workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
---|---|---|---|---|
elmlang | general | Thank you ! I was just missing the `port` declaration on top of my module :slightly_smiling_face: | 2019-03-01T10:02:31.164300 | Toni |
elmlang | general | the issue already exists, actually : <https://github.com/elm/elm-lang.org/issues/798> | 2019-03-01T10:38:25.166000 | Toni |
elmlang | general | Hey, I'm still trying to publish. I've waited long and even deleted and re-created the GitHub repo, but it is still saying that the tag is not there. Can it be some local cache? | 2019-03-01T10:38:38.166100 | Dorsey |
elmlang | general | Nah, I just tried running it from a container and it's the same. | 2019-03-01T10:43:19.166500 | Dorsey |
elmlang | general | can someone point me to an in-dept explaination of the elm.json? <https://github.com/elm/compiler/blob/master/docs/elm.json/application.md> is not enough for me | 2019-03-01T10:45:00.167300 | Daysi |
elmlang | general | what are you missing ? | 2019-03-01T10:45:55.168000 | Mindy |
elmlang | general | because there isn't anything more to elm.json than this and the <https://github.com/elm/compiler/blob/master/docs/elm.json/package.md> version | 2019-03-01T10:46:37.169100 | Mindy |
elmlang | general | why?
A: I get "module not found" errors when `elm make`ing all Elm-apps for my big, multipage app. The "source-directories" in my elm.json are just like in the old elm-package.json. But now the compiler doesn't find modules anymore ... | 2019-03-01T10:47:35.170100 | Daysi |
elmlang | general | Aw, silly me. The repository name did not match package name! | 2019-03-01T10:48:15.170300 | Dorsey |
elmlang | general | None | 2019-03-01T10:48:32.170500 | Daysi |
elmlang | general | "source-directories": [
"app/fas_ui",
"app/fas_ui/demo"
], | 2019-03-01T10:49:07.170900 | Daysi |
elmlang | general | so it should just work. After all Elm starts to compile files in the `app/fas_ui` directory but doesn't find modules which live in the same directory and subdirectory. | 2019-03-01T10:50:10.172400 | Daysi |
elmlang | general | I think the message from `elm publish` could be improved and say something like `no repository by this name on <http://github.com|github.com>`. | 2019-03-01T10:51:16.173500 | Dorsey |
elmlang | general | Do those directories directly contain your elm files ? | 2019-03-01T10:51:31.174200 | Mindy |
elmlang | general | does Elm try to look not in the project root but takes the node_modules/.bin/elm as root and looks for node_modules/.bin/elm/app/fas_ui or something? | 2019-03-01T10:51:38.174700 | Daysi |
elmlang | general | Maybe you should put `app/fas_ui/src/elm` in your `source-directories` | 2019-03-01T10:51:54.175100 | Mindy |
elmlang | general | the elm files are in subdirectories | 2019-03-01T10:51:56.175200 | Daysi |
elmlang | general | of these directories | 2019-03-01T10:52:04.175500 | Daysi |
elmlang | general | I think you should put the whole path up to the topmost elm module | 2019-03-01T10:53:02.176000 | Mindy |
elmlang | general | that changed the error message at least :smile: | 2019-03-01T10:54:21.176200 | Daysi |
elmlang | general | now it doesn't find a module which lives in a symlinked directory (we do translations this way by symlinking to the different language directories and building for every language) | 2019-03-01T10:55:22.176400 | Daysi |
elmlang | general | None | 2019-03-01T10:56:04.176700 | Daysi |
elmlang | general | ..../elm/I18n is a symlink to a folder | 2019-03-01T10:56:53.177500 | Daysi |
elmlang | general | so maybe I have to adjust the bash `find` command in `node_modules/.bin/elm make $(find app/fas_ui/src/elm/Fas/Apps -name '*.elm' -type f)` | 2019-03-01T10:57:42.178000 | Daysi |
elmlang | general | no idea though how to do that | 2019-03-01T11:00:33.178300 | Daysi |
elmlang | general | isn't there an easy way to say "hey elm: you see that folder? Find all apps in there and bundle them individually into another folder." | 2019-03-01T11:01:15.179200 | Daysi |
elmlang | general | So I just had a random stupid idea… but has anyone seen this sort of thing before or are there good / bad reasons to structure code like this?
