flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Justin Mclean <jus...@classsoftware.com>
Subject Re: [DISCUSS] No RC process for releases
Date Mon, 01 Dec 2014 00:08:35 GMT

Look like we now have two versions of the no RC process. The one Eric describes is not the
same process we used on TourDeFlex. Perhaps this process should be documented somewhere so
we are all on the same page. This will avoid conflict to what people think the process is
and isn't.

> It's time to call the vote when the RM thinks it's time to call the vote.

Which I did (after asking for feedback on the release and waiting a week) and was told I shouldn't
have as it was not the no RC way.

>  By this I mean that any issues raised that are not blocker will be deferred to the next

This is where the issue lies, for instance the 3rd Party stuff in TourDeFlex was a non issue
as far as a couple of PMC members were concerned, but was a blocker to others. A vote was
not allowed to be called until that was resolved as we were using the no RC process.  Spelling
mistakes and other minor issues have been considered blockers in the past. Given the limited
number of people who vote this means the 3 +1's can be very hard to get.

> and that any vetoes will have to have a pretty convincing argument as well as a reason
why it
> wasn't raised as an issue during the

It seems there are always issues that are raised during the RC process, it theory RTC and
the CI test should catch them but it's happened on numerous occasions even where there been
discussion on the list before about the subject eg 3rd party code in TourDeFlex.

View raw message