workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
elmlang | general | hmm^^ | 2019-04-12T05:51:11.028700 | Emilee |
elmlang | general | Also it turns out that firefox *does* actually still support `xlink:href`, so I could've ignored this whole weirdness^^ but it's still nice to follow the current standard :slightly_smiling_face: | 2019-04-12T05:51:57.028900 | Emilee |
elmlang | general | thinking how to make an API with lots of options, where some are optional and others are not :thinking_face:
it'd be really nice if the old extensible records were still in, then I could have done something like:
```
type alias Options =
{ foo : String --required
, bar : Maybe String --optional
}
defaults = { bar = Nothing }
```
and then you would use it like:
```
Foo.view {Foo.defaults | foo = "foo"}
```
that means you would get a type error if you didn't specify all required options, but the optional ones you could leave out | 2019-04-12T06:23:21.033300 | Nana |
elmlang | general | You can't do this anymore? | 2019-04-12T06:24:03.033500 | Danika |
elmlang | general | Oh wait I can see why | 2019-04-12T06:24:24.034000 | Danika |
elmlang | general | no, you're not allowed to change the type of a record - it was removed to enable easier garbage collection in the future | 2019-04-12T06:24:50.034700 | Nana |
elmlang | general | ```
type alias Required =
{ a
| foo : String
}
type alias Optional a =
{ a
| bar : Maybe String
}
type alias Foo = Optional Required
defaults = { bar = Nothing }
withDefault : Required -> Foo
withDefault req =
{ defaults | req }
```
Does this work, or something similar? | 2019-04-12T06:29:10.043700 | Danika |
elmlang | general | I guess I could achieve the same thing by having my function accept two records, one with the required stuff and the other with optional stuff, and expose a default record for the optional ones | 2019-04-12T06:29:16.044100 | Nana |
elmlang | general | i dont think `{ defaults | req }` is valid syntax though | 2019-04-12T06:30:08.045700 | Danika |
elmlang | general | :thinking_face: | 2019-04-12T06:31:54.047100 | Nana |
elmlang | general | <@Nana> You might want to look how the generated code for <https://package.elm-lang.org/packages/dillonkearns/elm-graphql/4.2.0/> handles optional arguments. I’m AFK right now, so maybe best look at the examples. | 2019-04-12T06:40:21.052400 | Jin |
elmlang | general | <@Nana> You could have `defaults : Required -> Optional Required`, and then have `Foo.view (Foo.defaults { foo = "abc" } |> Foo.setBar "123")`. Would that help? | 2019-04-12T06:40:55.052600 | Kenya |
elmlang | general | I think that would achieve the same thing as having the function take two records, like:
```
Foo.view {foo = "foo"} {Foo.defaults | bar = Just "bar"}
``` | 2019-04-12T06:43:50.054000 | Nana |
elmlang | general | ah, so the way elm-graphql does it would be like this?
```
Foo.view {foo = "foo"} (\defaults -> {defaults | bar = Just "bar"})
``` | 2019-04-12T06:46:30.056400 | Nana |
elmlang | general | and you can use `identity` if you don't want to change any defaults | 2019-04-12T06:47:19.056700 | Nana |
elmlang | general | ```
> import Foo
> Foo.withDefaults { a = "Hello World!", b = 0.5, c = 10 }
{ a = "Hello World!", b = 0.5, c = 10, d = Nothing, e = Nothing, f = Nothing } : Foo.Options
``` | 2019-04-12T06:50:57.057300 | Danika |
elmlang | general | I don't think I'd recommend doing this...
```
type alias Required =
{ a : String
, b : Float
, c : Int
}
type alias Optional required =
{ required
| d : Maybe String
, e : Maybe Int
, f : Maybe Float
}
type alias Options = Optional Required
withDefaults : Required -> Options
withDefaults r =
{ a = r.a
, b = r.b
, c = r.c
, d = Nothing
, e = Nothing
, f = Nothing
}
``` | 2019-04-12T06:51:32.057600 | Danika |
elmlang | general | Exactly | 2019-04-12T06:52:24.057700 | Huong |
elmlang | general | <@Danika> that doesn't let you specify some of the optional arguments though? | 2019-04-12T06:54:06.058800 | Nana |
elmlang | general | indeed | 2019-04-12T06:54:21.059000 | Danika |
elmlang | general | <@Nana> Do you have functions that work on different records of optional arguments? Then the two argument approach would be the only one that works, or? | 2019-04-12T06:56:37.061000 | Kenya |
elmlang | general | but I guess having two functions like `view : Required -> Html msg` and `withExtra: Required Optional -> Html msg` could be nice as a convenience | 2019-04-12T06:57:12.061800 | Nana |
elmlang | general | Damn, Slate.js requires React | 2019-04-12T08:55:25.062700 | Hoa |
elmlang | general | Well, Quill might fit the task equally well. | 2019-04-12T08:58:00.063200 | Hoa |
elmlang | general | the only editor I've used before is TinyMCE, which I think was like 600 kb by itself :grin: not exactly "Tiny"
but it had good features for restricting and cleaning up html | 2019-04-12T08:58:03.063400 | Nana |
elmlang | general | I’ve used Quill a few years ago with React. I had a good impression back then :slightly_smiling_face: | 2019-04-12T08:59:44.063600 | Hoa |
elmlang | general | Oh wait. It looks ProseMirror is maintained too. | 2019-04-12T09:01:44.063800 | Hoa |
elmlang | general | <@Nana> Have you considered a builder pattern? Something like
```
module Thing
...