<https://ellie-app.com/4RGFp2Snr8ca1> | 2019-03-01T11:59:49.179900 | Buffy |
elmlang | general | Yeh update just becomes function application :confused: <https://ellie-app.com/4RGJFvfGbzSa1> ??? | 2019-03-01T12:05:36.181400 | Buffy |
elmlang | general | Someone tell me this is stupid before I go and build an app like this, kek. | 2019-03-01T12:06:02.181900 | Buffy |
elmlang | general | Why would you do that? What's the intended benefit? | 2019-03-01T12:07:21.182700 | Dorsey |
elmlang | general | <@Buffy> yeah I had that idea too and people told me it was bad :sweat_smile: | 2019-03-01T12:07:26.183100 | Nana |
elmlang | general | The benefit would be reduced boilerplate I think | 2019-03-01T12:08:18.183200 | Nana |
elmlang | general | Well, IMO the readability of the code is very much reduced that way. This is of course subjective, but the impression is strong on me. | 2019-03-01T12:09:16.183500 | Dorsey |
elmlang | general | No good reason, the idea just popped into my head last night and I was just curious.
I watched this… and was curious.
<https://www.youtube.com/watch?v=jVeGoqyq5Fs> | 2019-03-01T12:09:40.183800 | Buffy |
elmlang | general | What about this: <https://ellie-app.com/4RGMXYWs5gka1> | 2019-03-01T12:10:11.184000 | Rosalee |
elmlang | general | It’s sort of like how in elm-ui, there used to be a “Seperation” of “styles” and “layout”, but after a while mdgriffith decided against the separation and moved toward co-locating, with a nicer api, styles and layout? Or something? | 2019-03-01T12:11:36.184200 | Buffy |
elmlang | general | <@Rosalee> did you see the second link I posted?
<https://ellie-app.com/4RGJFvfGbzSa1> | 2019-03-01T12:11:53.184400 | Buffy |
elmlang | general | I’m just wondering what sorts of issues might crop up from colocating business logic and views.
Seems very anti SOC but I’m curious if I’ve been brainwashed by OOP to think a certain way and actually it might be better for some things? | 2019-03-01T12:13:18.184600 | Buffy |
elmlang | general | <@Dorsey> heard. | 2019-03-01T12:13:58.184800 | Buffy |
elmlang | general | I like it :slightly_smiling_face: | 2019-03-01T12:15:09.185000 | Rosalee |
elmlang | general | Not sure I do… but I’d be curious to see what it looks like a little bigger… | 2019-03-01T12:16:34.185200 | Buffy |
elmlang | general | Nice thing about MSG being a union is you can do different things based on the type, but you could still do that here…
I’m just not sure how confusing it’d be having business logic mixed in with the views…
Like the Elm compiler would still be nice to work with but I’d see business logic issues colocated with view logic issues? | 2019-03-01T12:18:27.186100 | Buffy |
elmlang | general | I’m wondering if anyone who knew elm before like v0.16 could chime in on this, I have to imagine this sort of thing is something people would’ve already tried? | 2019-03-01T12:19:03.187500 | Buffy |
elmlang | general | Two problems to look out for:
1) The debugger doesn't (presently) have good visibility into closures, so using them in messages or the model can make debugging harder.
2) You must be disciplined in never caching model state inside a closure as part of the update, because there can be delays from when messages are sent to when update handles them and you don't want race conditions. | 2019-03-01T12:19:14.187600 | Dede |
elmlang | general | Bam! :exploding_head: there it is! FTW <@Dede>! | 2019-03-01T12:20:10.188400 | Buffy |
elmlang | general | I almost feel a bit of remorse for asking as if I might’ve brought up memories of past pains. Apologies if that’s the case. | 2019-03-01T12:21:06.189500 | Buffy |
elmlang | general | The debugger problem is just an artifact of the moment; hopefully it will go away in time. The other one is more subtle and pernicious :wink: | 2019-03-01T12:21:22.189800 | Dede |
elmlang | general | I went down a similar road recently, no major pain incurred before redirecting. | 2019-03-01T12:21:38.190300 | Dede |
elmlang | general | And to clarify, when you say caching… are you referring to using like Html.lazy or the like? | 2019-03-01T12:22:02.190800 | Buffy |
elmlang | general | Yeah I tried something similar and encountered race conditions | 2019-03-01T12:22:11.190900 | Nana |
elmlang | general | No, something slightly different. | 2019-03-01T12:22:34.191400 | Dede |
elmlang | general | Wow fascinating. I wonder if either of y’all could provide an ellie example? Maybe we could start like an anti-patterns page in Elm that people could look up for answers to similar questions like this in the future? | 2019-03-01T12:23:11.191600 | Buffy |
