workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
elmlang | general | Otherwise the dead code elimination will remove it. | 2019-03-12T11:21:01.857600 | Jin |
elmlang | general | The whole App? | 2019-03-12T11:21:14.857900 | Corinne |
elmlang | general | The port, so it won’t appear in `app.ports`. | 2019-03-12T11:22:02.858500 | Jin |
elmlang | general | but `app.ports` is `undefined` | 2019-03-12T11:22:13.858800 | Corinne |
elmlang | general | Unless.. .that's expected? | 2019-03-12T11:22:31.859400 | Corinne |
elmlang | general | you gotta wire them up on the Elm side | 2019-03-12T11:23:43.859700 | Rosa |
elmlang | general | Ah, right so a completely empty object for `app` is normal | 2019-03-12T11:25:11.861000 | Corinne |
elmlang | general | If you never use any incoming port in your `subscriptions` function used in `main` nor any outgoing port in `init` or `update`, then `app.ports` will be undefined because the elm compiler will have optimized those away (as they are never called) | 2019-03-12T11:25:55.861900 | Mindy |
elmlang | general | Cool, thanks. I guess I was expecting _something_ to be in that object, or maybe a `[]` for `ports`, but I see now | 2019-03-12T11:26:58.862700 | Corinne |
elmlang | general | Basically, `app` is just an object that may or may not have a `ports` field and `app.ports` is an object with one field per port called in the elm side.
If it is an outgoing port, it will have access to a `subscribe` and an `unsubscribe` method. If it is an incoming port it will have a `send` method. | 2019-03-12T11:30:55.865500 | Mindy |
elmlang | general | using `/` as the italic symbol in combination with `/` as division and `//` in url's isn't possible normally. Why is `/` the symbol for italic (thought it was a single `*`, or `_`) | 2019-03-12T11:39:27.865900 | Virgie |
elmlang | general | None | 2019-03-12T11:42:42.866100 | Leoma |
elmlang | general | <@Virgie> <https://bear.app> parses this way :point_up:. It’s more intuitive for people who don’t understand markdown. | 2019-03-12T11:42:47.866500 | Leoma |
elmlang | general | In my implementation, modifiers (/*) only count if the modifiers have whitespace on the left-but-not-right to open, and right-but-not-left to close. | 2019-03-12T11:43:52.867000 | Dede |
elmlang | general | It covers most cases. | 2019-03-12T11:43:55.867200 | Dede |
elmlang | general | `(5 / 4) ` also parses fine with that. | 2019-03-12T11:44:09.867400 | Dede |
elmlang | general | My math is a bit off. :blush: | 2019-03-12T11:44:27.867700 | Leoma |
elmlang | general | my guess is then that the url is parsed differently | 2019-03-12T11:44:33.867900 | Virgie |
elmlang | general | as in, a url is not text, it is its own thing | 2019-03-12T11:44:45.868200 | Virgie |
elmlang | general | Yes. | 2019-03-12T11:44:58.868400 | Leoma |
elmlang | general | But it can also be styled. | 2019-03-12T11:45:44.868600 | Leoma |
elmlang | general | None | 2019-03-12T11:45:50.868800 | Leoma |
elmlang | general | does something like `ftp://` also parse as a url? or just `http[s]`? | 2019-03-12T11:46:00.869200 | Virgie |
elmlang | general | I mean, once one has a the code for one, the other is easy to add. My target is very simple markdown I can explain to non-savvy users is a single example. `"*this will be bold* /this will be italic/" <http://clickme.com>"` I think most people won’t know or care about ftp. | 2019-03-12T11:48:10.870100 | Leoma |
elmlang | general | Anyway, I have some pseudo code I like. Off to the battlefield to see where this goes. Back in a bit and thanks to everyone for their help so far. | 2019-03-12T11:50:51.872600 | Leoma |
elmlang | general | is there any way to "namespace" a bunch of type constructors? I want to import a type `S` but have the type constructors accessed on `S`, e.g. `S.Value`. Maybe there's a clever trick that I don't know...
