workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
elmlang | general | apart from that, I assume that easiest way to handle association would be to store associations via IDs (UUIDs) and simply pull'em from the store for render purposes | 2019-01-07T17:10:19.738000 | Floy |
elmlang | general | ? | 2019-01-07T17:10:21.738200 | Floy |
elmlang | general | <@Floy> you should store the values once and store IDs to refer to them | 2019-01-07T17:14:20.738400 | Earlean |
elmlang | general | so will my assumption be correct that in order to correctly handle, let's say `delete`, I ought to have dedicated functions to know exactly what to delete (class) and where (classId in all students)? | 2019-01-07T17:18:48.738700 | Floy |
elmlang | general | yeah, I think this might be correct answer | 2019-01-07T17:28:19.739000 | Floy |
elmlang | general | You can certainly do that. But invalid IDs are fine to have. | 2019-01-07T17:29:15.739200 | Earlean |
elmlang | general | if you have time, I wonder about one extra thing - will I be correct to assume that each view wouldn't exactly store the model (i.e. copy or "lens" of the store). I should pass that to view/update only and maybe just have some convenient function to materialize slice of the store for view purposes? | 2019-01-07T17:30:21.739400 | Floy |
elmlang | general | i.e. `viewStudent : Store -> List StudentView` ? | 2019-01-07T17:30:53.739600 | Floy |
elmlang | general | Yep, that's a common strategy | 2019-01-07T17:34:19.739800 | Earlean |
elmlang | general | :+1: | 2019-01-07T17:41:04.740000 | Floy |
elmlang | general | ok, I will try to find time and build something against those recommendations, if I may call them so | 2019-01-07T17:41:28.740200 | Floy |
elmlang | general | thx a lot for quick chat | 2019-01-07T17:41:33.740400 | Floy |
elmlang | general | Is it possible to do something like this: `user = { model.user | firstName = "John" }`? My intuition says it should be, but my compiler says it is not. | 2019-01-07T21:11:57.741700 | Marcus |
elmlang | general | Feels more succinct to do it that way than to break out `model.user` as an intermediate variable, which works. | 2019-01-07T21:12:58.742400 | Marcus |
elmlang | general | ```
user = model.user |> \u -> { u | firstName = "John" }
```
I’ve seen someone suggest this before. | 2019-01-07T21:31:05.742600 | Lizabeth |
elmlang | general | Ah, not a bad approach. Thanks. | 2019-01-07T21:39:34.742800 | Marcus |
elmlang | general | I'm wondering if there is a more idiomatic way to do what I'm doing below. This is one of the case statements in my update function:
```
NewName name ->
let
newModel = { model | name = newName }
in
( newModel
, persistModel newModel
)
```
I have this pattern a lot, where I create an intermediate variable for the model, then return that and a command that uses is. | 2019-01-07T22:25:14.745600 | Marcus |
elmlang | general | I guess I could do something like:
```
{ model | name = newName }
|> (\m -> (m, persistModel m))
```
That doesn't really feel much clearer though. | 2019-01-07T22:26:46.747200 | Marcus |
elmlang | general | what does `persistModel` do? | 2019-01-07T22:34:09.747700 | Ruthann |
elmlang | general | i ask because the name implies it's doing something you're already doing by returning `newModel` in the first part of the tuple | 2019-01-07T22:34:54.748400 | Ruthann |
elmlang | general | I simplified the actual code so it isn't domain specific. In my case, it passes the model to an API. | 2019-01-07T22:36:05.749300 | Marcus |
elmlang | general | e.g., for backend storage | 2019-01-07T22:36:24.749600 | Marcus |
elmlang | general | ic and returning `newModel` would be a sort of 'optimistic update' in that case | 2019-01-07T22:37:47.750200 | Ruthann |
elmlang | general | in theory, that looks fine to me, you're updating your model locally (temp) and sending it away (perm)
we do this in places as well. the only other thing i could see is if you only updated your local model when the backend task returned with the true state of the backend after the save | 2019-01-07T22:39:15.751700 | Ruthann |
elmlang | general | It's strictly fire-and-forget. I wasn't crazy about always creating the intermediate variable to hold the updated model, but I don't think there's a better way. | 2019-01-07T22:40:16.752600 | Marcus |
elmlang | general | if you just want it to look cleaner:
```
{ model | name = newName }
|> (,)
|> Tuple.mapSecond persistModel
``` | 2019-01-07T22:41:01.753200 | Ruthann |