elmlang | general | Now that I look closer, you implemented the race condition I had in mind :wink: | 2019-03-01T12:25:48.192800 | Dede |
elmlang | general | Imagine someone mashed the button 10 times. | 2019-03-01T12:25:56.193200 | Dede |
elmlang | general | So fast that all 10 happened before the first message made it to update. | 2019-03-01T12:26:04.193500 | Dede |
elmlang | general | The closure is referring to the external model, not to the argument m. | 2019-03-01T12:26:18.193900 | Dede |
elmlang | general | So you only wind up incrementing 1 time instead of 10. | 2019-03-01T12:26:28.194200 | Dede |
elmlang | general | `(\m -> {m | count = model.count + 1}) ` should not refer to `model` but to `m` to avoid this. | 2019-03-01T12:26:54.194600 | Dede |
elmlang | general | If some of your closures capture model info and are delayed until some slow http request returns, you could wind up applying really stale bits of the model, out of order, when the closures finally do run. | 2019-03-01T12:28:26.195900 | Dede |
elmlang | general | Does that make sense? | 2019-03-01T12:28:29.196100 | Dede |
elmlang | general | So an anti-pattern is certainly closing over `model` in `view`. | 2019-03-01T12:28:45.196500 | Dede |
elmlang | general | Man, that’s clear as a summer day! Makes perfect sense! | 2019-03-01T12:38:32.197100 | Buffy |
elmlang | general | That’d be a super easy mistake to make… In which case giving each function their own signature ala the constructors in the Msg union eliminates that ambiguity. | 2019-03-01T12:39:38.198200 | Buffy |
elmlang | general | Yay! More proof TEA is awesome! | 2019-03-01T12:40:03.198800 | Buffy |
elmlang | general | See also the discussion on this discourse thread: <https://discourse.elm-lang.org/t/message-types-carrying-new-state/2177> | 2019-03-01T13:04:31.199000 | Carman |
elmlang | general | Richard Feldman's response in particular is worth reading | 2019-03-01T13:05:07.199300 | Carman |
elmlang | general | I started writing Elm in v0.15. Happy to answer questions about how things were done in "the old days" | 2019-03-01T13:07:30.199500 | Carman |
elmlang | general | In elm/time, why is `now` a `Task x Posix` and not a `Task Never Posix`? | 2019-03-01T14:42:00.200300 | Angella |
elmlang | general | It makes it easier to combine with other tasks | 2019-03-01T14:47:22.201000 | Carman |
elmlang | general | A `Task x Posix` can never fail (similarly to how a `Maybe a` must be nothing and a `List a` must be empty) | 2019-03-01T14:48:33.201900 | Carman |
elmlang | general | Taking `List a` as an example:
```
empty1 : List a
empty1 =
[]
empty2 : List Never
empty2 =
[]
``` | 2019-03-01T14:49:24.202600 | Carman |
elmlang | general | `empty1` can more easily be combined with other types of lists. You could easily say `[1, 2, 3] ++ empty1` or `List.sum empty1` | 2019-03-01T14:50:51.203800 | Carman |
elmlang | general | while if you wanted to do that with `empty2` you'd have to first map the `never` function | 2019-03-01T14:51:43.204300 | Carman |
elmlang | general | This is the same reason Html snippets with no event handlers are usually given the type `Html a` rather than `Html Never` | 2019-03-01T14:52:29.205100 | Carman |
elmlang | general | It would be nice if we could write a version of Task.perform that could somehow understand that. How might that work? I guess the type would need to somehow say, "I'll only accept something with an unspecified type parameter"? I'm trying to wrap my brain around that. | 2019-03-01T15:08:53.210700 | Angella |
elmlang | general | It does? | 2019-03-01T15:17:50.210900 | Carman |
elmlang | general | `Task.perform` works with `Time.now` even though it is a `Task x Posix` | 2019-03-01T15:18:11.211100 | Carman |
elmlang | general | See <https://ellie-app.com/4RKDqH89v4Sa1> for a live demo | 2019-03-01T15:18:19.211300 | Carman |
elmlang | general | the autolocking of topics on Elm's Discourse after 10 days seems a bit aggressive. example: i just watched Dan Abrams' talk in Oslo from 1 week ago and was compelled to explore his ideas further, but couldn't post in his reference topic that's just 1 month old: <https://discourse.elm-lang.org/t/elm-media-control-api-proposal/3006> | 2019-03-01T16:56:01.214400 | Jarvis |
elmlang | general | another example is that i used someone's Parent-Child proposal heavily in production and i had some feedback, found the Discourse topic where they debuted it, but was unable to provide feedback because it was a couple months old. | 2019-03-01T17:10:03.218100 | Jarvis |
