workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
elmlang | general | Adding it in general cases means you can't check for exhaustiveness | 2019-02-28T11:58:37.066600 | Kris |
elmlang | general | I'm referring to just the cases where the function is a compile time constant value though | 2019-02-28T11:59:19.068100 | Jae |
elmlang | general | because of stuff like
```
foo a b =
case (a, b) of
(x, x) -> ..
``` | 2019-02-28T11:59:25.068300 | Kris |
elmlang | general | also you'd need some way to disambiguate if you mean binding `a` to `someConstantValue` or having it equal to `someConstantValue` | 2019-02-28T11:59:26.068400 | Shondra |
elmlang | general | (syntactically) | 2019-02-28T12:00:00.068800 | Shondra |
elmlang | general | Ah, I understand. Yeah that makes sense | 2019-02-28T12:00:18.069500 | Jae |
elmlang | general | Yeah, there's no reason for not allowing that for compile time stuff, other than it being ugly | 2019-02-28T12:00:31.069900 | Kris |
elmlang | general | erlang/elixir have that feature, they use the `^` operator | 2019-02-28T12:01:01.070300 | Shondra |
elmlang | general | `case x of ^y` means if x equals y, but then again, in erlang/elixir there is no exhaustive checking | 2019-02-28T12:01:52.071200 | Shondra |
elmlang | general | You can do it in haskell with template haskell
```
someQ = ... something equivalent to 5...
case x of
$(someQ) -> "5"
``` | 2019-02-28T12:02:51.072800 | Kris |
elmlang | general | (TemplateHaskell is Haskell's compile time macro system) | 2019-02-28T12:03:17.073300 | Kris |
elmlang | general | I understand the desire to minimize the amount of syntax but right now I wish I had something like that so that I can reuse the same constant string in multiple case statements to reduce the risk of typos | 2019-02-28T12:03:53.074100 | Jae |
elmlang | general | Would it make sense to use `if ... then ... else` ? | 2019-02-28T12:04:56.074700 | Carman |
elmlang | general | Yeah I can sympathize | 2019-02-28T12:05:00.074800 | Kris |
elmlang | general | Perhaps you can factor that out in some other way | 2019-02-28T12:05:12.075300 | Kris |
elmlang | general | <@Carman> That's a good point. In this instance I can do that. Thanks! | 2019-02-28T12:05:48.076000 | Jae |
elmlang | general | Hey there -- I'm a bit disappointed that the built in elm/browser and elm/html mouse events don't decode the various fields available on the JS event and am considering which package to use as an alternative. Is there one folks here prefer or recommend? I think I'm leaning towards <https://github.com/mpizenberg/elm-pointer-events> | 2019-02-28T12:47:25.079500 | Elvis |
elmlang | general | Haskell has another way to do this: the PatternSynonyms extension. See <https://ocharles.org.uk/posts/2014-12-03-pattern-synonyms.html> | 2019-02-28T14:14:09.079700 | Paula |
elmlang | general | So you could write:
```
someConstant = 5
pattern SomeConstant = someConstant
case x of
SomeConstant -> whatever
``` | 2019-02-28T14:20:04.080000 | Paula |
elmlang | general | Ah, true, I’d forgotten about those. They’re great | 2019-02-28T14:29:15.080600 | Kris |
elmlang | general | There is something that has been bothering me for a long time, is there a way to model my Elm app as a state machine so that the compiler can know that the app is already in some state and I can skip checking for impossible states? For example: (1) an action can be dispatched only if the user presses a button, (2) that button is shown only if the data has been loaded from the server, (3) therefore I don't have to check if the data is loading in the update for that action. I know doing that in Elm is probably impossible but I'm really interested in learning about tools that enable that. | 2019-02-28T15:34:55.084000 | Dayna |
elmlang | general | It's tricky if your system is asynchronous like Elm. Consider the following scenario:
1. A user clicks a button when the app is in state `A`. This fires a (slow) HTTP request in the background
2. While the request is happening, the user clicks another button, transitioning the app to state `B`
3. The HTTP request finally completes, triggering a message | 2019-02-28T15:39:43.086000 | Carman |
