workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
elmlang | general | Here's what I found as well, Elm looks very promising. | 2019-04-08T14:32:10.742500 | Dayna |
elmlang | general | TBH, I am a bit surprised how well Elm fares in those benchmarks. :elm: | 2019-04-08T14:32:36.742700 | Timika |
elmlang | general | Take a look at bundle size. Elm compiles to a very tiny resource. I cannot get our CSS close to the small size of our elm code. | 2019-04-08T14:36:46.742900 | Charity |
elmlang | general | For context, I took one of our apps and added a service worker and a manifest and it passed all of the PWA audits on the first run. | 2019-04-08T14:37:26.743100 | Charity |
elmlang | general | Btw I also love how Elm makes your css way smaller as well. If you use elm-css and minify the code all your css properties will become single letters. | 2019-04-08T14:38:52.743400 | Dayna |
elmlang | general | I've seen several different ways of structuring larger applications, but I am not satisfied with any of them so far. Richard's SPA example is okay for a stateless application with mutually exclusive "pages", but a stateful one will need to take a different approach. The approach taken by [this game](<https://github.com/Lattyware/massivedecks/blob/master/client/src/MassiveDecks/Scenes/Start.elm>) is not bad. I combined these two approaches for my current application, but I still find it lacking. For example, I don't want to load the actual application until the user is logged in, which means I'll need to create a sum type `type Screen = LoginScreen Login.Model | AppScreen App.Model`, and this continues throughout the application, resulting in many layers of nesting inside of sum types in order to guarantee certain conditions are met. This also avoids the problem where everything is wrapped in `Maybe`, but I still feel like there should be a better way to do it.
For reference, you can find the application I am building here <https://github.com/ericnething/roll2d6-client>. It's ~14k loc right now.
I also considered using a state stack (like in games or iOS) where only the state on the top of the stack gets updated and rendered, which would of course allow me to push/pop different models and swap them out. I noticed in the game linked above, localStorage is used to pass data around between "pages", but that creates a global mutable state, which is undesirable. Has anyone seen any better approaches for structuring a "desktop-like" stateful application? | 2019-04-08T15:26:35.755100 | Isaiah |
elmlang | general | Inline Styled elm-css slows down Elm a lot though at the moment. If you have a lot of styled nodes, beware and benchmark before. Stylesheets are much faster. | 2019-04-08T15:27:25.755200 | Velia |
elmlang | general | I’m taking a similar approach of `type Screen = LoginScreen Login.Model | AppScreen App.Model` with a relatively small but quite complex app right now. I’m fairly excited about this approach, but if you or anyone else finds a better one, I’d love to hear about it. | 2019-04-08T15:39:12.755600 | Chae |
elmlang | general | I’m taking what I think is a distinctive approach in an app I’m working on. <https://github.com/jhbrown94/experimental-elm-spa-skeleton> | 2019-04-08T15:43:15.755800 | Dede |
elmlang | general | (this isn’t the app, it’s an extract from it, but it highlights some of the key stuff.) | 2019-04-08T15:43:52.756000 | Dede |
elmlang | general | Gotta go now, but happy to discuss if it looks interesting later on. | 2019-04-08T15:43:59.756200 | Dede |
elmlang | general | elm-css generates a stylesheet and inserts it at the top of the root element, are you talking about an old version of it? | 2019-04-08T15:54:32.756500 | Dayna |
elmlang | general | No current version. Elm-css currently performs some checks at runtime (this will be improved once/if the phantom branch is merged) and the hashing used to generate the classes (at each view) can become heavy. At last `Lazy` renders a style node at the DOM level of the lazy call, so you may end with a lot of duplicate style nodes in some cases.
This may not matter though depending of your use case.