elmlang | general | can anyone point me on the right direction, I am trying to convert someone else's library that never was updated from 0.18. In the 0.18 code they were encoding a list of Value (JSON "Value") using Encode.list [ Value, Value etc. ]. The 0.19 version of Encode requires you to supply the encoder and list. What encoder would be used to encode a list of JSON Value? | 2019-03-01T17:24:42.223800 | Flossie |
elmlang | general | <@Pam> E.list takes (a -> Value) so wouldn't identity function be sufficient? | 2019-03-01T17:27:34.224300 | Jarvis |
elmlang | general | you have List Value, so you need a function (Value -> Value) | 2019-03-01T17:27:53.224700 | Jarvis |
elmlang | general | I'm not sure .. I will try it. | 2019-03-01T17:28:26.225300 | Flossie |
elmlang | general | Yeah, that did satisfy the compiler. No idea yet if the end result is correct, but thanks. At least I have something to work with. | 2019-03-01T17:30:50.226500 | Flossie |
elmlang | general | it's one thing i like about this sort of type system. you see the signature `list : (a -> Value) -> List a -> Value`, you know you have a List Value, so you substitute the other `a` with `Value` and see you just need a function `Value -> Value`. | 2019-03-02T02:25:10.229000 | Jarvis |
elmlang | general | as opposed to shopping for a method on the datastructures that might do what you want | 2019-03-02T02:25:57.229900 | Jarvis |
elmlang | general | for some reason it took me a while to appreciate it | 2019-03-02T02:26:52.230100 | Jarvis |
elmlang | general | Hi, I'm using the elm/svg package and I want to render an image inside some svg. I figured something like this would work
`Svg.image [ Html.Attributes.src "image.png", Html.Attributes.width 100, Html.Attributes.height 100 ] []` but nothing appears. I looked at some html examples online and they seem to be including an `xlink:href="image.png"` attribute in their image. The `Html.Attributes` module doesn't seem to have that so what do I do instead? | 2019-03-02T05:19:27.231600 | Jae |
elmlang | general | <@Jae> you can use `Html.Attributes.attribute` to put any arbitrary attribute | 2019-03-02T05:20:36.232300 | Lynne |
elmlang | general | Ah, I didn't notice that. Thanks! | 2019-03-02T05:20:55.232600 | Jae |
elmlang | general | Oh, also there is a Html.Attributes.href (not xlink:href) but when I use that I get a runtime exception with the message "TypeError: Cannot assign to read only property 'href' or object '#<SVGImageElement>'". Is this a known bug? | 2019-03-02T05:23:01.234700 | Jae |
elmlang | general | I don't know much about SVG but I can imagine that one is required putting a namespaced attribute (xlink:href) and hence setting just href may be prohibited | 2019-03-02T05:26:40.235600 | Lynne |
elmlang | general | There is also `Svg.Attributes.xlinkHref` - that might by something you are actually after <@Jae> | 2019-03-02T05:28:38.236200 | Lynne |
elmlang | general | That can't be used inside Svg.image though. Svg.image expects Html.Attributes | 2019-03-02T05:30:17.236800 | Jae |
elmlang | general | Both `Html.Attribute` and `Svg.Attribute` are type aliases for `VirtualDom.Attribute`, so that can't be true <@Jae> | 2019-03-02T05:32:28.237600 | Lynne |
elmlang | general | Dang, I never realized that. That solved my problem then, the image appears. Thanks! | 2019-03-02T05:33:28.238200 | Jae |
elmlang | general | I'm looking for approaches to lazily populating my model based on the view. In short, one of my views needs to display the asciidoc-generated HTML from a string in my model. The natural place to request the (relatively expensive) generation is when the view actually needs it, but since generation requires sending a message out a port, the view - as far as I can tell - can't make the request. Other approaches I can think of (e.g. having my update determine when a URL change will require the generation) require putting knowledge about specific views in places besides the view. Any ideas on how I can make the generation request only when the view actually needs that data? | 2019-03-02T06:25:23.242700 | Winnie |
elmlang | general | Do you parse URL in your `view` function now and render something depending on what you parsed? Or how does your app work currently? | 2019-03-02T06:33:44.243100 | Lynne |
elmlang | general | yes, the URL is stored in the model and I sort out what to render in the view based on that. It's all taken pretty directly from the navigation documentation for 0.19. | 2019-03-02T06:36:37.243300 | Winnie |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.