builder: RequiredParameters -> Builder
...
build: Builder -> Thing
...
setOptA: Int -> Builder -> Builder
...
setOptB: Bool -> Builder -> Builder
```
And use looks like:
```
Thing.builder(requiredParameters) |> Thing.setOptA 17 |> Thing.setOptB False |> Thing.build
``` | 2019-04-12T09:45:26.067600 | Dede |
elmlang | general | <@Dede> that just seems more verbose though :thinking_face: | 2019-04-12T09:46:37.067900 | Nana |
elmlang | general | Yeah, it’s a pattern I see commonly in languages that let you attach operations to datatypes, where you’d write
```
Thing::builder(required).a(17).b(False).build()
``` | 2019-04-12T09:49:04.069000 | Dede |
elmlang | general | That said, if there are dependencies between the optional arguments, this gives you a way to enforce them. | 2019-04-12T09:49:34.069800 | Dede |
elmlang | general | You should be able to copy <https://pursuit.purescript.org/packages/purescript-options/4.0.0/docs/Data.Options> and use a similar idiom | 2019-04-12T09:50:19.070900 | Niesha |
elmlang | general | E.g. setters can return different builder types so that your early choices condition what your later options are. | 2019-04-12T09:50:24.071100 | Dede |
elmlang | general | I like this approach as well. I use it here <https://github.com/z5h/component-result/blob/1.1.0/src/ComponentResult.elm#L73>
Then use it like
```
withModel myModel
|> withCmd myHttpGet
|> withCmd myPortCmd
|> withExternalMsg LoadingData
``` | 2019-04-12T09:56:53.071900 | Leoma |
elmlang | general | Oh yeah, wordiness varies a lot depending on whether you need to keep module prefixing or not. | 2019-04-12T09:57:48.072400 | Dede |
elmlang | general | `builder(requiredParameters) |> setOptA 17 ...` is not so bad. | 2019-04-12T09:58:11.073300 | Dede |
elmlang | general | (Ha, I almost always keep the module prefix… Just put down that example to keep it simple). | 2019-04-12T09:58:49.074300 | Leoma |
elmlang | general | This is an area where Rust-ish operation-type associations have a code density advantage. | 2019-04-12T09:59:18.075000 | Dede |
elmlang | general | …which may or may not be a readability disadvantage :wink: | 2019-04-12T09:59:29.075300 | Dede |
elmlang | general | Readability > all, imo. :slightly_smiling_face: | 2019-04-12T10:01:10.075600 | Rochell |
elmlang | general | :100: . Although sometimes brevity helps, brevity isn’t necessarily the rule. | 2019-04-12T10:02:02.075700 | Leoma |
elmlang | general | but people will always disagree on what is more or less readable :stuck_out_tongue: | 2019-04-12T10:02:55.075900 | Nana |
elmlang | general | If brevity is concise and easily readable, then awesome. But if you have to start thinking about how the code is phrased, rather than why it's phrased that way, then it's a stumbling block to understanding. :slightly_smiling_face: | 2019-04-12T10:02:59.076100 | Rochell |
elmlang | general | Yeah, that's true enough. | 2019-04-12T10:03:05.076300 | Rochell |
elmlang | general | It's an inexact art :slightly_smiling_face: | 2019-04-12T10:03:29.076500 | Rochell |
elmlang | general | and also, keeping types simple | 2019-04-12T10:04:07.076700 | Nana |
elmlang | general | json-decode-pipeline for example is famously hard for beginners to understand, even though it looks pretty readable | 2019-04-12T10:05:34.076900 | Nana |
elmlang | general | The Rust attitude to keeping type simple is to clone the Elm type system, then do a complete tableflip by adding Traits and automatic coercions :wink: | 2019-04-12T10:06:02.077100 | Dede |
elmlang | general | (I spend way more time fighting the type system than the borrow checker, even though the latter is what looks scary at first.) | 2019-04-12T10:06:33.077300 | Dede |
elmlang | general | hello | 2019-04-12T10:09:18.077700 | Jefferson |
elmlang | general | Rust just cannot be as easy as Elm | 2019-04-12T13:43:39.079500 | Myrna |
elmlang | general | The non-trait type system is more or less identical to Elm’s type system. But after that, there’s just a whole lot more stuff. | 2019-04-12T13:44:24.079700 | Dede |
elmlang | general | multi-paradigm, lifetime, mixed of side effects | 2019-04-12T13:44:28.079900 | Myrna |
elmlang | general | One may think it must be good for a language can do all of this | 2019-04-12T13:44:53.080100 | Myrna |
elmlang | general | But it is just opposite. Like C++, for a team using C++, half of team will not understand full spec of the language for just using it, not counting the domain specific problems... | 2019-04-12T13:46:00.080300 | Myrna |