See <https://rawgit.com/webbhuset/test-elm-performance/master/results/charts.html>
or test it yourself:
<https://rawgit.com/webbhuset/test-elm-performance/master/0.19/index.html> | 2019-04-08T16:05:45.756800 | Velia |
elmlang | general | I’ve (only) had problems with elm-css while loading an infinite list of items containing 5 elements each. Losing fps while scrolling after 200+ items | 2019-04-08T16:09:53.757400 | Lea |
elmlang | general | Yes it's usually only noticeable with hundreds or more nodes with complex styles. | 2019-04-08T16:11:21.757600 | Velia |
elmlang | general | It is definitely fine for generic use though | 2019-04-08T16:11:26.757800 | Lea |
elmlang | general | Sure, and there are some workarounds for extreme cases, but still, if 60 FPS is wanted, this might be an issue (typically animations). | 2019-04-08T16:12:20.758000 | Velia |
elmlang | general | Exactly, it’s fine for static pages, that may even have quite a bit of content. But if you want to move animate/transition stuff regular css is a lot less risky | 2019-04-08T16:13:29.758300 | Lea |
elmlang | general | can someone help me decode this union type into a list?
```
decoder : Decode.Decoder Permissions
decoder =
Decode.succeed Permissions
|> Pipeline.required "owner" (Decode.list operationDecoder)
|> Pipeline.required "group" (Decode.list operationDecoder)
|> Pipeline.required "other" (Decode.list operationDecoder)
operationDecoder : Decode.Decoder Operation
operationDecoder =
Decode.string
|> Decode.andThen
|> toOperation
toOperation : String -> Decode.Decoder Operation
toOperation operation =
case operation of
"read" -> Decode.succeed Read
"write" -> Decode.succeed Write
"execute" -> Decode.succeed Execute
_ -> Decode.fail "Unknown operation in permissions"
``` | 2019-04-08T16:43:29.759300 | Vilma |
elmlang | general | I keep getting this error:
```
-- TYPE MISMATCH -------------------------------- src/Operations.elm
This function cannot handle the argument sent through the (|>) pipe:
27| Decode.string
28| |> Decode.andThen
^^^^^^^^^^^^^^
The argument is:
Decode.Decoder String
But (|>) is piping it a function that expects:
a -> Decode.Decoder b
``` | 2019-04-08T16:44:03.759800 | Vilma |
elmlang | general | the `andThen` needs an argument, so I think `|> Decode.andThen toOperation` will do what you want | 2019-04-08T16:45:16.760400 | Virgie |
elmlang | general | I believe this could be an issue from `hecrj/html-parser` that may not perform enough checks before creating nodes attributes. I would open a bug there with your SSCCE.
If this can help until it's confirmed, it works using `elm-explorations/markdown` `toHtmlWith` (html is valid markdown), for example:
<https://ellie-app.com/5cFKS9GLsrxa1>
By the way, I'm also working on an RSS reader in my spare time, and I began going the whole rabbit hole, starting with <https://package.elm-lang.org/packages/dmy/elm-imf-date-time/latest/> then working on some xml and rss libraries. So what are you using to parse RSS feeds and do you intend to publish something? I would not like to duplicate some work if possible. | 2019-04-08T16:46:31.760500 | Velia |
elmlang | general | oh my | 2019-04-08T16:47:19.760900 | Vilma |
elmlang | general | this is just me being tired | 2019-04-08T16:47:40.761300 | Vilma |
elmlang | general | that’s so obvious lol | 2019-04-08T16:47:45.761500 | Vilma |
elmlang | general | thanks <@Virgie> | 2019-04-08T16:47:48.761700 | Vilma |
elmlang | general | I spent 20 min looking at this | 2019-04-08T16:48:00.761900 | Vilma |
elmlang | general | Maybe <@Tamra> could confirm. | 2019-04-08T16:56:48.762400 | Velia |
elmlang | general | Concerning the loop, I believe it is because the Virtual DOM diffing algorithm notices that the attribute has still not be created, and tries to create it again (which is not possible because of the invalid character), raising the error at each diffing. You most likely won't be able to fix this without fixing the attribute creation. | 2019-04-08T16:58:29.762600 | Velia |
elmlang | general | I'm thinking about how to debug `elm/parser` Parsers (read: I'm trying to write a parser and failing miserably). What about a version of the `elm/parser` package that collects "logs" about what it did during the parsing - eg. something like here: <https://rosettacode.org/wiki/Monads/Writer_monad#Haskell> ? Would that make sense? | 2019-04-08T17:08:07.765200 | Florencia |
elmlang | general | (Or, does anybody have a better idea?) | 2019-04-08T17:08:24.765500 | Florencia |
elmlang | general | In my experience, using `inContext` and writing a little thingymajob to visualise the context stack for deadends has been super helpful | 2019-04-08T17:11:55.765700 | Huong |
elmlang | general | I don't have any code handy, tho, sorry! | 2019-04-08T17:12:04.765900 | Huong |
elmlang | general | It's true I haven't used `inContext` yet at all.
