rspeele / fparsec-pipes Goto Github PK

Here is some sample
not sure if I have configured one of the pipes wrong. also can we add a p<string> or other means of capturing a whole line?

Users of my library graphql-net are attempting to deploy to SharePoint and having troubles because FParsec-Pipes is not signed on NuGet:

chkimes/graphql-net#69

You probably know what I mean:

let num: Parser<int,unit> =
    %% +.digits 1 -- +.digits 2 -- +.digits 3 -- +.digits 4 -- +.digits 5 -- +.digits 6 -- +.digits 7
    -%> fun a b c d e f g -> f

I've been studying the library for a while now and though I'd ask how is it possible to get that precise -%> fun a b c d e f g -> f dependent on the parsers used previously without pattern matching on tuples like in Fparsec?

Actually, I was sure being able to take the correct number of arguments at the end as if accumulating them would be something impossible in F#, but you proved me wrong. Of course, I've been studying the code, but it is not really obvious what is happening there and I am not sure how to come up with a minimalist example. I would not mind if you gave me a pointer. Thanks!

Edit: Actually, what the linkBeyond5 example does just dawned on me. It is similar to this:

let a = "asd"
let b = 2
let c = 'a'

let w = a |> (fun a b c -> c)
let e = b |> w
let r = c |> e

But still, the Pipes library takes all the arguments at the end, not at the start. How is that possible?

Edit2: I figured this out too. I asked because I thought I would not be able to do it, but I guess I did.

let a = "asd"
let b = 2
let c = 'a'

let w f = a |> f
let e f = b |> w f
let r f = c |> e f

r (fun a b c -> c)

You just pass in an argument acting as an accumulator. This does not explain everything, but now I understand what Pipes is doing. I feel like I have a firm basis now. Thanks for the novel idea. I'll definitely be using Pipes to write the parser I am about to do.

Yes, a deliberate pun on the #7's title :)

I am wondering why support for the framework had to go with the commits de52172 and 69fd979. FParsec itself still targets as low as net40 (I think F# 4.1 may generate code only for net45 and above, though). It also seems that these changes did not require any APIs beyond 1.6, if I am not mistaken, so 2.0 seems too high for a library.

One of these commits did away with the README.md and LICENSE files by the way. It's not clear if that was intentional.

I was having issues with the intro code in VSCode on Windows with the Ionide plugin.

Specifically:
// Parses count digits and returns the result as an integer. let digits (count : int) = %% +.(qty.[count] * digit) -%> (String >> Int32.Parse)
was complaining about the (String >> Int32.Parse) construct needing a type annotation to allow the String constructor to be resolved.

I ended up (after some assistance on fsharp slack) going with:
let digits (count : int) = %% +.(qty.[count] * digit) -%> (fun (chars:char[]) -> String(chars) |> Int32.Parse )

which worked for me

The included example project, Examples.SQLite, takes 25 seconds to compile on my PC. Most F# projects compile in 1-2 seconds.

This also makes Intellisense extremely sluggish. Finally, third party tools using F# compiler services, such as the F# Power Tools extension for Visual Studio, take ages to perform simple tasks.

I am so sorry to raise an issue. I can't, for the life me me, get even the most basic code to work. I have a horrible feeling that this has nothing to do with your code.

#I @"../packages/fparsec/lib/netstandard1.6"
#r @"fparseccs.dll"
#r @"fparsec.dll"
#I @"../packages/fparsec-pipes/lib/netstandard1.6"
#r @"fparsec-pipes.dll"

open FParsec
open FParsec.Pipes
open System

// Parses `count` digits and returns the result as an integer.
let digits (count : int) =
    %% +.(qty.[count] * digit)
    -|> (String >> Int32.Parse)

and I get this output when I run it in FSI

error FS0193: Could not load file or assembly 'FParsecCS, Version=1.0.3.0, Culture=neutral, PublicKeyToken=null' or one of its dependencies. A strongly-named assembly is required. (Exception from HRESULT: 0x80131044)

Hi,
is it possible to support .NET Standard 2.0?
The project graphql-net depends on this library, i try to add support for .NET Standard 2.0 (see chkimes/graphql-net#34).

Started a PR as well: #6