I am new to ELM and I want to create a modal without the use of any libraries such as Bootstrap or ELM-UI. I found this simple example online which is also using JSON Decode. Is there a possibility to have the modal work simply without any framework/library and JSON Decode? How can I modify the code to simply get a working modal?
module Main exposing (main)
import Browser
import Html exposing (Html, Attribute, button, div, span, text)
import Html.Events exposing (onClick, on)
import Html.Attributes exposing (class, style)
import Json.Decode as Decode
type alias Model =
{ isVisible : Bool, count : Int }
initialModel : Model
initialModel =
{ isVisible = False, count = 0 }
type Msg
= Show
| Hide
| Increment
| Decrement
update : Msg -> Model -> Model
update msg model =
case msg of
Show ->
{ model | isVisible = True }
Hide ->
{ model | isVisible = False }
Increment ->
{ model | count = model.count + 1 }
Decrement ->
{ model | count = model.count - 1 }
view : Model -> Html Msg
view model =
div []
[ button [ onClick Show ] [ text "Show!" ]
, if model.isVisible then
div
([ class dialogContainerClass
, on "click" (containerClickDecoder Hide)
]
++ dialogContainerStyle
)
[ div dialogContentStyle
[ span [] [ text "Click anywhere outside this dialog to close it!" ]
, span [] [ text "Clicking on anything inside of this dialog works as normal." ]
, div []
[ button [ onClick Decrement ] [ text "-" ]
, text (String.fromInt model.count)
, button [ onClick Increment ] [ text "+" ]
]
]
]
else
div [] []
]
dialogContainerClass : String
dialogContainerClass = "dialog-container-class"
containerClickDecoder : msg -> Decode.Decoder msg
containerClickDecoder closeMsg =
Decode.at [ "target", "className" ] Decode.string
|> Decode.andThen
(\c ->
if String.contains dialogContainerClass c then
Decode.succeed closeMsg
else
Decode.fail "ignoring"
)
dialogContainerStyle : List (Attribute msg)
dialogContainerStyle =
[ style "position" "absolute"
, style "top" "0"
, style "bottom" "0"
, style "right" "0"
, style "left" "0"
, style "display" "flex"
, style "align-items" "center"
, style "justify-content" "center"
, style "background-color" "rgba(33, 43, 54, 0.4)"
]
dialogContentStyle : List (Attribute msg)
dialogContentStyle =
[ style "border-style" "solid"
, style "border-radius" "3px"
, style "border-color" "white"
, style "background-color" "white"
, style "height" "120px"
, style "width" "440px"
, style "display" "flex"
, style "flex-direction" "column"
, style "align-items" "center"
, style "justify-content" "center"
]
main : Program () Model Msg
main =
Browser.sandbox
{ init = initialModel
, view = view
, update = update
}
If I understand your question correctly, the problem you're trying to solve is clicking outside the modal to close it. Decoding the event object to get information about the DOM is a bit of a hack in Elm – I think you're right to try to avoid it, unless necessary. One way to achieve the same thing is to add a click event handler with stop propagation to your modal contents – this stops the click event from firing on the container when it originates from within the modal.
I've put your example code in an Ellie and made some small changes: https://ellie-app.com/b9gDPHgtz2ca1
This solution uses Html.Events.stopPropagationOn, which is like on but does a call to event.stopPropagation(). This function does require you to supply a decoder, so I'm afraid you can't get away from importing Json.Decode, but we are using the simplest possible decoder – Decode.succeed – and only to satisfy the parameters of the function.
I've added a NoOp variant to Msg, as there is nothing to do when the modal is clicked; simply attaching this event handler stops the Hide event from firing when we don't want it to.