The thing is, I don't have dead ends. It parses successfully, but wrongly :sweat_smile: | 2019-04-08T17:19:02.766100 | Florencia |
elmlang | general | I can try to force something like that though | 2019-04-08T17:19:17.766300 | Florencia |
elmlang | general | with `Parser.end` or something | 2019-04-08T17:19:21.766500 | Florencia |
elmlang | general | `Parser.problem` is always nice, if you know some place it reaches but shouldn't be reaching | 2019-04-08T17:23:27.766700 | Huong |
elmlang | general | <@Krista> At last I forgot to mention it, but the lack of attributes check in elm/html is known:
<https://github.com/elm/html/issues/46>
But most likely won't be fixed mainly for performance reasons. See also:
<https://github.com/elm/html/issues/47#issuecomment-229853447> | 2019-04-08T17:26:20.766900 | Velia |
elmlang | general | Yeah, it's the comparison between the type checking that I was interested in. This is from the perspective of someone who's been doing Elm for three years and is taking on a Typescript project for the first time. I've done some investigation and I've written a gist. I'll post once I've finished. As an example, I was suprised that it can do exhaustiveness checking of switch statements, but it's very clunky. | 2019-04-08T17:29:29.767100 | Rheba |
elmlang | general | ah, wow, that's a lot of great insights, thanks a lot <@Velia>! I figured also that I should rather find a way to fix it in hecrj/html-parser (to check before the node creation) | 2019-04-08T17:33:17.767300 | Krista |
elmlang | general | regarding the rss, it's "just" a elm ui (<https://github.com/magopian/rerss-elm-ui>) as a side project, for the rerss project of some friends: <https://framagit.org/ybon/rerss/tree/master/rerss/app> | 2019-04-08T17:34:15.767500 | Krista |
elmlang | general | <@Florencia> I've used this technique every now and then for debugging `elm-markup`: <https://github.com/mdgriffith/elm-markup/blob/bd14defa26dd88bf77435fe614c7cf1cbd591b8f/src/Mark/Internal/Parser.elm#L1493> | 2019-04-08T17:35:33.768400 | Cornell |
elmlang | general | Just logs the source of what it parsed | 2019-04-08T17:36:15.768800 | Cornell |
elmlang | general | I did not give it much thought, but a `Parser.Debug` could make sense, maybe using `andThen`, `getSource` and `getOffset` after each function of the API and producing some nicely formatted logs that show the parser progression in the string with each returned value. It could be slow but this does not matter much, and maybe an `Advanced` version would also be needed. Actually a Context in addition would be much better, as the parser cannot guess it.
Not sure if it should use `Debug.log` or accumulate the logs.
