workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
elmlang | general | i don't need the loader | 2019-01-24T04:38:42.872000 | Desire |
elmlang | general | screwed up yesterday, removed like 50 loaders due to the switch to parcel but forgotten the important ones | 2019-01-24T04:39:44.872300 | Desire |
elmlang | general | Happens to the best. :smile: | 2019-01-24T04:39:54.872500 | Timika |
elmlang | general | I need to focus an `input` element. Is there a way to do this in elm?
`autofocus` attribute only works for page-loads, so thats not an option | 2019-01-24T05:46:59.873500 | Selene |
elmlang | general | <https://package.elm-lang.org/packages/elm/browser/latest/Browser-Dom#focus> | 2019-01-24T05:47:27.873700 | Timika |
elmlang | general | ah thx, the dom package moved into browser :thumbsup: | 2019-01-24T05:48:06.874100 | Selene |
elmlang | general | Hi! Is there a way to subscribe to scroll events? There is the `Browser.Dom.getViewport` task. I could combine it with `onAnimationFrame`. Is there a more straight forward solution? | 2019-01-24T05:56:05.876100 | Ayesha |
elmlang | general | <@Ayesha> is this something you would (in JS) use the `scroll` event for? | 2019-01-24T05:57:58.876600 | Virgie |
elmlang | general | Yes. I need a Msg every time the viewport changes its position. | 2019-01-24T05:58:35.877100 | Ayesha |
elmlang | general | there is this package that handles the `wheel` event which is similar <https://package.elm-lang.org/packages/mpizenberg/elm-pointer-events/latest/Html-Events-Extra-Wheel> (from a quick mdn search there are some differences though)
otherwise you can implement your own event listener to `scroll` using <https://package.elm-lang.org/packages/elm/html/latest/Html-Events#on> | 2019-01-24T06:02:36.878800 | Virgie |
elmlang | general | Thanks. I think the custom `on` event is the way to go. | 2019-01-24T06:05:34.879300 | Ayesha |
elmlang | general | so I looked into `elm-ui`a bit earlier, and really liked the padding/spacing/alignment model... but I've been fiddling with CSS for 10 years, and didn't really want to have to learn a new styling system at the same time as learning Elm. I've also had quite a mixed experience with CSS frameworks - you often end up fighting the system when you want something a bit custom...
but watching <@Cornell>'s talk right now, and it sounds really good! :grinning_face_with_star_eyes: I think I'm going to take the plunge and start converting some views into it | 2019-01-24T07:12:41.888400 | Nana |
elmlang | general | Update us on how it goes. Because I have seen the talks and been thinking about trying them out too :slightly_smiling_face: | 2019-01-24T07:15:33.888600 | Renda |
elmlang | general | did you meat to write elm-ui? | 2019-01-24T07:36:46.889000 | Desire |
elmlang | general | <@Desire> ah right thanks :smile: elm-css is what I've been using so far and it's been a little bit annoying actually with some of the limitations it has | 2019-01-24T07:39:53.889300 | Nana |
elmlang | general | interesting | 2019-01-24T07:57:38.889500 | Desire |
elmlang | general | and yes, elm-ui is exciting | 2019-01-24T07:58:34.889700 | Desire |
elmlang | general | Having spent most of my day on tweaking css with display: inline-block, flexbox and the occasional negative margin, I agree! | 2019-01-24T09:38:54.889900 | Renda |
elmlang | general | yeah even when you know all the tricks, CSS is still fiddly. <@Leoma> | 2019-01-24T10:04:40.890300 | Nana |
elmlang | general | I am using `elm-json-decode-pipeline` and I am trying to use optional but with a default value | 2019-01-24T10:11:29.891300 | Sadie |
elmlang | general | If I use `optional "planning" (Decode.map Just decodePlanning) (Just Dict.empty)` | 2019-01-24T10:11:44.891700 | Sadie |
elmlang | general | It tells me ```The argument is: Decoder (Maybe Planning -> b)``` | 2019-01-24T10:12:39.892200 | Sadie |
elmlang | general | Ok can I make it a maybe.with default ? | 2019-01-24T10:13:11.892500 | Sadie |
elmlang | general | I guess I should just remove all the Just and I should be fine | 2019-01-24T10:15:25.893200 | Sadie |
elmlang | general | `optional "planning" decodePlanning Dict.empty` | 2019-01-24T10:15:44.893500 | Sadie |
elmlang | general | would recommend elm-ui, it's awesome! writing CSS often comes down to how well you know all the tricks and loopholes... which is not how things should be :exploding_head: | 2019-01-24T10:17:27.893800 | Glenda |
elmlang | general | I can also recommend <https://github.com/circuithub/elm-semantic-ui> | 2019-01-24T10:23:22.894000 | Zachary |
elmlang | general | Even though elm-ui certainly has some gaps in features and a few bugs, I’m still 100% sold on it.
