I wrote in my previous mails. > > >> > […] >> Basically, when a non-committer propose a patch, a committer will RTC, >> and for most apache project, when a committer has a patch, then it's a >> CTR except for some critical projects, where another committer has to >> validate the patch (thius this is a RTC) > > This may have been true in the 1990s with SCCS, RCS, and SVN, but this > should not be true in a properly managed project. This remains true when some legal protection is at stake, as at Apache. > All contributions should > be reviewed before commit to the mainline no matter what. Non, that's a project's decision to make. Most of the Apache projects don't review commits, because it's extremelly costly, and we do consider that committers have already careful enough not to break the project. This is also why tests are so critical. But as I already said, the project's policy can be more strict, and enforce a review by a pair. It's a choice. It's not mandatory. > RTC is all there > is, anythong else is hacking of the worst order. Man, you are living in a world of hacking of the worst order projects :-) ! Most of the Apache Projects, and I'm ready to bet #M$ that most of the OSS projects out there are not using RTC... > Obviously here I am > talking about big projects with multiple committers, for one person > projects things are different. And I'm talking about Apache Projects, which are not small...
