Calling `zeroCreateUnchecked` from an F# program - f#

I can see in the FSharp.Core source code, that their library functions often call Microsoft.FSharp.Primitives.Basics.Array.zeroCreateUnchecked count to create a zero initialized array. In order to do some performance comparisons with the core libraries I would like to do the same, but it is a private method that I can't seem to access. I can see the definition of the function is:
let inline zeroCreateUnchecked (count:int) =
(# "newarr !0" type ('T) count : 'T array #)
But that doesn't appear to be valid in a normal F# program either. Is there a way to make this work?

Related

Why does reference equality for functions which are bound to the same variable return false?

let f x = System.Object.ReferenceEquals(x,x)
f id // false
I thought that at first it might because a function could be converted to a closure multiple times, but the above disproves that. Why does that last line return false?
You likely have optimizations turned on. This time it's the opposite problem.
What happens if inlining is turned on?
id will be rewritten to an instance of idFuncClass <: FSharpFunc.
The whole expression will be rewritten to:
Object.ReferenceEquals(new fsharpfun1(), new fsharpfun1())
You can turning off inlining with:
[<MethodImpl(MethodImplOptions.NoInlining)>]
let f x = System.Object.ReferenceEquals(x,x)
You'll find that the comparison works again.
But the bigger take-away is this - comparing two functions in F# is undefined behavior. In fact a function type doesn't even implement equality.
let illegal = id = id //this won't compile
Here's the relevant section in the F# Spec:
6.9.24 Values with Underspecified Object Identity and Type Identity
The CLI and F# support operations that detect object identity — that is, whether two object references refer to the same “physical” object.
For example, System.Object.ReferenceEquals(obj1, obj2) returns true if the two object references refer to the same object. Similarly, GetHashCode() returns a hash code that is partly based on physical object identity ...
The results of these operations are underspecified when used with values of the following F# types:
Function types
Tuple types
Immutable record types
Union types
Boxed immutable value types
For two values of such types, the results of System.Object.ReferenceEquals and
System.Runtime.CompilerServices.RuntimeHelpers.GetHashCode are underspecified; however, the operations terminate and do not raise exceptions.
An implementation of F# is not required to define the results of these
operations for values of these types.
What the spec advises is to treat the actual function-type and its CLR implementations as a black-box.

What's the advantage of having a type to represent a function?

What's the advantage of having a type represent a function?
For example, I have observed the following snippet:
type Soldier = Soldier of PieceProperties
type King = King of PieceProperties
type Crown = Soldier -> King
Is it just to support Partial Application when additional args have yet to be satisfied?
As Fyodor Soikin says in the comments
Same reason you give names to everything else - values, functions,
modules, etc.
In other words, think about programming in assembly which typically does not use types, (yes I am aware of typed assembly) and all of the problems that one can have and then how many of those problems are solved or reduced by adding types.
So before you programmed with a language that supported functions but that used static typing, you typed everything. Now that you are using F# which has static typing and functions, just extend what you have been using typing for but now add the ability to type the functions.
To quote Benjamin C. Pierce from "Types and Programming Languages"
A type system is a tractable syntactic method for proving the absence
of certain program behaviors by classifying phrases according to the
kinds of values they compute.
As noted in "Types and Programming Languages" Section 1.2
What Type Systems Are Good For
Detecting Errors
Abstraction
Documentation
Language Safety
Efficiency
TL;DR
One of the places that I find named type function definitions invaluable is when I am building parser combinators. During the construction of the functions I fully type the functions so that I know what the types are as opposed to what type inferencing will infer they are which might be different than what I want. Since the function types typically have several parameters it is easier to just give the function type a name, and then use that name everywhere it is needed. This also saves time because the function definition is consistent and avoid having to debug an improperly declared function definition; yes I have made mistakes by doing each function type by hand and learned my lesson. Once all of the functions work, I then remove the type definitions from the functions, but leave the type definition as comments so that it makes the code easier to understand.
A side benefit of using the named type definitions is that when creating test cases, the typing rules in the named function will ensure that the data used for the test is of the correct type. This also makes understanding the data for the test much easier to understand when you come back to it after many months.
Another advantage is that using function names makes the code easier to understand because when a person new to the code looks at if for the first time they can spot the consistency of the names. Also if the names are meaningful then it makes understanding the code much easier.
You have to remember that functions are also values in F#. And you can do pretty much the same stuff with them as other types. For example you can have a function that returns other functions. Or you can have a list that stores functions. In these cases it will help if you are explicit about the function signature. The function type definition will help you to constrain on the parameters and return types. Also, you might have a complicated type signature, a type definition will make it more readable. This maybe a bit contrived but you can do fun(ky) stuff like this:
type FuncX = int -> int
type FuncZ = float -> float -> float
let addxy (x:int) :FuncX = (+) x
let subxy :FuncX = (-) x
let addz (x:float) :FuncZ =
fun (x:float) -> (fun y -> x + y)
let listofFunc = [addxy 10;addxy 20; subxy 10]
If you check the type of listofFunc you will see it's FuncX list. Also the :FuncX refers to the return type of the function. But we could you use it as an input type as well:
let compFunc (x:FuncX) (z:FuncX) =
[(x 10);(z 10)]
compFunc (addxy 10) (addxy 20)