Knowledge of CSS can help work around some of elm-ui’s current (and reasonably minor) issues/limitations. | 2019-01-24T10:24:52.894300 | Leoma |
elmlang | general | Removing loaders is the best part when switching to parcel :smile: | 2019-01-24T10:25:05.894500 | Zachary |
elmlang | general | <@Zachary> I think Elm-UI is way more interesting though, as it's more a language to replace CSS rather than just a set of predefined CSS classes | 2019-01-24T10:31:27.894700 | Nana |
elmlang | general | I guess you must return a Maybe Dict with Dict.empty. | 2019-01-24T10:36:18.896600 | Tracey |
elmlang | general | The way you're doing you are returning only the Just part, without the Nothing. | 2019-01-24T10:36:55.897600 | Tracey |
elmlang | general | Folks, another question.
I found myself in a situation that I think the compiler can help me to avoid.
In a mistake I have a function A that returns function B then returns function A again.
As a human I know that this will lead me to a infinite loop. But I'm wondering if there is a way we can make the compiler identify that misleading before it happen in runtime. | 2019-01-24T10:47:26.902800 | Tracey |
elmlang | general | <@Tracey> totality detection isn't possible with a Turing-complete language :disappointed: | 2019-01-24T11:01:41.904000 | Niesha |
elmlang | general | More context: <https://en.wikipedia.org/wiki/Halting_problem> | 2019-01-24T11:06:28.904800 | Timika |
elmlang | general | You can detect some programs that won’t halt, you just can’t guarantee you’ll detect all of them. | 2019-01-24T11:12:51.906000 | Simon |
elmlang | general | So the compiler could tell you if you’ve created something known to be infinitely recursive; some languages do do this already | 2019-01-24T11:13:28.906700 | Simon |
elmlang | general | Elm does this for values (e.g. decoders, where people run into this fairly regularly) | 2019-01-24T11:23:36.907800 | Huong |
elmlang | general | it doesn't do this for functions, and that's actually something used in the implementations of functions like `never : Never -> a` | 2019-01-24T11:24:15.908700 | Huong |
elmlang | general | (which is pretty much `never (Never v) = never v` given `type Never = Never Never`) | 2019-01-24T11:24:47.909300 | Huong |
elmlang | general | hello everyone! I’m having a weird error that I never encountered before:
```The `Http` module does not expose a `Request` type. These names seem close``` | 2019-01-24T11:28:35.910200 | Karrie |
elmlang | general | this is my code:
```
import Http
import Url.Builder exposing (QueryParameter)
request :
{ body : Http.Body
, expect : Http.Expect a
, headers : List Http.Header
, method : String
, timeout : Maybe Float
, url : String
}
-> Http.Request a
request =
Http.request
{ body = config.body
, expect = config.expect
, headers = config.headers
, method = config.method
, timeout = config.timeout
, url = config.url
}
``` | 2019-01-24T11:28:51.910500 | Karrie |
elmlang | general | There is no `Request` type in the Http module anymore: <https://package.elm-lang.org/packages/elm/http/latest/Http> | 2019-01-24T11:32:17.911400 | Earnest |
elmlang | general | functor is defined for as `for all a b.` of course you can use it in case where `a = b` but you can't say that thing that supports only this case is functor. if so then everything would be functor because this is true for any data type. It can be defined as `map _ a = a` | 2019-01-24T11:32:49.911500 | Zona |
elmlang | general | therefore it doesn't even make sense to talk about this property at all. | 2019-01-24T11:33:15.911900 | Zona |
elmlang | general | Check out the return type of `Http.request`: <https://package.elm-lang.org/packages/elm/http/latest/Http> | 2019-01-24T11:33:16.912200 | Earnest |
elmlang | general | thanks! | 2019-01-24T11:41:26.912600 | Karrie |
elmlang | general | is there a way to convert “2019-01-24T16:20:00Z” to a date time type? | 2019-01-24T11:41:44.913000 | Karrie |
elmlang | general | I like the never type, though it should probably get first class type system treatment vs an interesting hack on the inference system. Then it’d free up Elm to help folks a little bit more in the other dimensions | 2019-01-24T11:41:57.913100 | Simon |
elmlang | general | I think this is what you're looking for <https://package.elm-lang.org/packages/rtfeldman/elm-iso8601-date-strings/latest/Iso8601> | 2019-01-24T11:42:29.913500 | Virgie |
elmlang | general | brilliant! :slightly_smiling_face: | 2019-01-24T11:43:19.913800 | Karrie |
elmlang | general | how can I convert it to a readable datetime though? (from frontend) | 2019-01-24T11:49:47.914000 | Karrie |
elmlang | general | you mean more readable than YYYY-MM-DDTHH:mm:ss.SSSZ? | 2019-01-24T11:53:18.914200 | Virgie |
elmlang | general | it depends a little bit on what you want really. The readme of elm/time has an example <https://package.elm-lang.org/packages/elm/time/latest/> | 2019-01-24T11:54:12.914400 | Virgie |
elmlang | general | Is the following worth reporting as a bug?