Code
module Main exposing (main)
import Browser
import Html exposing (Attribute, Html, button, div, span, text)
import Html.Attributes exposing (class, style)
import Html.Events exposing (on, onClick)
import Json.Decode as Decode
type alias Model =
{ isVisible : Bool, count : Int }
initialModel : Model
initialModel =
{ isVisible = False, count = 0 }
type Msg
= Show
| Hide
| Increment
| Decrement
| NoOp
update : Msg -> Model -> Model
update msg model =
case msg of
Show ->
{ model | isVisible = True }
Hide ->
{ model | isVisible = False }
Increment ->
{ model | count = model.count + 1 }
Decrement ->
{ model | count = model.count - 1 }
NoOp ->
model
view : Model -> Html Msg
view model =
div []
[ button [ onClick Show ] [ text "Show!" ]
, if model.isVisible then
div
(onClick Hide
:: dialogContainerStyle
)
[ div
(onClickStopPropagation NoOp
:: dialogContentStyle
)
[ span [] [ text "Click anywhere outside this dialog to close it!" ]
, span [] [ text "Clicking on anything inside of this dialog works as normal." ]
, div []
[ button [ onClick Decrement ] [ text "-" ]
, text (String.fromInt model.count)
, button [ onClick Increment ] [ text "+" ]
]
]
]
else
div [] []
]
onClickStopPropagation : msg -> Html.Attribute msg
onClickStopPropagation msg =
Html.Events.stopPropagationOn "click" <| Decode.succeed ( msg, True )
dialogContainerStyle : List (Attribute msg)
dialogContainerStyle =
[ style "position" "absolute"
, style "top" "0"
, style "bottom" "0"
, style "right" "0"
, style "left" "0"
, style "display" "flex"
, style "align-items" "center"
, style "justify-content" "center"
, style "background-color" "rgba(33, 43, 54, 0.4)"
]
dialogContentStyle : List (Attribute msg)
dialogContentStyle =
[ style "border-style" "solid"
, style "border-radius" "3px"
, style "border-color" "white"
, style "background-color" "white"
, style "height" "120px"
, style "width" "440px"
, style "display" "flex"
, style "flex-direction" "column"
, style "align-items" "center"
, style "justify-content" "center"
]
main : Program () Model Msg
main =
Browser.sandbox
{ init = initialModel
, view = view
, update = update
}
I am referring to this example in http://elm-lang.org/examples/radio-buttons. I don't see anywhere whereby the state of the buttons is being managed.
In my own little Elm project I need to do something like
label []
[ input
[ type_ "radio"
, checked (model.choosenSize == size)
, onClick (SetSize size)
] []
, text (sizeToString size)
]
Without managing the checked attribute, all the radio buttons will remain checked after you click on it.
So what is the magic in the example?
The example you are referring is very simple. It doesn't explicitly manage the state of the buttons. Instead, their state is managed by the browser. In a real application, of course, you would better manage it explicitly. Something like:
view : Model -> Html Msg
view model =
div []
[ fieldset []
[ radio "Small" (model.fontSize == Small) (SwitchTo Small)
, radio "Medium" (model.fontSize == Medium) (SwitchTo Medium)
, radio "Large" (model.fontSize == Large) (SwitchTo Large)
]
, Markdown.toHtml [ sizeToStyle model.fontSize ] model.content
]
radio : String -> Bool -> msg -> Html msg
radio value isChecked msg =
label
[ style [("padding", "20px")]
]
[ input [ type_ "radio", checked isChecked, name "font-size", onClick msg ] []
, text value
]
(I added a Bool argument to radio)
How do you get the current focus in Elm? I know how to set focus with Elm, but I can't find any functionality to detect what currently has focus.
The elm-lang/dom package allows setting focus on an element given an ID but it does not allow you to fetch the currently focused element. It hints that you can use document.activeElement for this. To do that, you'll have to use ports.
Here is a contrived example. Let's say you have a Model that contains the currently selected id and a list of all ids of some textboxes we'll soon create.
type alias Model =
{ selected : Maybe String
, ids : List String
}
The Msgs we will use will be able to inquire about the focus as well as use the Dom library to set focus:
type Msg
= NoOp
| FetchFocused
| FocusedFetched (Maybe String)
| Focus (Maybe String)
For that, we will need two ports:
port focusedFetched : (Maybe String -> msg) -> Sub msg
port fetchFocused : () -> Cmd msg
The javascript calling these ports will report on the current document.activeElement:
var app = Elm.Main.fullscreen()
app.ports.fetchFocused.subscribe(function() {
var id = document.activeElement ? document.activeElement.id : null;
app.ports.focusedFetched.send(id);
});
The view displays the currently selected id, provides a list of buttons that will set the focus on one of the numbered textboxes below.