Why can't we satisfy F# static member constraints with type extensions?

I'd like to be able to extend types from other libraries with static methods to enable generic arithmetic. Take, for example, the newly minted SIMD-friendly fixed-size VectorN types from Microsoft. They define Zero, they define (+), they define (/), but I can't use Array.average on them because they don't define DivideByInt, which I'd be happy to:
open System.Numerics
type Vector2f with
static member DivideByInt (v:Vector2f) (i:int) = v / Vector2f(single i, single i)
let bigArray : Vector2f[] = readABigFile()
printf "the average is %A" (Array.average bigArray)
But it won't let me compile, complaining
error FS0001: The type 'Vector2f' does not support the operator 'DivideByInt'
Why does this limitation exist in the F# compiler?
(Edit: essentially the same question was asked previously.)
It is not currently possible to define an operator overload in a type extension. There is an F# language user voice item for this (with quite a lot of votes) and so this is something that might change in future versions of F# (I think it would be great addition that fits nicely with the F# design).
If you absolutely need something like this today, you can either create a lightweight wrapper for your type that adds the operators, or you can use a (somewhat scary) trick that lets you hide the standard operator with a new overloaded one. The following question has both examples: Global operator overloading in F#

When should we use FSharpFunc.Adapt?

Looking at the source in FSharp.Core and PowerPack, I see that a lot of higher-order functions that accept a function with two or more parameters use FSharpFunc.Adapt. For example:
let mapi f (arr: ResizeArray<_>) =
let f = FSharpFunc<_,_,_>.Adapt(f)
let len = length arr
let res = new ResizeArray<_>(len)
for i = 0 to len - 1 do
res.Add(f.Invoke(i, arr.[i]))
res
The documentation on FSharpFunc.Adapt is fairly thin. Is this a general best practice that we should be using any time we have a higher-order function with a similar signature? Only if the passed-in function is called multiple times? How much of an optimization is it? Should we be using Adapt everywhere we can, or only rarely?
Thanks for your time.
That's quite interesting! I don't have any official information (and I didn't see this documented anywhere), but here are some thoughts on how the Adapt function might work.
Functions like mapi take curried form of a function, which means that the type of the argument is compiled to something like FSharpFunc<int, FSharpFunc<T, R>>. However, many functions are actually compiled directly as functions of two arguments, so the actual value would typically be FSharpFunc<int, T, R> which inherits from FSharpFunc<int, FSharpFunc<T, R>>.
If you call this function (e.g. f 1 "a") the F# compiler generates something like this:
FSharpFunc<int, string>.InvokeFast<a>(f, 1, "a");
If you look at the InvokeFast function using Reflector, you'll see that it tests if the function is compiled as the optimized version (f :? FSharpFunc<int, T, R>). If yes, then it directly calls Invoke(1, "a") and if not then it needs to make two calls Invoke(1).Invoke("a").
This check is done each time you call a function passed as an argument (it is probably faster to do the check and then use the optimized call, because that's more common).
What the Adapt function does is that it converts any function to FSharpFunc<T1, T2, R> (if the function is not optimized, it creates a wrapper for it, but that's not the case most of the time). The calls to the adapted function will be faster, because they don't need to do the dynamic check every time (the check is done only once inside Adapt).
So, the summary is that Adapt could improve the performance if you're calling a function passed as an argument that takes more than 1 argument a large number of times. As with any optimizations, I wouldn't use this blindly, but it is an interesting thing to be aware of when tuning the performance!
(BTW: Thanks for a very interesting question, I didn't know the compiler does this :-))