Anyway it seems to be a good idea, but I may have overlooked some showstoppers. | 2019-04-08T17:37:36.768900 | Velia |
elmlang | general | Parsers are indeed sometimes tricky to debug (for example if you don't know that `chompUntil*` do not consume the end string but advances the position, which leads to hard to debug errors, <https://github.com/elm/parser/issues/20)> | 2019-04-08T17:42:11.769100 | Velia |
elmlang | general | Well, before anyone else loses as much time as I have, there’s a bug in Safari that causes onFocus to trigger twice in series, before update gets a chance to do it’s work on the first. | 2019-04-08T17:43:33.771100 | Leoma |
elmlang | general | On related note to the parsers, if I change a package in `~/.elm/0.19.0/packages/...`, does Elm recheck its hash or something or does it just use it blindly? | 2019-04-08T17:43:47.771400 | Florencia |
elmlang | general | Nice, thanks for your answer. | 2019-04-08T17:44:07.771700 | Velia |
elmlang | general | The elm forum is running on discourse. My company created a community and the FE is pure Elm. I could fork it, remove the banking bits, and make it open source.
Sorry to post this to general, but does anyone know who I can talk with to gauge whether this is interesting ? | 2019-04-08T17:49:56.773900 | Charity |
elmlang | general | The nice things about discourse are
- it's free (for sizeable FOSS projects like Elm)
- it has a bunch of builtin moderation tools and a community quite willing to listen to us
- it's reasonably familiar
- it's well tested.
So basically, it's something don't have to worry about _running_, only about moderating. If an alternative with similar features were to pop up, I'm sure that'd be very interesting, but the bar for the Elm community to move is set quite high! | 2019-04-08T17:56:30.774800 | Huong |
elmlang | general | That said, me personally and a large part of the community I'm sure would be super interested in having this available as open source! | 2019-04-08T18:02:07.775000 | Huong |
elmlang | general | I fully realize that's a non-trivial investment, though :sweat_smile: | 2019-04-08T18:02:29.775200 | Huong |
elmlang | general | Any good resources on creating an autocomplete input box in Elm 0.19? Extra points if it's from an HTTP API | 2019-04-08T18:15:54.776200 | Luz |
elmlang | general | I'm having an issue with creating a file drop-zone in elm using ports. I was able to successfully create it once in its own repository (<https://github.com/SkySor44/elm-drop-zone>) but when I tried to implement it in our app at work I can't get the event.dataTransfer.files to be anything but empty. I am aware of the many instances of people trying to log the files in the console but that isn't my issue here. When I drop a file from my computer it doesn't put it in the dataTransfer.files FileList. All my code is exactly the same with just a little more occurring in surrounding elements and I can't figure out why it works in the above repository but not in my elm app. Any ideas? | 2019-04-08T18:17:28.776500 | Treasa |
elmlang | general | We run a community already in production. The main parts of the community are solid. I'm not sure what kind of moderation y'all would need and we can talk about that if it is a gap.
> it's free (for sizeable FOSS projects like Elm)
This would be free and OSS.
> it has a bunch of builtin moderation tools and a community quite willing to listen to us
There are gonna be gaps here and if we know what they are we can work them.
> it's reasonably familiar
familiar how? like UX? I guess this is personal preference, but our app is material design and installs as a PWA.
> it's well tested.
the base app has been tested and currently is live in production. The backend is a separate interface and has 1663 unit tests. | 2019-04-08T18:20:00.776600 | Charity |
elmlang | general | Again, I'm not looking to have this conversation here. It is certainly more complicated than
• "hey I wanna do this thing"
• "ok, go do this thing" | 2019-04-08T18:20:32.777200 | Charity |
elmlang | general | This package can help! <https://package.elm-lang.org/packages/ContaSystemer/elm-menu/latest/> | 2019-04-08T18:27:40.780400 | Lorilee |
elmlang | general | Hey I have a little problem in my animation frame subscription.
1) When the player hits 'go', I get a msg, and there I request some random data in a Cmd.
2) When the result of that randomness is received, I update the model to have animate=true
3) my subscription function checks for animate=true, and if so returns `onAnimationFrameDelta AniFrame`
The problem is it doesn't work unless I do something else like click a button. | 2019-04-08T18:31:44.781000 | Lindsey |
elmlang | general | Re discourse being free for the Elm community - I mean that it's a hosted instance, so it's also free as in beer. Discourse itself is also FOSS, but that's not super relevant.