elmlang | general | we use pipes `|>` *a lot* in our code, it doesn't always lead to more readable code so up to you to judge | 2019-01-07T22:41:44.754000 | Ruthann |
elmlang | general | The majority of the team is new to FP, so I'll probably stick with the intermediate variable over the pipes. Although, we could stand to embrace pipes more. | 2019-01-07T22:42:55.755200 | Marcus |
elmlang | general | Can you explain the line `|> (,)`? | 2019-01-07T22:43:12.755900 | Marcus |
elmlang | general | hum... actually i don't think that'll work `(,) : a -> b -> (a,b)` your anonymous function is better | 2019-01-07T22:43:15.756000 | Ruthann |
elmlang | general | i definitely found that pipes or too much transformations was a barrier to onboarding | 2019-01-07T22:43:47.756700 | Ruthann |
elmlang | general | ppl tended to hold the whole type in their heads rather than just looking at each transform in turn, or not being able to see the input/output straight away | 2019-01-07T22:44:24.757700 | Ruthann |
elmlang | general | It's an awesome idiom, but an unfamiliar syntax. | 2019-01-07T22:44:34.758000 | Marcus |
elmlang | general | We'll get there. | 2019-01-07T22:44:53.758200 | Marcus |
elmlang | general | Why not make your `persistModel` function return the tuple?
```
persistModel : Model -> (Model, Cmd Msg)
```
then
```
{ model | name = newName }
|> persistModel
```
or just
```
persistModel { model | name = newName }
```
An alternative is to make your anonymous function a generic one, maybe something like:
```
thenDo : (Model -> Cmd msg) -> Model -> (Model, Cmd msg)
thenDo f model =
(model, f model)
```
then
```
{ model | name = newName }
|> thenDo persistModel
```
Or if you want to keep the tuple structure more visible:
```
apply : (a, a -> b) -> (a, b)
apply (a, f) = (a, f a)
```
then
```
apply
( { model | name = newName }
, persistModel
)
```
the most difficult might be to find a good name :sweat_smile: | 2019-01-08T03:24:42.759600 | Velia |
elmlang | general | Hi folks, any good idea to cache http requests? I mean cache the GET request with url and params, so that next time fetch with same url and params, it should not trigger the request. | 2019-01-08T03:57:28.762700 | Raymonde |
elmlang | general | It should apply to all requests through the app. | 2019-01-08T04:00:12.763600 | Raymonde |
elmlang | general | <@Raymonde> I was thinking about this too
you could just use a Dict with the entire url as key I guess | 2019-01-08T04:01:07.764500 | Nana |
elmlang | general | (or the relavant part of the url) | 2019-01-08T04:02:35.765200 | Nana |
elmlang | general | Seems I have to store this Dict in global model, and pass all the way down where there is request command. | 2019-01-08T04:03:06.765700 | Raymonde |
elmlang | general | yup | 2019-01-08T04:03:35.765900 | Nana |
elmlang | general | I’ll think about it, thanks. | 2019-01-08T04:04:23.766200 | Raymonde |
elmlang | general | won't the browser do this for you “for free” if the http server responds with caching headers? | 2019-01-08T04:08:47.766800 | Lewis |
elmlang | general | <@Lewis> not when getting things through scripts I think? | 2019-01-08T04:09:26.767200 | Nana |
elmlang | general | ah. ok. | 2019-01-08T04:09:38.767400 | Lewis |
elmlang | general | Is it an intended/documented behaviour to not using cache when requesting data from scripts? It is plain old HTTP after all. Here, for example, it is said that resources downloaded through fetch are subject to the HTTP cache: <https://hacks.mozilla.org/2016/03/referrer-and-cache-control-apis-for-fetch/> | 2019-01-08T04:25:52.769600 | Lynne |
elmlang | general | XHR requests do adhere to the normal HTTP caching too. | 2019-01-08T04:32:29.771500 | Bert |
elmlang | general | You need to set no-cache headers for GET requests when you want to bypass the cache. | 2019-01-08T04:33:36.772900 | Bert |
elmlang | general | Which incidentally many times is what you want when dealing with an API. | 2019-01-08T04:34:30.774500 | Bert |
elmlang | general | And backend can be set up to prevent same requests from being sent by setting some response headers | 2019-01-08T04:34:59.775200 | Lynne |
elmlang | general | ah I see :sweat_smile: | 2019-01-08T04:46:08.775600 | Nana |
elmlang | general | Is that means it doesn’t need to cache the requests in model, if set cache control correctly? | 2019-01-08T04:52:13.776800 | Raymonde |
elmlang | general | Yes, that's how it is supposed to be done if you control both frontend and backend | 2019-01-08T04:52:37.777200 | Lynne |
