On Sat, Feb 24, 2018 at 10:55 AM, R0b0t1 <r030t1@gmail.com> wrote:

I am a stain upon this mailing list and my opinion is of no
consequence, but I think moving directly to 3.6 would be the best
option. Python 3.5 is a very strange situation and has some half
finished features despite being a full release. It is very likely the
release of 3.6 was rushed to fix some of what was "wrong" in 3.5
(mainly missing async features, but there are many).[1] These new
features are useful enough that they will likely outweigh any
additional work involved in adopting 3.6.

I agree. I feel like 3.5 is half-finished when it comes to its async implementation, and 3.6+ should be the goal. 

Let me put it this way to everyone -- I think we should standardize on 3.6+. It would be good if we could have consensus on that. I am pretty confident that this is the 'right answer'.

As to *when* we do this, that I do not have the answer to. That is why I raised it for consideration by others. 

However, I see lots of benefits of making this switch sooner rather than later. Consider that 2.7 stops being supported upstream in 2 years. 

I look at it this way -- the effort we spend maintaining and working within the code base that offers 2.7 support over the next 2 years could be used *instead* to remove lots of cruft from the code base and standardize on something we know we are going to need to standardize on *anyway*. That puts us 2 years ahead of where we would be if we delay the decision.

People are willing to do the work. Mgorny expressed concern on IRC that this would impact EAPI 7 efforts. I don't think that is the case. If anything, it puts us in a better position. 

In fact, if assurance is needed that EAPI 7 is not impacted, or that we would work extra hard on EAPI 7 to in turn gain support for dropping lower versions of python, I would see that as a possibility -- i.e. "we agree to move to 3.6+ with the understanding that it should not impact EAPI 7 development."

Best,

Daniel