elmlang | general | Normally you'd only expect the message to come in when the app is in state `A` but because messages are asynchronous it could definitely come in during state `B` (or `C`, etc) | 2019-02-28T15:40:44.087000 | Carman |
elmlang | general | I think that in a state machine with asynchronous events, any message _could_ happen in any state | 2019-02-28T15:41:49.087800 | Carman |
elmlang | general | Well if modelling such state transitions was possible then surely you could model the possible transitions that can happen during loading of the request | 2019-02-28T15:42:59.089700 | Dayna |
elmlang | general | if your `Model` is an enum you could do something like
```
case (model, msg) of
(ModelStateOne stateOne, MsgOne) ->
-- handle this case
(model, Cmd.none)
_ ->
(model, Cmd.none)
```
it’s not as clean as you’re hoping, but it could work | 2019-02-28T15:43:21.090200 | Alleen |
elmlang | general | enum => custom type | 2019-02-28T15:43:37.090400 | Alleen |
elmlang | general | i’m too used to Rust right now | 2019-02-28T15:43:42.090600 | Alleen |
elmlang | general | Sure, but I'm more interested in a way to model the transitions upfront and then have the compiler force me to prove that the model is in the correct state (request loaded etc) | 2019-02-28T15:44:38.091800 | Dayna |
elmlang | general | The only practical solution I can think of right now is to put the button click message inside the model and when the request is loaded to partially apply the response data to the message so that when the view renders the response: (1) view is forced to pattern match on the response to extract the message it should send and (2) the update will also get the response data from the message, not the model and it will be confirmed to exist. This is all very theoretical and hacky, of course, but I can see myself playing with the idea | 2019-02-28T15:49:12.096400 | Dayna |
elmlang | general | Right, so the issue is that you have two values - model and msg - and you need to combine their types in some way :thinking_face: | 2019-02-28T15:49:45.097400 | Nana |
elmlang | general | The more I think about it, the better it sounds :smile: I'd imagine a model with
`article : WebData (Article, { editArticle : Msg })` where the editArticle message constructor is actually `Article -> Msg` but the update function will partially apply the data when it's returned from the server. The view will pattern match on the article in the model to extract the message for the edit button and the update will grab the article by pattern matching the message. | 2019-02-28T15:56:52.101700 | Dayna |
elmlang | general | So, if your model only represents legal states, and you write a transition, you can’t possibly have a problem.
You can’t write a transition that accepts an illegal state, or creates one.
Can you tell us more about the problems you’re solving and why you focused on this idea of a state machine? | 2019-02-28T15:59:04.103400 | Leoma |
elmlang | general | The problem is that 90% of the code I write (and I imagine most boring CRUD apps are like that) is requesting data, showing stuff on the screen and sending requests when the user presses something. On every `XButtonClicked` message I have to pattern match on the request to get the data and since the request could be loading, I have to add a branch like in <@Alleen>’s example | 2019-02-28T16:01:10.105800 | Dayna |
elmlang | general | <@Dayna> you could get the data from the view like `onClick (DoStuff loadedData)` but I think that can cause some concurrency issues? | 2019-02-28T16:03:41.108900 | Nana |
elmlang | general | That's also a way to do it, but I think the theoretical API I showed above is a bit more reliable and also it doesn't involve the view knowing what data the update needs | 2019-02-28T16:04:32.109800 | Dayna |
elmlang | general | Seems it'd be nice if there were an easy generic way to escape variables in patterns though, to say that they should be evaluated rather than bound. | 2019-02-28T16:16:11.110500 | Paula |
elmlang | general | <@Dayna> you could also use websockets and subscribe to graphql queries :) | 2019-03-01T02:15:00.119000 | Valeria |
elmlang | general | How do I make this code typecheck?
