1.0 Feature Freeze Dec 15th


It’s getting real, folks! 0.7/1.0 feature freeze has been set for Dec 15, 2017 AoE (“Anywhere on Earth” time). That means that if there’s a feature you want in 1.0 it needs to be in by then or it’ll have to wait for 1.x (or 2.0 if it’s breaking). The current list of issues that are considered blocking can be found on the milestone overview. However, since the feature freeze is approaching we’ll be increasingly aggressive about kicking things off the milestone that can be done in a non-breaking fashion after 1.0 (even if that means leaving ugly old APIs around). It’s also unlikely that new things are going to get added to the milestone at this point. However the usual rules apply and even if it’s not on the milestone, if you can get it done before the feature freeze (or convince somebody else to get it done for you), it’ll be in the release.

After feature freeze, we’ll put out the 0.7-alpha release, which will be a good time for package authors to check compatibility with 0.7 and upgrade their packages, fix deprecations, etc. Hopefully femtocleaner will make that process somewhat simpler this time around.

If you’re assigned to any of the issues currently on the milestone, please take a look and see if you’ll be able to get your changes in. If not, please mention so on the issue to see if somebody else can pick it up in time.

The 0.6 release cycle was quite long. We don’t intend to repeat that for 1.0 ;).

Happy Thanksgiving (for U.S. folks)!

Why do `A_mul_Bc` and `A_mul_Bt` belong to `Base`, while `A_mul_B!` lives in Base.LinAlg?
Calculating cos(0.9528) in BigFloat

This is excellent news!

I have been wondering whether the lingering 0.5.x milestones are still of relevance. Some seem to be, although lots of, e.g., the doc issues should be 1.x.


Approximately how long is the transition from 0.7-alpha to 1.0 planned?


The plan for 0.6 was 2 months and actually it lasted 6 1/2 months. I’d guess this time the plan will be again 2 months …




To be fair, 0.6 had some major changes (mostly the type system, but a lot of others that required careful refinement of an API).


Just browse the label ‘breaking’ on the 1.0 milestone




Maybe this should be stickied to the top of the discourse page?



Will Pkg3 be a part of 0.7? or 1.0?


Yes. There’s a bunch of work still to do on that as well, but it’ll be part of 0.7/1.0. When/how the integration process works is still to be worked out.


Is it just me, or does it seem like we’re jumping the gun with the feature freeze here? Enormous changes like Pkg3 haven’t landed yet…


@stevengj I agree with you, for example I have an idea for a pull request on Pkg3, but it might take 1~2 weeks until I can prioritize actually working on it and taking a look. However, I’m neutral, the date is up to you

Gotta say Dec. 24/25th or New Years would make an interesting date though.

Maybe a pre-alpha on Dec. 15th for Pkg testing purposes.


Does the feature freeze include packages in stdlib?


What is your definition of ‘stdlib’ ?


Presumably https://github.com/JuliaLang/julia/tree/master/stdlib


Yeah, one of the efforts behind 0.7 has been moving functionality out of Base and into a standard library of modules that will be shipped with Julia.


That’s a good point, even the package interaction issue seems pretty big on its own. Maybe the plan is to do all of this in a non-breaking way after 1.0?


I am not a developer, I have doubt whether the goals can be achieved. Look at https://github.com/JuliaLang/julia/milestones, in the list of issues left, there are 4, 6, 9, 16 issues opened in 2013, 2014, 2015, 2016 respectively. If those issues could be resolved in this 20 days, it should have been resolved in the past three years.