elmlang | general | Another example is data and object (opaque data). | 2019-04-12T13:47:19.080600 | Myrna |
elmlang | general | Only the clean separation between two can yield a good program.. | 2019-04-12T13:47:45.080800 | Myrna |
elmlang | general | So, I’m enjoying using Rust, and I’m not really interested in a language slag-off. It’s very different from Elm, serving a different space. I’ll leave it at that. | 2019-04-12T13:49:42.081100 | Dede |
elmlang | general | yeah, it is just my opinion. My main work language is C. | 2019-04-12T13:50:53.081300 | Myrna |
elmlang | general | I am just have some rants to what I have seen in the industry code. My bad. | 2019-04-12T13:51:27.081500 | Myrna |
elmlang | general | I find Rust much more tractable than C++ has become. (I’m also fond of C for its purity of essence., but I’d write in Rust over C because I like the guardrails.) | 2019-04-12T13:51:57.081700 | Dede |
elmlang | general | what do you think has more overhead:
• creating single-rule style tags
• creating custom elements
context: we're currently using custom elements for our svg icons because they are colored with css variables which are not supported in elm. an alternative would be to use pure elm and dynamically create a style tag for every icon defining the css variable
i would like to use a pure elm approach, but i'm worried about the wonkyness and overhead of ad-hoc creating a lot of style tags | 2019-04-12T16:27:57.084800 | Emilee |
elmlang | general | <@Emilee> I think there may be a workaround with the CSS variables with
```
Html.Attributes.attribute "style" "--some-var: 10px"
``` | 2019-04-12T16:44:40.088200 | Alicia |
elmlang | general | we have SVG icons as well, we used `fill: currentColor` for them, and then we wrap them in a span or whatever to set the color | 2019-04-12T16:47:21.088800 | Alicia |
elmlang | general | like `span [ class "w-4 h-4 color-blue" ] [ Icon.decorativeIcon SomeIconName ]` | 2019-04-12T16:47:46.089300 | Alicia |
elmlang | general | hmm, the attribute workaround seems feasible. of course it can accidentally nuke other styles, but that should be avoidable
`currentColor` seems like a better choice for single-color icons than a css variable, but we can't change that right now since parts of our angular codebase depend on the way colors are set at the moment^^ | 2019-04-12T17:12:58.090900 | Emilee |
elmlang | general | Hi, I constantly feel that structure and presentation are two different respects of a HTML view | 2019-04-12T18:40:25.093400 | Myrna |
elmlang | general | mix css style directly inside the code make my html structure very hard to read | 2019-04-12T18:41:03.094200 | Myrna |
elmlang | general | Does anyone have suggestion for that situation? | 2019-04-12T18:41:42.094900 | Myrna |
elmlang | general | Right now, I use bootstrap classname in the attribute tag for html, and have separate css for it. But bootstrap requires me to modify my html structure as well.. | 2019-04-12T18:43:12.096200 | Myrna |
elmlang | general | We've been using tailwinds which does utility classes (e.g `p-4 font-semibold color-blue`) , I love it now and try to use as little css as possible. It has a little learning curve but once you are over that development is fluid | 2019-04-12T18:49:46.100400 | Augustus |
elmlang | general | Bulma is another one that is similar to tailwinds | 2019-04-12T18:51:46.101500 | Augustus |
elmlang | general | <@Augustus> I agree, utilities classes are cool | 2019-04-12T18:54:13.102000 | Myrna |
elmlang | general | very easy to manipulate as well | 2019-04-12T18:54:27.102400 | Myrna |
elmlang | general | but feels like it is kind of abbreviation or alias to write css style attribute in the html, isn't it | 2019-04-12T18:56:41.103500 | Myrna |
elmlang | general | of course, shorter, sweater, I looked up the grid css | 2019-04-12T18:58:24.104200 | Myrna |
elmlang | general | the two probably can combine together, one for visual layout, the other for visual style | 2019-04-12T18:58:58.104900 | Myrna |
elmlang | general | Separating layout and style was how `style-elements`(now `elm-ui`) started :smile: | 2019-04-12T19:31:34.105600 | Cornell |
elmlang | general | I’m actually working on this now. Currently on vacation, but hope to land this in in near future | 2019-04-13T03:01:13.108800 | Patricia |