elmlang | general | The question is, however, if Elm's HTTP implementation allows to NOT set no-cache headers - it may be hardcoded | 2019-01-08T04:54:18.778200 | Lynne |
elmlang | general | As <@Bert> mentioned many times when you dealing with API you want to bypass the cache | 2019-01-08T04:54:54.778900 | Lynne |
elmlang | general | Alright, I checked the api in my app, and noticed there is a request header `cache-control: no-cache`. According to the article you post, I can set cache mode to `force-cache` | 2019-01-08T04:59:30.780700 | Raymonde |
elmlang | general | There are no headers in `elm/http` requests by default. | 2019-01-08T05:00:10.781600 | Bert |
elmlang | general | But I don’t see a place to set cache mode with elm/http | 2019-01-08T05:00:30.782200 | Raymonde |
elmlang | general | `Http.request` accepts `headers` as a record field | 2019-01-08T05:01:19.782900 | Lynne |
elmlang | general | Please also note that setting up cache is not limited only to `cache-control` header. There are few other headers which needs to be set up in request and response for caching to be reliable. | 2019-01-08T05:02:27.784300 | Lynne |
elmlang | general | Ah, that seems many things to concern. | 2019-01-08T05:03:54.785200 | Raymonde |
elmlang | general | It is not that complicated but, yes, it does require some learning | 2019-01-08T05:04:36.785700 | Lynne |
elmlang | general | Some server-side frameworks make it easy by taking single configuration option and then setting up responses for you. So you might want to check docs for whatever is on your backend. | 2019-01-08T05:05:59.786500 | Lynne |
elmlang | general | I only control our frontend, I’ll learn http cache first. Thanks. | 2019-01-08T05:08:36.787500 | Raymonde |
elmlang | general | btw, something that's puzzled me a bit is all the talk about using Service Workers for caching
they always seem to treat caching as something completely new, never mentioning the traditional HTTP caching | 2019-01-08T05:54:43.789600 | Nana |
elmlang | general | service workers offer much more granular control over what and when resources are cached | 2019-01-08T05:57:51.790100 | Danika |
elmlang | general | <https://code.fb.com/web/web-performance-cache-efficiency-exercise/> here’s an interesting internal study at facebook | 2019-01-08T05:59:29.790600 | Danika |
elmlang | general | > After a few weeks of collecting data and letting caches fill up, we looked back over the past seven days’ worth of data. The initial results were surprising to us: 25.5% of all logged requests were missing the cache. We split the data by interface, desktop and mobile, and still saw the same breakdown: 24.8% of desktop requests and 26.9% of mobile were missing the cached image. Not what we expected, so we dug in more. | 2019-01-08T05:59:32.790900 | Danika |
elmlang | general | This is with standard HTTP caches | 2019-01-08T05:59:41.791100 | Danika |
elmlang | general | hmm are they saying that 25% should have had the image cached but didn't? or is it just that 25% hadn't seen the image before? | 2019-01-08T06:06:20.791800 | Nana |
elmlang | general | Should’ve been cached | 2019-01-08T06:08:08.792000 | Danika |
elmlang | general | why would that be though? because the browser decided to free up the harddrive space? | 2019-01-08T06:10:00.793000 | Nana |
elmlang | general | Browsers ultimately discard the cache when they feel like it. | 2019-01-08T06:10:34.793300 | Danika |
elmlang | general | like, surely it's not just "HTTP caching is buggy and fails 20% of the time" | 2019-01-08T06:10:48.793700 | Nana |
elmlang | general | Yes, its not that HTTP caching is *necessarily* the evil one, but its more that browsers treat service worker cache and HTTP cache differently | 2019-01-08T06:11:23.794400 | Danika |
elmlang | general | <https://jakearchibald.com/2016/caching-best-practices/> this article is getting a bit on the old side now, but using service worker and http cache in tandem is a reasonable strategy | 2019-01-08T06:11:56.795200 | Danika |
elmlang | general | just seems weird to use service workers for something such basic things, and especially for Google to promote it. couldn't they just change the browser settings? | 2019-01-08T06:12:31.796000 | Nana |
elmlang | general | or is it that too many websites are telling the browser to cache things, so you should use Service Workers to tell the browser you *really* want to cache it? | 2019-01-08T06:15:51.797000 | Nana |
elmlang | general | I mean ultimately service workers and http caches are for different things, right.
