elmish/urlparser

Turn URLs into statically-typed data

Elmish-urlParser: urlParser extras for Elmish apps.

Elmish-UrlParser implements parser combinators for turning URLs into strongly-typed data.

For more information see the docs.

Installation

paket add Fable.Elmish.UrlParser

Building

With .NET Core SDK v6:

fotnet fsi build.fsx

urlparser's Issues

[<RequireQualifiedAccess>] on functions

Yall opposed to adding [<RequireQualifiedAccess>] for functions like map, oneOf, parseHash, etc?

Also, any reason as to why so conservative when naming functions like s and i32?

Please don't take this as crapping on the library (it's pretty awesome) but I wonder if adoption is not as big as Feliz.Router... due to the cryptic nature of some of these function names...

And maybe placing </> in a module that one must open kind of like fstoolkit-errorhandling does it...

Again, just observations since I am comparing urlParser (elmish.browser) with Feliz.Router

Documentation/Hint on each function

I think I have suggested renaming s and str to something more userfriendly but got rejected =/
Can we instead provide documentation with /// on each of the functions? I think the documentation in the Elmish.UrlParser github page is very helpful, so maybe adding it to the function would also provide value...

* s combinator for a static string we expect to find in the URL,
* </> combinator to capture a slash in the URL,
* str combinator to extract a string,
* i32 combinator to attempt to parse an int.
* top combinator that takes nothing.
* remaining combinator to get the remaining parts of the URL

I tend to shadow like this to provide clarity:

// combinator for a static string we expect to find in the URL
let matchString = UrlParser.s
// combinator to extract a string
let parseString = UrlParser.str
// combinator to attempt to parse and Int32
let parseInt32 = UrlParser.i32

Fable 4.0 breaks parseParams

Before the Fable 4.0 update we did:

let urlParser location = parsePath pageParser location

but now we have to do

let urlParser (location:Browser.Types.Location) =
    parse pageParser location.pathname (parseParams (location.search.Replace("?","")))

It looks like the ? is no longer handled automatically.

/cc @et1975 @alfonsogarciacaro @MangelMaxime @tforkmann

parseParams doesn't handle empty or null case

Description

I am seeing this line throws when used via Navigation when query string is null. It assumes queryString has some value

urlParser/src/parser.fs

Line 357 in 014ab69

querystring.Substring(1).Split('&')

It just looks wrong to me.

Related information

• elmish version: 4.0
• fable-compiler version: 4.0
• fable-core version: 4.0
• Operating system: macos

Suggestion : parser that takes everything remaining ?

I recently had a case where I needed to consume the remaining elements of an url that might contain slashes. this is the parser I wrote :

    let remaining: Parser<(string -> 'a), 'a> = (fun s ->
        [
          { visited = List.concat [s.visited; s.unvisited]
            unvisited = []
            args = s.args
            value = s.value (String.Join("/", s.unvisited))}
        ])

usage:

type Route =
    | Home
    | Graph of string
// #/graph/lorem/ipsum
oneOf [ map Route.Graph (s "graph" </> remaining)
        map Route.Home top ]

that started with what looked like a bug in the parsing, but was due to the payload containing a /. I thought that might be worth documenting at least, maybe adding to the lib ?