```
type alias Foo = { hello : Int }
updateFoo : Foo -> Int -> Foo
updateFoo foo num = { foo | hello = num }
type Things =
Good Foo
| Bad Foo
makeUpdater : Things -> Int -> (a -> Int -> a) -> Things
makeUpdater things newNum updater =
case things of
Good foo
-> updater foo newNum
Bad foo
-> things
``` | 2019-03-01T03:20:29.120100 | Deane |
elmlang | general | The compiler tells me this:
```
The 1st argument to `updater` is not what I expect:
35| -> updater foo newNum
^^^
This `foo` value is a:
Foo
But `updater` needs the 1st argument to be:
a
``` | 2019-03-01T03:20:48.120400 | Deane |
elmlang | general | But updater IS being used correctly, no? | 2019-03-01T03:21:06.120700 | Deane |
elmlang | general | <@Deane> `type` is not the same as `type alias` | 2019-03-01T03:25:43.121200 | Lynne |
elmlang | general | You have defined new type and given it a `Maybe` constructor (shadowing the one imported implicitly) | 2019-03-01T03:26:33.122000 | Lynne |
elmlang | general | Also, even if you used `type alias` it would not work because you would be returning `Foo` but `makeUpdater` says it should be `Maybe Foo` | 2019-03-01T03:28:09.123200 | Lynne |
elmlang | general | right, my example doesn't demonstrate the issue I thought it did | 2019-03-01T03:28:38.123800 | Deane |
elmlang | general | I'll try to come up with another one | 2019-03-01T03:28:49.124100 | Deane |
elmlang | general | anyone got experience with an elm SPA and googles crawler?
I am seeing the following error in the console: `Uncaught SyntaxError: Unexpected token ,`
Would be nice if i could rule out elm as the problem here | 2019-03-01T03:30:29.125200 | Selene |
elmlang | general | <@Lynne> I've "corrected" the example to show the right error | 2019-03-01T03:31:31.125900 | Deane |
elmlang | general | <@Deane> I pointed to the problem in my last statement :slightly_smiling_face: You return `Foo` but `makeUpdater` says it should be `Things` (either `Good Foo` or `Bad Foo`) | 2019-03-01T03:34:20.126700 | Lynne |
elmlang | general | So you should have written, for example, `Good <| updater foo newNum` in your first branch | 2019-03-01T03:35:17.127400 | Lynne |
elmlang | general | And I think it still won't compile because `a` is too generic in your case and compiler will force you changing signature of `updater` to `(Foo -> Int -> Foo)`. If you want a generic signature you should write something like:
```
type Things a
= Good a
| Bad a
makeUpdater : Things a -> Int -> (a -> Int -> a) -> Things a
``` | 2019-03-01T03:37:41.129300 | Lynne |
elmlang | general | I agree with <@Johna> - if you are asking the question... roll your own. In my experience, extensible means trying to fit too many use cases. Is usually better to do exactly what you need. | 2019-03-01T04:04:43.129600 | Gale |
elmlang | general | Thank you ! | 2019-03-01T04:07:43.129900 | Yolando |
elmlang | general | :point_up: | 2019-03-01T04:11:56.130100 | Johna |
elmlang | general | Hi. I currently trying to upgrade from 0.18 to 0.19 and ran into a package dependency thing. I used elm-upgrade for the job and now I swapped the packages which don't support the new elm (sortable-data-tables and date-format) manually for their new forks:
"evancz/elm-sortable-table": "1.0.1 <= v < 2.0.0", -> "billstclair/elm-sortable-table": "1.1.1 <= v < 2.0.0",
"mgold/elm-date-format": "1.6.0 <= v < 2.0.0" -> "ryannhg/date-format": "1.6.0 <= v < 2.0.0" | 2019-03-01T04:52:08.132900 | Daysi |
elmlang | general | now I'm stuck in this situation: | 2019-03-01T04:54:57.133600 | Daysi |
elmlang | general | so elm-upgrade says "dependencies are fine", `elm make` says "dependencies are not good" | 2019-03-01T04:55:40.134600 | Daysi |
elmlang | general | can someone point me in the right direction? | 2019-03-01T04:56:00.134900 | Daysi |
elmlang | general | Help is appreciated | 2019-03-01T04:56:09.135300 | Daysi |
elmlang | general | <https://www.markuslaire.com/github/elm-dependencies-analyzer/> is a tool to help you work out dependency issues | 2019-03-01T04:56:25.136000 | Jin |
elmlang | general | <@Lynne> yes, you're correct when you say "And I think it still won't compile because `a` is too generic in your case and compiler will force you changing signature of `updater` to `(Foo -> Int -> Foo)`" | 2019-03-01T04:56:41.136400 | Deane |
elmlang | general | And yes looks like the solution is basically that `Things` will become `Things a` | 2019-03-01T04:57:29.137000 | Deane |
elmlang | general | thanks! | 2019-03-01T04:57:30.137200 | Deane |
elmlang | general | elm-upgrade will only check that your dependencies are Elm 0.19 ready, afaik. Elm will check if the dependency constraints fit. | 2019-03-01T04:58:10.137400 | Jin |
elmlang | general | You are welcome <@Deane> | 2019-03-01T04:58:21.137700 | Lynne |
elmlang | general | thanks! but the tool says "Only application elm.json is supported, not package elm.json.
