72 hours is a guideline. I think it’s reasonable - that’s what most projects use. But if you think it’s not enough in this case, make it 5 days or 7 days. Whatever. Doesn’t take that long to run the rat reports and see if it builds.
The “tested the release candidate” is probably what you’re worried about. What we’ve discussed and decided on is that the “3.0” release is a “technology preview” and shouldn’t be expected to be perfect or fit for any use. The release is simply attesting to the fact that it’s an Apache product, duly licensed and vouched for. Cheers, Greg Trasuk > On Jan 9, 2016, at 2:51 PM, Patricia Shanahan <p...@acm.org> wrote: > > Remember that a PMC member voting +1 is asserting that they have personally > downloaded, built, and tested the release candidate, as well as reviewing its > licensing. > > Do we have three PMC members who can do that within 72 hours? Anybody who > would vote -1 on that schedule? > > (I do not expect to be able to vote in favor in 72 hours from now, but will > not vote against unless someone reports a real problem. Most likely, I'll not > vote at all.) > > Patricia > > On 1/9/2016 11:42 AM, Greg Trasuk wrote: >> >> What I actually meant (sorry for not writing precisely) is why not call the >> 72-hour vote now and release it? >> >> Cheers, >> >> Greg. >>> On Jan 9, 2016, at 12:39 PM, Patricia Shanahan <p...@acm.org> wrote: >>> >>> Now that we have a release candidate (YEAH!), we need to sort out how >>> and when to vote on it. We have two proposals, copied from different >>> e-mails. >>> >>> Peter Firmstone: >>>> Voting on this release will commence in 4 weeks, to allow time for >>>> people to check they can reproduce these artifacts and test their >>>> code and report back with any issues. >>> >>> Greg Trasuk: >>>> Why not just go ahead and call the vote now? Once we have 3 ‘+1’s >>>> saying it meets the license requirements, then we can put it up on >>>> the main page. >>> >>> Each of these has possible problems. >>> >>> Peter's plan takes at least 31 days: 4 weeks followed by a minimum of 72 >>> hours for the vote. That may be longer than necessary. >>> >>> On the other hand, people who succeed in building and testing may be >>> ready to vote sooner than people who have problems, so we could hit >>> three +1 votes early, even if there would also have been three -1 votes. >>> A vote needs to have a definite time period. Also, releases are not >>> vetoed by -1 votes, but if anyone detects a serious problem we need to >>> stop and regroup, even if three PMC members have built and tested >>> successfully. >>> >>> I suggest a two phase process similar to Peter's plan, but without the >>> fixed time frame. Instead, anyone who plans to vote should record their >>> intent here, proceed with building and testing, and report their >>> results. Deal with any issues on a consensus basis - I'm hoping there >>> will be none because the issues have already been discussed. >>> >>> When most PMC members who plan to vote have reported success, we can >>> call an immediate 72 hour (or slightly longer) vote. >>> >>> Patricia >>
