https://pulumi.com logo
s

straight-cartoon-74589

09/12/2018, 9:57 PM
Can we expect a go sdk for kube anytime soon?
c

creamy-potato-29402

09/12/2018, 9:58 PM
It is not slated for next release, but if there was a huge amount of demand, that might change the ship schedule. For me personally? I would rather wait for Go2, as it will be hard IMO to deliver as good an experience as we have with JavaScript and TypeScript.
Oh wait, we might be doing it over the next 2 releases.
cc @incalculable-sundown-82514
s

straight-cartoon-74589

09/12/2018, 9:59 PM
The go snippet really grabbed our attention on the website. Was disappointed to see it missing from kube. But typescript is certainly not a deal breaker.
i

incalculable-sundown-82514

09/12/2018, 10:00 PM
Go is on the schedule, but not for another month at a minimum with the current plans. There is a lot of engineering work to be done to make it as good an experience as we have with TypeScript today.
c

creamy-potato-29402

09/12/2018, 10:00 PM
Yeah, my main concern is: can we deliver a not terrible experience in Go.
I don’t want people to try it and think it sucks.
s

stocky-spoon-28903

09/12/2018, 10:12 PM
Anecdotally I know of people who have tried Pulumi with Go and not seen the big deal. When pushed to try TypeScript instead they suddenly “get it”
c

creamy-potato-29402

09/12/2018, 10:13 PM
I believe that will be the common case until Go 2.
Personally.
Like all PL opinions this is 100% unscientific.
s

stocky-spoon-28903

09/12/2018, 10:14 PM
Hah yes. It does seem that the lack of sum types is a big problem for Go and the style required by Pulumi
c

creamy-potato-29402

09/12/2018, 10:15 PM
That’s exactly right. Actually for better or worse, I believe we are basically only theoretically multi-language. Our TS support is by far the most mature and (IMO) most compelling, and I can’t see how that will change.
s

stocky-spoon-28903

09/12/2018, 10:16 PM
I think there are other languages with a lot of potential: namely F# and Kotlin
c

creamy-potato-29402

09/12/2018, 10:16 PM
Maybe when Go2 ships, that will change, and that would be nice.
ha ha as a person who wrote a Python 3 compiler in Haskell for fun, I actually agree with that, but I don’t think people will use it. 🙂
s

stocky-spoon-28903

09/12/2018, 10:20 PM
Hmmm, Haskell support might be interesting 🤔
c

creamy-potato-29402

09/12/2018, 10:21 PM
It would be pretty cool IMO.
It would be more awkward in some ways for k8s, as record literals are more awkward than JSON literals in TS.
There’s also a bunch of stuff that is better suited to imperative languages, like shelling out
s

stocky-spoon-28903

09/12/2018, 10:23 PM
I can see that. TypeScript has a nice tradeoff between rigour and flexibility that this kind of thing wants
c

creamy-potato-29402

09/12/2018, 10:24 PM
The biggest advantage is real pattern matching, which makes it far easier to write “verbs” like “expose this pod” and stuff
Haskell is basically a DSL for tree-walking.
I’m not sure there is a really useful core set of combinators, so I’m not sure how much mileage you’d get out of monads.
s

straight-cartoon-74589

09/12/2018, 10:28 PM
I almost feel like having ts really solid with lots of examples would be better than lots of languages with partial engine support
c

creamy-potato-29402

09/12/2018, 10:28 PM
I completely agree.
s

stocky-spoon-28903

09/12/2018, 10:29 PM
I tend to agree also, especially given that there are lots of helper libraries and `ComponentResource`s etc
The one I really want to do (and have in progress) is HCL however, so that Terraform configuration can be run unmodfied
c

creamy-potato-29402

09/12/2018, 10:34 PM
cc @microscopic-florist-22719
m

microscopic-florist-22719

09/12/2018, 10:34 PM
Yep, already aware 🙂
c

creamy-potato-29402

09/12/2018, 10:34 PM
lol
@stocky-spoon-28903 I was thinking of defining a set of Babel extensions that restrict JS to a customizable subset (e.g., conditional or not, loops or not). That way, opsy people can start with no-code and flip the switch if they ever need to.