```
module Values exposing (S(..))
type S = A | B | C
...
module Main
import Values exposing (S)
defaultModel =
S.A
``` | 2019-03-12T11:51:33.873300 | Kymberly |
elmlang | general | To my knowledge constructors are on the same level of visibility of the type. | 2019-03-12T11:53:37.874000 | Hoa |
elmlang | general | If you constructors have generic names consider to add a suffix or prefix to them. E.g. `type BlendMode = MultiplyMode | OverlayMode | ... ` instead of generic `Multiply` or `Overlay`. | 2019-03-12T11:55:20.875600 | Hoa |
elmlang | general | In your example I would write `Values.A` | 2019-03-12T11:56:20.876200 | Hoa |
elmlang | general | Constructors are namespaced by the name of their module, not their type :+1: | 2019-03-12T11:58:54.876900 | Carman |
elmlang | general | Any package for AES Symmetric Encryption? | 2019-03-12T12:07:20.878700 | Brady |
elmlang | general | I want to exchange keys with node.js and send encrypted messages between server and Elm app | 2019-03-12T12:08:23.879600 | Brady |
elmlang | general | I can use ports, but I'd like to do that in Elm if possible | 2019-03-12T12:08:49.880200 | Brady |
elmlang | general | Is there any tool that checks if dependencies listed in `elm.json` are actually used? | 2019-03-12T12:11:21.881400 | Dorsey |
elmlang | general | <@Dorsey> <https://github.com/stil4m/elm-analyse> | 2019-03-12T12:11:51.882000 | Brady |
elmlang | general | Elm parser question: how would one implement a grammar fragment of this form?
```
textSquence = textNoWhitespace | textSequence whitespace textNoWhitespace
``` | 2019-03-12T12:12:21.882800 | Dede |
elmlang | general | (idiomatically?) | 2019-03-12T12:12:33.883200 | Dede |
elmlang | general | <@Brady> Other ppl are using this: <https://package.elm-lang.org/packages/billstclair/elm-crypto-aes/latest/> | 2019-03-12T12:12:38.883400 | Hoa |
elmlang | general | Not expert, tho. Might well be it has limits :slightly_smiling_face: | 2019-03-12T12:13:06.884000 | Hoa |
elmlang | general | <@Hoa> I was looking at that but cannot figure how to import shared key, but thanks! | 2019-03-12T12:13:40.884600 | Brady |
elmlang | general | Thanks <@Brady> (and <@Shea> of course) | 2019-03-12T12:29:42.884700 | Dorsey |
elmlang | general | Test case question: `/* hello*/` Does this get italics or not? :wink: | 2019-03-12T12:30:46.884900 | Dede |
elmlang | general | So, I’m not super fussy about edge cases, but I suspect the simplest parser that mostly works will parse that as `Italic "* hello*"` (or however the type is represented) | 2019-03-12T12:35:48.885200 | Leoma |
elmlang | general | Broader question: Elm Parser mostly advertises itself as a replacement for regexps, rather than, say, for BNF-driven parser generators (e.g. yacc/bison.) Do people use it in the latter manner as well, or mostly hand-write parsing once the Parser has handled tokenization? | 2019-03-12T12:58:37.886700 | Dede |
elmlang | general | Update: I believe I’ve found a fast manual way to to this without Parser or Regexp. Will share if/when it works. (Very Fermat-esque statement.) | 2019-03-12T14:20:42.886800 | Leoma |
elmlang | general | parser combinators are usually used to parse straight into an AST, so no separate lexer/parser, thing. So yeah, elm/parser type parsers are definitely used as a replacement for BNF-driven parser generators | 2019-03-12T14:52:37.887100 | Huong |
elmlang | general | Thanks. Formally does that wind up being in the LL top-down family of parsers, then? | 2019-03-12T14:57:28.887400 | Dede |
elmlang | general | Yeah, pretty much | 2019-03-12T15:05:26.887600 | Huong |