Re familiarity: I mean for people coming from other communities that use discourse, such as Elixir.
I think the best course of action might be to message the Discourse moderation team directly (<https://discourse.elm-lang.org/about> -> ElmModerationTeam -> Message)
My main purpose was really just to make it clear that changing our community infrastructure needs significant reason. It being built in Elm is absolutely amazing (and I'm super stoked by this! :smile: ), but not very high on the priority list. | 2019-04-08T18:35:48.781100 | Huong |
elmlang | general | FYI, there’s a programmatic subscription bug in Elm. <https://github.com/elm/compiler/issues/1776> | 2019-04-08T18:43:54.782400 | Leoma |
elmlang | general | Ok thanks. It works if I change my 'animate' bool in the event where I receive the button message. | 2019-04-08T18:47:56.782600 | Lindsey |
elmlang | general | Personally, I didn't feel the `Response` or `Error` types were a great fit so I ended up rewriting both of them and yes, the mapping between them too | 2019-04-08T19:39:28.783100 | Kimiko |
elmlang | general | Do you have a link to this issue? | 2019-04-08T19:41:33.783300 | Kimiko |
elmlang | general | is there some easy way of getting elm 0.18 on ubuntu 18.04 other than building from source? | 2019-04-08T20:01:40.784100 | Carter |
elmlang | general | `npm install [email protected]` ? | 2019-04-08T20:09:35.784700 | Ruthann |
elmlang | general | <https://www.npmjs.com/package/elm> | 2019-04-08T20:09:43.784900 | Ruthann |
elmlang | general | I appreciate pragmatism.
Also, where is all of this free beer? | 2019-04-08T20:27:13.785300 | Charity |
elmlang | general | Thank you! I'll check it out! | 2019-04-08T20:46:08.785500 | Luz |
elmlang | general | <https://bugs.chromium.org/p/chromium/issues/detail?id=101137&q=pushstate&colspec=ID%20Pri%20Mstone%20ReleaseBlock%20Area%20Feature%20Status%20Owner%20Summary> | 2019-04-09T00:35:52.786000 | Brady |
elmlang | general | Is there any updated word on websocket library migration to 0.19? | 2019-04-09T00:39:20.786400 | Nga |
elmlang | general | much as I want it too, I hope that priority is being given to the compiler bugs | 2019-04-09T02:35:48.786600 | Lynn |
elmlang | general | We started thinking about that yesterday too - that might create opportunities to model thinks more specifically for our use case | 2019-04-09T02:42:29.786800 | Lynn |
elmlang | general | The project is some kind of social RSS reader where you share (or re-rss, like a retweet). What about your project? | 2019-04-09T02:51:06.787000 | Krista |
elmlang | general | Trying to install create-elm-app on WSL/Ubuntu, and I get this from binwrap-install: `ERR Error extracting <https://github.com/stoeffel/elmi-to-json/releases/download/0.19.1/elmi-to-json-0.19.1-linux.tar.gz> - Error: EACCES: permission denied, mkdir '/usr/lib/node_modules/create-elm-app/node_modules/elmi-to-json/unpacked_bin`
I had the same issue with installing both elm and create-elm-app on plain Ubuntu at work; eventually had to just use nvm, which is a huge pain. | 2019-04-09T05:17:23.788800 | Gertrude |
elmlang | general | <@Charity> Thank you for the gist! | 2019-04-09T05:24:09.789000 | Miguelina |
elmlang | general | you can solve this issue by owning /usr/lib/node_modules <@Gertrude> im not sure if its possible to change default global lib location of npm but a temporary fix would be just `sudo chown <YOURUSERNAME> /usr/lib/node_modules` | 2019-04-09T05:28:43.790900 | Sunni |
elmlang | general | and probably you will need also need to somehow give yourself write access to `/usr/bin` | 2019-04-09T05:29:37.791800 | Sunni |
elmlang | general | what i did was `sudo chown root:nikos /usr/bin && sudo chmod g+w /usr/bin` | 2019-04-09T05:30:27.792600 | Sunni |
elmlang | general | <https://docs.npmjs.com/resolving-eacces-permissions-errors-when-installing-packages-globally#manually-change-npms-default-directory> this is the recommended way to handle this | 2019-04-09T05:43:46.792900 | Huong |
elmlang | general | Messing with permissions outside of `/home` is generally not a great idea | 2019-04-09T05:44:23.793800 | Huong |
elmlang | general | Try not to install packages globally. I understand the urge, I love to do the same, but this leads to a lot of pain. Best is to install in your project dir (which is the default) | 2019-04-09T05:48:00.794000 | Syble |
elmlang | general | ```sudo npm install -g create-elm-app --unsafe-perm=true```
Should do the trick...