If you would find it useful to also support package elm.json,
please open an issue at GitHub repository." when I upload my elm.json | 2019-03-01T04:58:36.137800 | Daysi |
elmlang | general | I thought elm-upgrade automatically upgraded the file to the correct structure but maybe there is an error? | 2019-03-01T04:59:32.138000 | Daysi |
elmlang | general | You have to put in the elm.json of the full application, not of one of the depemdencies. | 2019-03-01T04:59:33.138200 | Jin |
elmlang | general | I did | 2019-03-01T04:59:41.138400 | Daysi |
elmlang | general | maybe, yes. your elm.json must be of type application | 2019-03-01T04:59:52.138600 | Jin |
elmlang | general | and your versions there must be fixed, no ranges | 2019-03-01T05:00:10.138800 | Jin |
elmlang | general | If nothing helps, remove all dependencies from your `elm.json` and install them again. Not ideal but not that long. | 2019-03-01T05:00:36.139000 | Velia |
elmlang | general | Hey beautiful people. I have the task at hand of given a sentence in a string, get all urls in the sentence.
I found out that elm/Url parser accepts things like <http://word> as valid urls, which doesn't work for us.
In writing a Url parser I arrived at a point where I have a big list of domain extension names (.com, .org, .gov) and need to match against that.
I'd like to do it without backtracking to keep things fast, so I'm thinking of building a *prefix tree* out of my list and will create my parser out of that.
Is there a better way to do the matching against the list? If not, could anyone recommend a library to build prefix trees? The rluiten/trie doesn't seem to allow the mapping from Trie -> Parser that I need. | 2019-03-01T05:41:00.144300 | Tawnya |
elmlang | general | <@Tawnya> Is not `Parser.oneOf [keyword ".com", keyword ".org", keyword ".gov"]` sufficient? | 2019-03-01T06:06:45.145200 | Lynne |
elmlang | general | But from what I understand `Parser.oneOf [ keyword ".com", keyword ".<http://co.uk|co.uk>" ]` would fail for `.<http://co.uk|co.uk>` values | 2019-03-01T06:10:53.146400 | Tawnya |
elmlang | general | Unless I used backtracking. | 2019-03-01T06:11:02.146600 | Tawnya |
elmlang | general | Is that right? | 2019-03-01T06:11:16.146800 | Tawnya |
elmlang | general | Not sure, why would it? .com and .<http://co.uk|co.uk> don't have a common "superstring" | 2019-03-01T06:13:40.147600 | Lynne |
elmlang | general | Meaning there is no string for which both .com and .<http://co.uk|co.uk> are substrings | 2019-03-01T06:14:01.148000 | Lynne |
elmlang | general | From what I see from `Parser.Advanced.keyword`'s definition, it does not progress if there is no match, so no backtracking is needed | 2019-03-01T06:16:35.148800 | Lynne |
elmlang | general | :thumbsup: Nice. Thanks! | 2019-03-01T06:24:50.149100 | Tawnya |
elmlang | general | Hello everyone! I have an image that I am loading with `img [ src "/images/myImage.png"] []`. When I `Browser.Navigation.pushUrl` for a new url, the link breaks. Why might this be the case? | 2019-03-01T08:07:40.150800 | Donya |
elmlang | general | ah, nm, it started working all of a sudden. strange! | 2019-03-01T08:10:04.151100 | Donya |
elmlang | general | Hey! I'm trying to publish my package, but I'm getting this: ```
I will now verify that everything is in order...
