On 7/26/13 11:58 PM, "Justin Mclean" <jus...@classsoftware.com> wrote:
>Hi, > >> I want to release. I agree that a release is long over due. I think we >> should, in the future, release earlier and more often. >That's what the release voting should encourage - yes you want to release >quality software but it's OK to ship with a bug or two and fix them up >next release. Agreed, except if they are regressions. I hate going backwards and it sort of implies that we don't have our quality control act together. > >IMO None of the current outstanding issues are show stopper/blockers they >effect a small proportion of the the SDK users and/or have work arounds. I still don't understand the workaround for the verify error on DataList. The bug author's workaround doesn't make sense. > >Choices are: >1. Fix the issues (we're still not certain how to fix at least one >issue), make yet another release candidate (not an insignificant effort >btw), call another vote (wait a minimum 3 days to have it pass), have >committers test and vote again. Will those committers who voted before be >willing to put in time and effort to vote again? I will. >2. Make this the release and follow up as quickly as we can with a >4.10.1, fixing those issues and any others that might show up once 4.10.0 >gets into users hands. This seems like more work. > >We do seem to be forgetting that there are a lot of features and bug >fixes in the release that would be good to actually put into people >hands. The mustella tests do pass, and the quality of the 4.10 RC 3 >release is better than 4.9.1, 4.8 or 4.6. > >During the next 3 weeks I'm travelling overseas and attending a >conference so that will also slow the process down if we need to make >another RC or 2. If we revert Ilist, I can have that and the resource module fix in tomorrow. Not sure what to do about the "new project" problem. You are right that the code parses correctly so I'm still digging to find out where it is going wrong. > >Thanks, >Justin
