Size: a a a

ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)

2020 June 05

Oℕ

Oleg ℕizhnik in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Yep, that's an experimental work by @kshelopugin .
Still there will be types that look more like ContextT.
So "context" usually considered as "global semantical property", it isn't considered to very from one abstract method to another
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
I have not delved into tofu too deeply. Was comparing your implementation with Izumi BIO and ZIO. And was interested to see you had some nice abstractions
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Like Env and has context
источник

Oℕ

Oleg ℕizhnik in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Yeah similar abstractions for bifunctors are coming
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Hence my question, I think I have not seen it any fundamental laws that also include environment
источник

λ

λoλcat in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Izumi.BIO inherits all the issues with the cats effect hierarchy
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
And I was wondering if profunctor laws would be suitable somehow. I am a category theory wannabe, 😊
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
λoλcat
Izumi.BIO inherits all the issues with the cats effect hierarchy
Can you elaborate?
источник

Oℕ

Oleg ℕizhnik in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Tofu usually understands "Context" just as "accessible" value.
So consider
case class Foo(
 bar: Bar,
 baz1: Baz,
 baz2: Baz,
)

When you have F HasContext Foo you may also get F HasContext Bar and two different version of F HasContext Baz
In that sense tofu approach is lawless, whatever you want to be provideable that way, you'll get.
Tofu just helps to deliver F[Context] automatically whenever it needed
источник

w

welcometotheclubbudd... in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Oleg ℕizhnik
Tofu usually understands "Context" just as "accessible" value.
So consider
case class Foo(
 bar: Bar,
 baz1: Baz,
 baz2: Baz,
)

When you have F HasContext Foo you may also get F HasContext Bar and two different version of F HasContext Baz
In that sense tofu approach is lawless, whatever you want to be provideable that way, you'll get.
Tofu just helps to deliver F[Context] automatically whenever it needed
20/10!
источник

Oℕ

Oleg ℕizhnik in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Thank you buddy
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Pretty cool
источник

λ

λoλcat in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
rafael saraiva
Can you elaborate?
Too powerfull typeclasses that can do anything
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
λoλcat
Too powerfull typeclasses that can do anything
That's cats effect problems alright. Coupled with the fact that you are bound to throwable as your error type as soon as you get to Sync
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Oleg. So an effect can depend on a context right? Like I need A with B with C
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Is it possible to partially provide the context? Let's say, I have A now, at some point later someone will give you B and C
источник

Oℕ

Oleg ℕizhnik in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Izumi is a great tool by great minds, but when they designed they BIO-TF toolkit, it seems they followed by the law of least surprise, so it really looks like bifunctorialized versions of cats-effect plus nice classes for Concurrent data primitives, which are also included in tofu.
I think we can come up with more principled version of Bifuntor monads and things, that are based on them.
источник

rs

rafael saraiva in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Oleg ℕizhnik
Izumi is a great tool by great minds, but when they designed they BIO-TF toolkit, it seems they followed by the law of least surprise, so it really looks like bifunctorialized versions of cats-effect plus nice classes for Concurrent data primitives, which are also included in tofu.
I think we can come up with more principled version of Bifuntor monads and things, that are based on them.
👏👏👏👏👏👏👏👏👏👏👏👏
источник

λ

λoλdog in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
О
источник

λ

λoλdog in ПОКА ОДЕРСКИ НЕ ВИДИТ (спидран Олега по тофу)
Я думал я чат перепутал
источник