● Found README.md
● Found LICENSE
● All packages start at version 1.0.0
✗ Version 1.0.0 is not tagged on GitHub!
-- NO TAG ----------------------------------------------------------------------
Packages must be tagged in git, but I cannot find a 1.0.0 tag.
These tags make it possible to find this specific version on GitHub.
To tag the most recent commit and push it to GitHub, run this:
git tag -a 1.0.0 -m "new release"
git push origin 1.0.0
The -m flag is for a helpful message. Try to make it more informative!
``` The tag is there (<https://github.com/tad-lispy/elm-springs/releases/tag/1.0.0>) so what's going on? | 2019-03-01T08:45:56.152500 | Dorsey |
elmlang | general | I believe this is some sort of caching issue with the github api. | 2019-03-01T08:49:20.153400 | Virgie |
elmlang | general | Shall I just try again later? | 2019-03-01T08:52:08.153700 | Dorsey |
elmlang | general | try that first, you can also force untag and then tag again | 2019-03-01T08:55:29.154300 | Virgie |
elmlang | general | i use 0.19 to `elm build` a module and it tells me | 2019-03-01T09:06:01.155000 | Daysi |
elmlang | general | can one not use explicit imports of that kind anymore? | 2019-03-01T09:06:40.155900 | Daysi |
elmlang | general | Elm told me to write `import Bla exposing (Blubb(..))` instead of `(Blubb(Foo))` just before that. I corrected it but now Elm says something isn't right in an Elm package that should work ... So I wanted to ask. | 2019-03-01T09:08:36.158100 | Daysi |
elmlang | general | in 0.19 you either import all variants/constructors, or none of them. | 2019-03-01T09:09:16.158600 | Virgie |
elmlang | general | this makes sense if you pattern match on a value of the custom type: you need all constructors to make the pattern match exhaustive | 2019-03-01T09:09:47.159300 | Virgie |
elmlang | general | on the other hand it is now a little harder to see where a name comes from | 2019-03-01T09:10:17.159800 | Virgie |
elmlang | general | Hello folks, I was trying to write a `port` and did not write it correctly I guess, so the compiler told me :
```
Hint: Ports are not a traditional FFI for calling JS functions directly. They
need a different mindset! Read <https://elm-lang.org/0.19.0/ports> to learn how
to use ports effectively.
```
But the page looks empty to me. Is it just me or ?
Thank you :slightly_smiling_face: | 2019-03-01T09:48:14.161500 | Toni |
elmlang | general | There are several links in the compiler hints that are broken :disappointed: | 2019-03-01T09:49:11.161900 | Carman |
elmlang | general | Here's the official guide page on ports: <https://guide.elm-lang.org/interop/ports.html> | 2019-03-01T09:49:27.162500 | Carman |
elmlang | general | oh, sorry to hear that, can I open an issue or help in some way ? | 2019-03-01T09:49:36.162800 | Toni |
elmlang | general | thanks for the link | 2019-03-01T09:49:51.163000 | Toni |
elmlang | general | Hmm looking at the repo it looks like it's present but empty :thinking_face: <https://github.com/elm/elm-lang.org/blob/master/src/pages/0.19.0/ports.elm> | 2019-03-01T09:55:23.163500 | Carman |
elmlang | general | <@Toni> here's a basic example :slightly_smiling_face:
<https://ellie-app.com/4HH7pySpmpVa1> | 2019-03-01T10:01:46.163800 | Nana |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.