Pointer to generic type

In the process of transforming a given efficient pointer-based hash map implementation into a generic hash map implementation, I stumbled across the following problem:
I have a class representing a hash node (the hash map implementation uses a binary tree)
THashNode <KEY_TYPE, VALUE_TYPE> = class
public
Key : KEY_TYPE;
Value : VALUE_TYPE;
Left : THashNode <KEY_TYPE, VALUE_TYPE>;
Right : THashNode <KEY_TYPE, VALUE_TYPE>;
end;
In addition to that there is a function that should return a pointer to a hash node. I wanted to write
PHashNode = ^THashNode <KEY_TYPE, VALUE_TYPE>
but that doesn't compile (';' expected but '<' found).
How can I have a pointer to a generic type?
And adressed to Barry Kelly: if you read this: yes, this is based on your hash map implementation. You haven't written such a generic version of your implementation yourself, have you? That would save me some time :)
Sorry, Smasher. Pointers to open generic types are not supported because generic pointer types are not supported, although it is possible (compiler bug) to create them in certain circumstances (particularly pointers to nested types inside a generic type); this "feature" can't be removed in an update in case we break someone's code. The limitation on generic pointer types ought to be removed in the future, but I can't make promises when.
If the type in question is the one in JclStrHashMap I wrote (or the ancient HashList unit), well, the easiest way to reproduce it would be to change the node type to be a class and pass around any double-pointers as Pointer with appropriate casting. However, if I were writing that unit again today, I would not implement buckets as binary trees. I got the opportunity to write the dictionary in the Generics.Collections unit, though with all the other Delphi compiler work time was too tight before shipping for solid QA, and generic feature support itself was in flux until fairly late.
I would prefer to implement the hash map buckets as one of double-hashing, per-bucket dynamic arrays or linked lists of cells from a contiguous array, whichever came out best from tests using representative data. The logic is that cache miss cost of following links in tree/list ought to dominate any difference in bucket search between tree and list with a good hash function. The current dictionary is implemented as straight linear probing primarily because it was relatively easy to implement and worked with the available set of primitive generic operations.
That said, the binary tree buckets should have been an effective hedge against poor hash functions; if they were balanced binary trees (=> even more modification cost), they would be O(1) on average and O(log n) worst case performance.
To actually answer your question, you can't make a pointer to a generic type, because "generic types" don't exist. You have to make a pointer to a specific type, with the type parameters filled in.
Unfortunately, the compiler doesn't like finding angle brackets after a ^. But it will accept the following:
TGeneric<T> = record
value: T;
end;
TSpecific = TGeneric<string>;
PGeneric = ^TSpecific;
But "PGeneric = ^TGeneric<string>;" gives a compiler error. Sounds like a glitch to me. I'd report that over at QC if I was you.
Why are you trying to make a pointer to an object, anyway? Delphi objects are a reference type, so they're pointers already. You can just cast your object reference to Pointer and you're good.
If Delphi supported generic pointer types at all, it would have to look like this:
type
PHashNode<K, V> = ^THashNode<K, V>;
That is, mention the generic parameters on the left side where you declare the name of the type, and then use those parameters in constructing the type on the right.
However, Delphi does not support that. See QC 66584.
On the other hand, I'd also question the necessity of having a pointer to a class type at all. Generic or not. they are needed only very rarely.
There's a generic hash map called TDictionary in the Generics.Collections unit. Unfortunately, it's badly broken at the moment, but it's apparently going to be fixed in update #3, which is due out within a matter of days, according to Nick Hodges.

Resources