elmlang | general | As a nontrivial example of something that's parsed using parser combinators: our dearly beloved `elm` compiler parses Elm using parser combinators :slightly_smiling_face: | 2019-03-12T15:06:39.887800 | Huong |
elmlang | general | :wink: | 2019-03-12T16:02:38.888100 | Dede |
elmlang | general | In Haskell or elm? | 2019-03-12T16:48:12.888500 | Dede |
elmlang | general | Haskell, though Elm can be parsed using `elm/parser`, in Elm, too - `stil4m/elm-syntax` does just that | 2019-03-12T16:49:33.888700 | Huong |
elmlang | general | I'm trying to get the error message out of `Json.Decode.fromValue`, and the docs _say_ it's a `Result String a` but when I `Debug.log` it it's `Err (Failure "Expecting an OBJECT with a field named `accessTokens`" <internals>)` | 2019-03-12T16:51:47.890000 | Kymberly |
elmlang | general | are the docs misleading or am I doing something wrong there? | 2019-03-12T16:52:13.890400 | Kymberly |
elmlang | general | ```
-- decodeValue : Decoder a -> Value -> Result String a
decodedJson =
userJson
|> D.decodeValue
(D.map3 User
(D.field "name" D.string)
(D.field "id" D.string)
(D.field "username" D.string)
)
updatedModel =
case Debug.log "decodedJson" decodedJson of
Ok value -> -- great!
Err error -> -- error should be a string, right?
``` | 2019-03-12T16:54:02.891600 | Kymberly |
elmlang | general | if I add `D.Failure` to the pattern, I get some clues...
```
Missing possibilities include:
Err (Field _ _)
Err (Index _ _)
Err (OneOf _)
``` | 2019-03-12T16:56:04.892100 | Kymberly |
elmlang | general | <@Kymberly> <https://package.elm-lang.org/packages/elm/json/latest/Json-Decode#Error> | 2019-03-12T16:58:05.892500 | Earlean |
elmlang | general | You must be looking at the wrong package docs | 2019-03-12T16:58:21.893100 | Earlean |
elmlang | general | dang I must've checked that I was on latest a dozen times - but prob on another package page. | 2019-03-12T17:02:27.893500 | Kymberly |
elmlang | general | thx! | 2019-03-12T17:02:29.893700 | Kymberly |
elmlang | general | Huh, I was looking at <https://package.elm-lang.org/packages/elm-lang/core/latest/Json-Decode> | 2019-03-12T17:03:03.894200 | Kymberly |
elmlang | general | Yep, that's docs for the core package for Elm 0.18 | 2019-03-12T17:05:17.895200 | Earlean |
elmlang | general | oh yikes - I'll have to keep an eye for "packages/elm-lang" - it shows up in a google search. :confused: | 2019-03-12T17:08:00.896100 | Kymberly |
elmlang | general | If you search via <https://packages.elm-lang.org> you'll only get 0.19 packages | 2019-03-12T17:14:34.897200 | Earlean |
elmlang | general | Older packages will come up in Google searches | 2019-03-12T17:15:02.897800 | Earlean |
elmlang | general | <@Kymberly> <https://github.com/elm/package.elm-lang.org/issues/275> | 2019-03-12T17:31:54.898100 | Millie |
elmlang | general | Thanks. | 2019-03-12T17:32:09.898200 | Dede |
elmlang | general | About to have dinner, but I’ll leave you with this <https://gist.github.com/z5h/493423f87f77bbdcf886159b21a3dc87>
Should I feel bad I’m now using regexs? I think I’m using them responsibly. | 2019-03-12T17:33:20.898400 | Leoma |
elmlang | general | Does it work? Is it maintainable? If so, it’s good enough for prod :wink: | 2019-03-12T17:50:47.898700 | Dede |
elmlang | general | Eventually yes and yes. But I’m mildly um bothered by the fact the elm/parser didn’t make it easier. | 2019-03-12T18:39:19.900200 | Leoma |
elmlang | general | I don't think that such a markdown can have very simple rules and be very practical at the same time.
For example have a look at the commonMark specification that tries to give a synthesis of the reasonable rules:
<https://spec.commonmark.org/0.28/#emphasis-and-strong-emphasis>
Not that simple...