```$ elm make src/Router.elm
elm: ./Data/Vector/Generic/Mutable.hs:703 (modify): index out of bounds (3,3)
CallStack (from HasCallStack):
error, called at ./Data/Vector/Internal/Check.hs:87:5 in vector-0.12.0.1-IfIQAgrX0q07Xkx6bkYLRX:Data.Vector.Internal.Check
elm: thread blocked indefinitely in an MVar operation
``` | 2019-01-24T12:51:28.915400 | Dede |
elmlang | general | My plan is to report it unless someone here says it's known/fixed upstream in which case I'll ignore it. | 2019-01-24T12:51:59.915500 | Dede |
elmlang | general | does it happen consistently? | 2019-01-24T12:52:28.915700 | Virgie |
elmlang | general | 100% of the time with the file at this point. | 2019-01-24T12:52:38.915900 | Dede |
elmlang | general | (that is relevant because MVars are used in distributing work over multiple cores) | 2019-01-24T12:53:24.916100 | Virgie |
elmlang | general | if there's not a duplicate then sure report it. | 2019-01-24T12:53:58.916400 | Virgie |
elmlang | general | and if you have a good reproducable example than share it regardless | 2019-01-24T12:54:15.916600 | Virgie |
elmlang | general | I had initially thought there wasn't a duplicate, but focusing on the MVar helped. This is probably what's going on: <https://github.com/elm/compiler/issues/1748>
Thanks. | 2019-01-24T12:58:15.916800 | Dede |
elmlang | general | <@Karrie> this is an excellent package for converting Time.Posix to Strings:
<https://package.elm-lang.org/packages/ryannhg/date-format/latest/> | 2019-01-24T13:20:58.917200 | Nana |
elmlang | general | <@Nana> I will have a look into it! Thanks! | 2019-01-24T13:57:03.917700 | Karrie |
elmlang | general | Basically, one thread crashes with an exception because someone is using an unsafe operation, and then the deadlock detection kicks in - that's the MVar part. | 2019-01-24T16:38:35.918500 | Niesha |
elmlang | general | Usually I've seen Evan tell people to create an issue if an SSCCE is available regardless whether one already exists in order to have more data on the matter. | 2019-01-24T17:02:45.918800 | Freda |
elmlang | general | Is it considered bad form to have an `init` which says, e.g. {model = []}, but also has a command which does some HTTP and sends back the ‘actual’ values for `model`? | 2019-01-24T19:20:54.920700 | Bernardo |
elmlang | general | Because: I’m having trouble with :arrow_up: , where the ‘actual’ values are loaded and sent back through a port (AFAICT from the JS side), but the model is never updated :disappointed: | 2019-01-24T19:21:49.921900 | Bernardo |
elmlang | general | It occurs to me now I could just defer initializing the Elm app and send the ‘actual’ values through as a flag, but I wanted to see if there’s a consensus on this? :thinking_face: | 2019-01-24T19:22:43.923000 | Bernardo |
elmlang | general | <@Bernardo> that's fine. It's not uncommon to have `init` produce a value for the model that shows it in a loading state and http requests to fetch data | 2019-01-24T19:28:01.925400 | Earlean |
elmlang | general | <@Earlean> okay. So now the mystery begins: I switched to the ‘just send an initial value in flags’ approach and it works fine, so:
Are there any known issue with having commands in initializers tripping over initial model values? | 2019-01-24T19:35:02.927900 | Bernardo |
elmlang | general | If I fire off the command via e.g. a button, it works just fine, it just fails to do anything when I use it in `init` :thinking_face: | 2019-01-24T19:38:51.929100 | Bernardo |
elmlang | general | Is that the root `init`? If inside a module, maybe is not wired up with the top init | 2019-01-24T19:40:10.930400 | Carrol |
elmlang | general | Yep, root `init`, I’m a bit of newbie, so this is my first Elm project, but still just one file and one `init` :slightly_smiling_face: | 2019-01-24T19:41:03.931300 | Bernardo |
elmlang | general | I’ll try to make a repro. | 2019-01-24T19:41:08.931600 | Bernardo |
elmlang | general | <@Bernardo> commands aren't going to see you model value, the command will produce a Msg value that will be passed to your `update` function | 2019-01-24T19:49:11.933700 | Earlean |