You http cache to save redownloading assets that havent changed. But service workers power PWAs which should/need to function like native apps. The expectation is different: pwas need to be functional offline but a cached website doesn’t offer the same assurance. | 2019-01-08T06:21:06.798600 | Danika |
elmlang | general | Service Workers can be used to make the app run and have content offline. | 2019-01-08T09:29:59.799900 | Bert |
elmlang | general | Oh, I didn't see the newest message. Yes! | 2019-01-08T09:31:21.801200 | Bert |
elmlang | general | I’d argue thats what they *should* be used for, not just that they _can_ be | 2019-01-08T09:31:23.801300 | Danika |
elmlang | general | yeah | 2019-01-08T09:31:25.801600 | Danika |
elmlang | general | Has anyone experimented with elm-ui ? If so what's your opinion about it? and how do you handle media queries and animation? | 2019-01-08T10:27:54.802800 | Quintin |
elmlang | general | this elm-ui <https://package.elm-lang.org/packages/mdgriffith/elm-ui/1.1.0/> | 2019-01-08T10:28:17.803200 | Quintin |
elmlang | general | I've used elm-ui on my game (<https://github.com/JoelQ/muses>) but haven't done animation or media queries yet :slightly_smiling_face: | 2019-01-08T10:31:16.804300 | Carman |
elmlang | general | <@Carman> did you enjoy elm-ui? or did you ran into problems? | 2019-01-08T10:32:07.804800 | Quintin |
elmlang | general | I did enjoy it! Layout in particular is much more intuitive to me than standard CSS | 2019-01-08T10:33:02.805500 | Carman |
elmlang | general | I agree with the premise that css and app ui is not really a great match, but looking at the docs, I still get the feeling that elm-ui doesnt entirely replace elm-css :disappointed: | 2019-01-08T10:35:10.806300 | Quintin |
elmlang | general | but I might be missing something | 2019-01-08T10:35:16.806500 | Quintin |
elmlang | general | what are you missing? please also be aware of <#C4F9NBLR1|elm-ui> | 2019-01-08T10:44:08.807600 | Desire |
elmlang | general | <@Desire> Im still looking at the docs, but out of the box, im missing how to deal with desktop/mobile components (media queries) and animations/transitions, I might be missing something ofc :smile: And thanks for the channel tip, I had no idea it existed | 2019-01-08T10:51:41.809000 | Quintin |
elmlang | general | animatons are in a different package | 2019-01-08T10:53:31.809800 | Desire |
elmlang | general | and media queries are supported afaik | 2019-01-08T10:54:21.810600 | Desire |
elmlang | general | if they are, then im missing something (but im glad they are supported, so I just need to search a bit better :smile: ) | 2019-01-08T10:56:25.811300 | Quintin |
elmlang | general | im moving this conversation to <#C4F9NBLR1|elm-ui> to avoid more noise thanks! :smile: | 2019-01-08T10:56:39.811700 | Quintin |
elmlang | general | Packages aren’t allowed to use native code, but am I right in thinking that `ports` are fair game? | 2019-01-08T11:00:13.813000 | Danika |
elmlang | general | Port are your interface to write native code | 2019-01-08T11:02:58.813500 | Agustin |
elmlang | general | I don’t think you can ship ports in a package though if thats what you are asking | 2019-01-08T11:03:16.814000 | Agustin |
elmlang | general | That’s a shame, that is what i was asking yes | 2019-01-08T11:04:23.814600 | Danika |
Subsets and Splits