But changing the global install folder to a folder in your home directory is probably the best option.. | 2019-04-09T05:52:34.794300 | Monnie |
elmlang | general | `/usr` is ok if its your machine | 2019-04-09T05:55:54.795100 | Agustin |
elmlang | general | `/usr/local` is especially ok on osx | 2019-04-09T05:56:15.795700 | Agustin |
elmlang | general | Anything is "ok" if it's your machine. Doesn't make it a great idea. | 2019-04-09T05:56:49.795800 | Huong |
elmlang | general | <@Dede> your approach make writing spa more easily for a beginner like me :smile: , thank you very much! | 2019-04-09T06:04:36.796200 | Shantell |
elmlang | general | I put my stuff in your framework, it works! until the one use subscriptions. I can not figure out how to get it to work. I can not fully understand your framework so far. do you have any example on subscriptions? | 2019-04-09T06:11:06.796500 | Shantell |
elmlang | general | well, for something like create-elm-app, it's sort of half the use case to have it globally. ditto the elm compiler really. | 2019-04-09T06:14:41.796900 | Gertrude |
elmlang | general | real problem is the insistence by Ubuntu that npm be installed such as to require root to do it | 2019-04-09T06:15:13.797100 | Gertrude |
elmlang | general | <@Huong> here we are just changing the folder permissions nothing inside of it and we dont own the folder in the example above i just added myself to the group and let the group be able to write | 2019-04-09T06:19:27.798300 | Sunni |
elmlang | general | if it was chmod -R then yes you would run into trouble since you would mess with the permissions of sudo | 2019-04-09T06:20:43.799100 | Sunni |
elmlang | general | Yes, and now _everything_ running as your current user has permission to globally install stuff. | 2019-04-09T06:20:44.799200 | Huong |
elmlang | general | If you trust your own user unconditionally, you might as well be running as root :shrug: | 2019-04-09T06:21:19.799500 | Huong |
elmlang | general | well if theres a way of setting npm global dir that would be a nicer thing of course but this is a work around | 2019-04-09T06:21:26.799700 | Sunni |
elmlang | general | Yes, there is, and that's exactly what I linked to | 2019-04-09T06:21:41.799900 | Huong |
elmlang | general | and globally installing things are not bad if they need some root stuf still they cannot do that | 2019-04-09T06:22:14.800100 | Sunni |
elmlang | general | I don't mind globally installing things, I do mind globally installing potentially untrusted stuff, and giving potentially untrusted stuff permission to globally install more stuff :smile: | 2019-04-09T06:22:52.800300 | Huong |
elmlang | general | Oh than that is cooler than my temporary patch :slightly_smiling_face: but can you explain please why do you thing being able to install things in path to be a bad thing ? | 2019-04-09T06:23:06.800500 | Sunni |
elmlang | general | I don't. My point is more that you're giving your user permission to globally install stuff. The result of which is that any process you run as your user can now globally install stuff, without your knowledge. | 2019-04-09T06:24:01.800700 | Huong |
elmlang | general | yes but thats not a extremely dangerous thing thats what im trying to say :slightly_smiling_face: they still cannot perform untrusted operations and even if you dont allow global invocation they would be able to put things under some location and call by path | 2019-04-09T06:25:33.800900 | Sunni |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.