view : Model -> Html Msg
view model =
div []
[ div [] [ text ("Currently selected: " ++ toString model.selected) ]
, div [] (List.map viewButton model.ids)
, div [] (List.map viewInput model.ids)
]
viewButton : String -> Html Msg
viewButton id =
button [ onClick (Focus (Just id)) ] [ text id ]
viewInput : String -> Html Msg
viewInput idstr =
div [] [ input [ id idstr, placeholder idstr, onFocus FetchFocused ] [] ]
The update function ties it all together:
update : Msg -> Model -> ( Model, Cmd Msg )
update msg model =
case msg of
NoOp ->
model ! []
FetchFocused ->
model ! [ fetchFocused () ]
FocusedFetched selected ->
{ model | selected = selected } ! []
Focus (Just selected) ->
model ! [ Task.attempt (always NoOp) (Dom.focus selected), fetchFocused () ]
Focus Nothing ->
{ model | selected = Nothing } ! [ fetchFocused () ]
Here is a working example on ellie-app.com.
I'm trying to make a content editable tag that uses enter to update the model.
My code is below, and here is a version that you can play around with on Ellie.
The on "blur" attribute works and updates the model when you click away. But I want to get the same 'update' functionality when an enter is pressed.
view : Model -> Html Msg
view model =
let
attrs =
[ contenteditable True
--, on "blur" (Json.map UpdateTitle targetTextContent)
, onInput2 UpdateTitle
, onEnter EnterPressed
, id "title"
, class "title"
]
in
div []
[ h1 attrs [ text model.existing ]
, text "Click above to start editing. Blur to save the value. The aim is to capture an <enter> and interpret that as a blur, i.e. to save the value and blur the field"
, p [] [ text <| "(" ++ model.existing ++ ")" ]
]
targetTextContent : Json.Decoder String
targetTextContent =
Json.at [ "target", "textContent" ] Json.string
onInput2 : (String -> msg) -> Attribute msg
onInput2 msgCreator =
on "input" (Json.map msgCreator targetTextContent)
onEnter : (Bool -> msg) -> Attribute msg
onEnter enterMsg =
onWithOptions "keydown"
{ stopPropagation = False
, preventDefault = False
}
(keyCode
|> Json.andThen
(\ch ->
let
_ =
Debug.log "on Enter" ch
in
Json.succeed (enterMsg <| ch == 13)
)
)
This code seems to be updating the model ok, but the DOM is getting messed up. For example if I enter enter after "blast" I see this
I tried switching to Html.Keyed and using "keydown" but it did not make any difference or just created different issues.
Solved! The key point is the filter function that uses Json.Decode.fail so that only <enter> is subject to preventDefault. See https://github.com/elm-lang/virtual-dom/issues/18#issuecomment-273403774 for the idea.
view : Model -> Html Msg
view model =
let
attrs =
[ contenteditable True
, on "blur" (Json.map UpdateTitle targetTextContent)
, onEnter EnterPressed
, id "title"
, class "title"
]
in
div []
[ h1 attrs [ text model.existing ]
, text "Click above to start editing. Blur to save the value. The aim is to capture an <enter> and interpret that as a blur, i.e. to save the value and blur the field"
, p [] [ text <| "(" ++ model.existing ++ ")" ]
]
targetTextContent : Json.Decoder String
targetTextContent =
Json.at [ "target", "textContent" ] Json.string
onEnter : msg -> Attribute msg
onEnter msg =
let
options =
{ defaultOptions | preventDefault = True }
filterKey code =
if code == 13 then
Json.succeed msg
else
Json.fail "ignored input"
decoder =
Html.Events.keyCode
|> Json.andThen filterKey
in
onWithOptions "keydown" options decoder
I have a simple form with one field. I would like to clear the field on form submit. I am clearing my model in my update function, but text remains in the text input.
type alias Model =
{ currentSpelling : String }
type Msg
= MorePlease
update : Msg -> Model -> ( Model, Cmd Msg )
update msg model =
case msg of
MorePlease ->
( log "cleared spelling: " { model | currentSpelling = "" }
, fetchWord model.currentSpelling )
view : Model -> Html Msg
view model =
div []
[ Html.form [ onSubmit MorePlease ]
[ input [ type' "text"
, placeholder "Search for your word here"
, onInput NewSpelling
, attribute "autofocus" ""
] []
, text model.currentSpelling
, input [ type' "submit" ] [ text "submit!" ]
]
]
The text displaying model.currentSpelling clears out when I empty it with the update function, but the text input box remains populated. Any idea how to clear it?
fetchWord makes an HTTP call, but it's omitted here.
add value model.currentSpelling into Attributes of the
input element. That's how you can control the string
inside of input element in html.