I would not go the regexs route for performance reasons (you should test what you did on the biggest/more complex document supported by the software) and because you will have tons of edge cases that you will fix, progressively making the regexs more and more complicated, and their maintenance impossible without breaking something.
Anyway, just to show that something is possible with elm/parser:
<https://ellie-app.com/4XGvnPg32fYa1>
The rules are very strict, and therefore most likely not very practical:
• Delimiters are '/', '*' and '_' for italic, bold and underline (the '/' choice is quite risky in my opinion)
• You can nest them but if one is not closed inside another, it will be cancelled
That's it. I did not implement any space before/after things, line breaks or url handling.
You could add an url parser before every others to handle them without conflicting with the styles.
However I did not really understand what you wanted about accepting some divisions like "5/2" without conflicting.
The code is a little obfuscated by the usage of `loop` to avoid stack issues (so it uses tail-call elimination),.
Anyway this is just for demo purposes, and there may be bugs. The performance should be reasonable though. | 2019-03-12T18:55:50.900400 | Velia |
elmlang | general | I just pushed a big overhaul to <https://github.com/jhbrown94/elmish-minimarkdown> It’s still not a pure Elm Parser implementation, but it’s a lot cleaner than the first pass.s | 2019-03-12T19:23:06.900700 | Dede |
elmlang | general | (As before, right now it just supports italic and bold.) | 2019-03-12T19:23:57.900900 | Dede |
elmlang | general | But it handles all the goofy cases reasonably well IMHO. | 2019-03-12T19:24:10.901100 | Dede |
elmlang | general | Think this is a readable function?
```
is18 : Int -> String -> Bool
is18 currentTime dobStr =
let
( month, day, year ) =
dobStr |> String.split "/" |> listToTuple3
in
case Iso8601.toTime (year ++ "-" ++ month ++ "-" ++ day) of
Ok value ->
-- 568025136000 is 18 years in miliseconds
568025136000 < (currentTime - Time.posixToMillis value)
_ ->
False
listToTuple3 : List String -> ( String, String, String )
listToTuple3 list =
case list of
[ a, b, c ] ->
( a, b, c )
_ ->
( "", "", "" )
```
where `dobStr` is a `mm/dd/yyyy` string
I suppose I could ensure that that string is formatted as a iso8601 string first, but wanna start here… | 2019-03-12T19:27:35.902200 | Jeanene |
elmlang | general | I might do something like
```
Iso8601.toTime (year ++ "-" ++ month ++ "-" ++ day)
|> Result.map (\value -> 568025136000 < (currentTime - Time.posixToMillis value))
|> Result.withDefault False
``` | 2019-03-12T19:31:18.902700 | Hoyt |
elmlang | general | Although I would probably pull that lambda into an actually function like `greaterThan18` to make it more clear what that is comparing | 2019-03-12T19:33:35.903900 | Hoyt |
elmlang | general | Or at least something like
```
millisecondsIn18years = 568025136000
``` | 2019-03-12T19:34:04.904500 | Hoyt |
elmlang | general | I’ll try this out tomorrow and report back! | 2019-03-12T19:35:18.904900 | Jeanene |
elmlang | general | <@Jeanene> that milliseconds seems wrong. For example 568,015,222,000 < 568,025,136,000, but it's exactly 18 years ago today | 2019-03-12T20:06:51.907000 | Chae |
elmlang | general | today != now…I think I got the ms from `18.years.ago.to_i` in ruby :smile: :sweat_smile: | 2019-03-12T20:08:17.907700 | Jeanene |
elmlang | general | That's a variable number due to leap years | 2019-03-12T20:09:58.908400 | Chae |
elmlang | general | None | 2019-03-13T00:55:41.908900 | Leoma |
elmlang | general | The regex parser lives: <https://gist.github.com/z5h/493423f87f77bbdcf886159b21a3dc87>
Parsing and rendering in 216 lines, even after a painful lesson <https://github.com/elm/regex/issues/10> | 2019-03-13T00:56:04.909300 | Leoma |
