On 04/23/2014 11:07 AM, Abigail wrote:
> On Tue, Apr 22, 2014 at 10:41:44AM -0400, Ricardo Signes wrote:>> * Karl Williamson <pub...@khwilliamson.com> [2014-04-21T13:46:05]>>> If there are no objections, I propose to put this patch into 5.20>>>> I tentatively object.>>>>> -from the core in the next stable release series, though we may not. As of>>> +from the core two stable releases after the first deprecation, though we>>>> I guess the question is: do we ever think a breakage is so minor that it needs>> only one deprecated release, but not minor enough that it's "just a bugfix.">> I'm pretty sure that in the past we have broken code which wasn't a> bugfix with *zero* deprecated releases, where the only known broken code> was from obfuscated code.>> I'm wondering whether it wouldn't be better that for any breakage a> judgement call is made on how long it must warn before it's released. IMO,> there isn't much benefit if, say, we can make a big improvement in> performance, only having to wait several years because it may break> someones JAPH.>> Of course, making such a judgement call isn't easy, and can be lead to never> ending discussions.>>
I agree with this. I do think the 2 release guideline should be the
default.