elmlang | general | I see you mentioned you are using a port. Have you subscribed to the port in your `subscriptions`? | 2019-01-24T19:50:00.935600 | Earlean |
elmlang | general | Yeah, so I _presume_ `update` gets the message back via the port call in JS (which is triggered by the command fired off in the init) | 2019-01-24T19:50:07.935800 | Bernardo |
elmlang | general | And that bit of `update` logic updates the model, but view doesn’t seem to change :grimacing: | 2019-01-24T19:50:39.937100 | Bernardo |
elmlang | general | You send on a port with a command and receive from ports via `subscriptions` | 2019-01-24T19:50:55.937600 | Earlean |
elmlang | general | Yep | 2019-01-24T19:51:05.937800 | Bernardo |
elmlang | general | What’s extra weird is: <https://elmlang.slack.com/archives/C0CJ3SBBM/p1548376731929100> | 2019-01-24T19:51:05.938000 | Bernardo |
elmlang | general | Do both the command and the button click produce the same Msg value? | 2019-01-24T19:52:08.939100 | Earlean |
elmlang | general | Look at the case for the Msg value that your command would produce in your `update` | 2019-01-24T19:54:15.940400 | Earlean |
elmlang | general | Yeah, same Msg | 2019-01-24T19:58:26.940600 | Bernardo |
elmlang | general | I’m just whipping up a repro in Ellie | 2019-01-24T19:58:37.940900 | Bernardo |
elmlang | general | Of course I’m sure I won’t be able to repro :joy: | 2019-01-24T19:58:47.941200 | Bernardo |
elmlang | general | Ah, are you sending to the port in JS immediately after initialising the Elm app? | 2019-01-24T20:01:50.942300 | Earlean |
elmlang | general | Sending to a port before Elm is ready doesn't work. | 2019-01-24T20:03:02.943700 | Earlean |
elmlang | general | <@Earlean> so that’s what I was doing at first, and I had the same behavior, which made me thing, _I know, I’ll have a command fire in `init` to trigger the HTTP and port send back, that way I can be sure Elm is ready_. But that didn’t work :disappointed: | 2019-01-24T20:05:42.945100 | Bernardo |
elmlang | general | Aaaand, it works as expected in my Ellie app :grimacing: <https://ellie-app.com/4xNRkmBRp2Va1> | 2019-01-24T20:05:55.945500 | Bernardo |
elmlang | general | Who knows, I must have done something silly in my actual app. | 2019-01-24T20:06:33.945900 | Bernardo |
elmlang | general | Sorry and thanks for putting up with me! | 2019-01-24T20:06:40.946200 | Bernardo |
elmlang | general | This: <https://github.com/elm/compiler/issues/774>, specifically not being able to use a `type alias` as the key in a `Dict` seems crazy to me. Am I missing something? | 2019-01-24T23:14:41.947000 | Bernardo |
elmlang | general | <@Bernardo> It's not related to `type alias` at all, it's about which types are `comparable` | 2019-01-24T23:20:49.948100 | Earlean |
elmlang | general | some of the built in types are `comparable`, they have some reasonable ordering. `Int`, `String` etc. are `comparable` | 2019-01-24T23:21:42.949200 | Earlean |
elmlang | general | By user defined types and records don't have a reasonable default ordering so you'd need some way to specify some how they should be ordered | 2019-01-24T23:22:59.950600 | Earlean |
elmlang | general | Most do have a reasonable ordering, it's just not the way it works | 2019-01-24T23:23:56.951600 | Kris |
elmlang | general | <@Bernardo> the way similar languages do this is by having types implement some kind of interface (a type class) that defines how values of that type should be ordered. Which leads us in to the discussion of why Elm doesn't have user definable or implementable type classes. | 2019-01-24T23:25:49.953200 | Earlean |
elmlang | general | There are a few packages that provide wrappers around `Dict` that take a function of `(k -> comparable)` so you can specify some function for converting your key values from some type to a type that is comparable. eg. <https://package.elm-lang.org/packages/turboMaCk/any-dict> | 2019-01-24T23:32:30.955200 | Earlean |
Subsets and Splits