elmlang | general | Is there an easy way to get a zero value'd record without making it myself? i.e.: ```type alias Thing = { x : Int, y : Maybe String }```, then ```getZeroValue Thing``` returns ```{x=0, y=Nothing}```. I strongly feel like the answer is no, but I'm praying otherwise :stuck_out_tongue: | 2019-03-13T03:08:32.911400 | Nga |
elmlang | general | 503 God's service not available at this time | 2019-03-13T03:10:01.911900 | Lynne |
elmlang | general | The answer is no, sorry :slightly_smiling_face: | 2019-03-13T03:10:36.912200 | Lynne |
elmlang | general | awwww, heh. thanks | 2019-03-13T03:11:14.912400 | Nga |
elmlang | general | <@Nga> a 'zero value' isn't meaningful for a lot of types | 2019-03-13T03:17:34.913200 | Earlean |
elmlang | general | What would you use this behaviour for? | 2019-03-13T03:18:27.914200 | Earlean |
elmlang | general | <@Earlean> I have a bunch of record types that define API responses. When I don't have an API response, I have it as `Maybe ApiRespObj`, but once I've received the object I transition my state from a `Maybe ApiRespObj` to just a `ApiRespObj`, so that it's easier to work with in my views. The conversion from the Maybe to the not-maybe is where I'm having a headache, because when I'm casing on my `MaybeApiRespObj` to determine whether I can convert it, if it's Nothing, I need to fall back to some zero-initd default value. It's programmed in a way that the case is never hit, but it still needs to be there because Elm needs to cover all cases. | 2019-03-13T03:21:03.916500 | Nga |
elmlang | general | Ah, that is a common mistake. Don't convert a Maybe to a nonsense default value. Just use `Maybe.map` etc. to work with the Maybe | 2019-03-13T03:24:32.918700 | Earlean |
elmlang | general | Right. it's a bit of a shame because I've modeled the data in such a way that it can only ever have a correct value at that point (we have a union type that declares the state of the app, whether its Loading, Loaded, an error, etc). By using Maybe's as my Loaded values it doesn't make a lot of sense since that's not actually something that can ever happen. Not sure where this indicates the real problem is with the approach we've taken. | 2019-03-13T03:35:54.920500 | Nga |
elmlang | general | Or if it's just a limitation problem that we have to accept. | 2019-03-13T03:36:33.920800 | Nga |
elmlang | general | A `Maybe a` that can't be Nothing should just be an `a`, right?
I'm just guessing of course, but I'd look higher up if the Maybe can be avoided before giving it to the function that doesn't expect any Nothings. | 2019-03-13T03:58:52.921100 | Sharon |
elmlang | general | <@Sharon> We have something like: `{settings = Maybe ApiSettingsResponse, profile = Maybe ApiProfileResponse}`. Can't set them to just `a` because then it's the same problem again (zero initialization before the API request returns). | 2019-03-13T04:01:40.922400 | Nga |
elmlang | general | Settings and Profile seems like something which you get once at the very start. If that is the case it may be a good idea to postpone initialization of page/view which depends on these and only init it when you receive them. In this case you won't need `Maybe`s. | 2019-03-13T04:06:02.923600 | Lynne |
elmlang | general | <@Nga> so if they can't be Nothing when you're `Loaded` then that's the time to unwrap them, if any of them are Nothing then you're still in the `Loading` state | 2019-03-13T04:18:26.926200 | Earlean |
elmlang | general | I succeeded in uploading a file together with other data, I thought this might be interesting. <https://discourse.elm-lang.org/t/how-i-uploaded-a-file/3309?u=hermanverschooten> | 2019-03-13T05:10:29.927500 | Salvador |
elmlang | general | I finally upgraded to 0,19. I tried to remove npx from my building script `npx elm make Apps/AccountDetails.elm --output ../app/assets/javascripts/passare_elm.js`, but I get an error about elm-package.json not existing (I had removed it). What do I need to do for the upgrade to be permanent? I had understood that npx was for a temporary change | 2019-03-13T09:19:56.930400 | Donnetta |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.