elmlang | general | It’s hard stuff. I’ve been using CSS since forever and there’s always a bit overlapping. Bootstrap-like frameworks are good solution since they do most of the job for you. You pay some price for it though, since they require you write HTML and add cSS classes to match their patterns. I personally do’t like tailwind-like solutions but they have a lot of fans. Given that we have flexbox and CSS grids are starting to be more approachable I would think to start with something basic like Normalize.css and build from there with Sass and build “components” using BEM like you would in React. So you end up having something like `Card.scss` holding `Card`, `Card__heading`, `Card__content`, etc. Utility classes like `text-right` are nice to have. Separate layout and styles may be good too, for theming. Sites like <https://css-tricks.com> gives a lot of insights to cope with HTML/CSS complexity. | 2019-04-13T05:26:08.109800 | Hoa |
elmlang | general | How do you all go about benchmarking elm code? I'm writing a game and the logic running in each animation frame is taking 40ms when it should be less than 16ms. | 2019-04-13T06:51:25.114400 | Jae |
elmlang | general | I know it has to be a CPU bottleneck because there was no performance improvements when I modified my view function to not make an dom changes. | 2019-04-13T06:52:15.114500 | Jae |
elmlang | general | I looked at the performance tab in Chromes dev tools but it just tells me A4, anonymous, and garbage collection are taking most of the time. It doesn't give me any understandable function names | 2019-04-13T06:53:09.114700 | Jae |
elmlang | general | Is my only option to slowly remove different parts of my code until I notice a speedup? | 2019-04-13T06:54:13.114900 | Jae |
elmlang | general | the bottom-up tab gives functions with the most self time. you can click on them to go to their definition where you can see the actual name | 2019-04-13T06:55:05.115100 | Virgie |
elmlang | general | do you by any chance do record updates in your loop? they are known to be slow | 2019-04-13T06:56:16.115300 | Virgie |
elmlang | general | Bottom up looks like this for me. | 2019-04-13T06:57:51.115500 | Jae |
elmlang | general | I probably do have record updates all over the place | 2019-04-13T06:58:09.115900 | Jae |
elmlang | general | The alternative to record updates is to recreate the record manually instead of using the update syntax? | 2019-04-13T06:58:50.116100 | Jae |
elmlang | general | yes | 2019-04-13T07:01:22.116300 | Virgie |
elmlang | general | try that first, I think that will help a lot. Then the debugger takes a lot of time (so don't do performance measurement while in --debug) | 2019-04-13T07:02:37.116500 | Virgie |
elmlang | general | perhaps you could let it run for a little while to get more robust numbers | 2019-04-13T07:02:52.116700 | Virgie |
elmlang | general | Definitely turn off the debugger if you want to have any clue about performance. It serializes the entire state, in a not-super-performant way, on every message. If you have a subscription to animation-frames and non-trivial state, that'll kill performance. | 2019-04-13T07:05:28.116900 | Huong |
elmlang | general | Turning off the debugger completely fixed the issue. | 2019-04-13T07:06:33.117100 | Jae |
elmlang | general | `--optimize` can also improve performance a bunch, especially when you have lots of boxed types (custom types with a single constructor) | 2019-04-13T07:08:11.117300 | Huong |
elmlang | general | True. I want to avoid that though because then I lose hot reloading and Debug.log | 2019-04-13T07:09:22.117500 | Jae |
elmlang | general | In the future, if I have the debugger off and I've avoided using the update syntax inside loops, is there a good way of figuring out what elm functions are taking a lot of CPU time? | 2019-04-13T07:11:53.117700 | Jae |
elmlang | general | I'm worried I'm going to run into performance issues and then have to resort to picking apart my app when trying to find the hot path so I can know what to optimize | 2019-04-13T07:13:02.117900 | Jae |
elmlang | general | I've found the bottom up tab to be helpful (click on the blue links/line numbers to go to the actual functions and see their name) | 2019-04-13T07:13:45.118100 | Virgie |
elmlang | general | another useful thing is using what <@Virgie> describes to figure out hot code-paths, picking out the largest offender, benchmarking it and figuring out how to optimize that. If you ever get to that point, feel free to poke me and I'll gladly try and help you out! | 2019-04-13T07:16:07.118300 | Huong |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.