I find exposing functions/types at the top gives a nice summary of what is
exposed when reading a new module. I can usually get a good idea of how
the module works by seeing the names of the things that are exposed next to
each other. That would be harder to do if the exposed annotation is
This is a very educational discussion! Thank you for sharing links and
examples.
As perhaps a minor point, I believe the elm-sortable-table does NOT belong
in the list of "encapsulated state", "components", "triplets".
It does have a State type defined. But a notable difference with various
Austin, you're exactly correct. Cmd is meant to be passed out to the runtime,
that's it.
The reason Http allows you to skip Task is for learnability. You can make a
simple request without learning what a Task is. But we shouldn't be hobbled by
that when we want to compose different types of
Hey Eric, please elaborate on this one. How do you envision leaf component
triggering global state update?
> The second problem becomes: *how do we load this global state/business
> data into the application and cache it?* This requirement is now fairly
> easy to wire-up in elm since the
The low-level API just became a lot less attractive when I realized that
the listener API for sockets is basically only constructed to work with
effects managers.
Mark
On Wed, Nov 30, 2016 at 6:45 AM Mark Hamburg wrote:
> We've been building an app that communicates with
A bigger issue going forward is that with effects managers, commands aren't
just shims around tasks. So, the notion that a module should expose both a
command-based interface and a task-based interface doesn't hold up well in
the long term. That leads me back to the notion that if commands are the
It maps to the native XMLHttpRequest.withCredentials. It applies only to
cross-site requests.
https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest/withCredentials
On Wednesday, 30 November 2016 11:56:34 UTC-3, Rupert Smith wrote:
>
> On Tuesday, November 29, 2016 at 4:44:01 PM UTC,
>
> Thank you for this example. But I don't see how to properly trigger my
> "ratings" storage reload action if user state changed.
However you please! The caller has the necessary User info now, so UserInfo
is out of the picture and you can do whatever would normally make sense.
Here's one
`withCredentials` is a normal Javascript AJAX argument, it means to pass
credential information to the remote server if you are connecting to a
remote server; without it then credentials are not passed, which is better
for security but without it you may not be able to access things without
This is something that is probably solved by having better IDE/editor tools
to manage imports/exports, rather than changing the language.
The import syntax has gone through many iterations, so I wouldn't be
surprised if it's pretty stable where it is.
On Wed, Nov 30, 2016 at 8:33 AM, Matthew
There's been some ongoing discussion about components (web components,
simplifying design with component re-use, etc.) and I saw this introduction
to Svelte today:
https://svelte.technology/blog/frameworks-without-the-framework/
It seems like a compiler that produces bare-metal components by
The module's interface is an entity unto itself. When I am visiting
unfamiliar code, I often want to know what its API looks like. Having
everything there at the top is way easier than having to scan through
search results for Ctrl-F "exposed".
It also helps with design. Seeing the entire
Maybe this would be helpfull:
https://www.npmjs.com/package/elm-version-manager
On Tuesday, November 29, 2016 at 3:54:06 PM UTC+2, Rupert Smith wrote:
>
> I've been trying to follow these instructions:
>
> http://www.lambdacat.com/how-to-setup-local-development-for-elm/
>
> but getting some
This is what it comes down to for me:
Richard F talking about Maybes: https://youtu.be/IcgmSRJHu_8?t=9m21s
He says "those will go from a series of Nothings to a series of Justs".
Thing a | Nothing would mean he'd say "those will go from a series of
Nothings to a series of Things" which makes
To expand on what simon said:
If you're using multiple Elm versions in production, use the version
installed via a local npm and make all your tooling refer to
node_modules/.bin/elm-make and friends. That way you can have a global
install of elm at a version, but still use 0.16 for a particular
>
> Put another way, *Elm does not have anything similar to components!*
>
> But how about elm-mdl, elm-autocomplete, elm-ui, elm-datepicker,
> elm-form, elm-sortable-table even spinner elm-spinner? Thay all use theirs
> model, update, view "triplet" to be reusable and hide implementation
>
On Monday, November 28, 2016 at 2:38:13 PM UTC, Rupert Smith wrote:
>
> On Monday, November 28, 2016 at 2:01:34 PM UTC, sergii.k...@zalando.de
> wrote:
>>
>> I feel it will be too messy to keep all this inter-triplet logic here. Is
>> there a better way? Maybe something similar to pub/sub or
@Richard
Thank you for this example. But I don't see how to properly trigger my
"ratings" storage reload action if user state changed.
Another question, What do you think about pages separation?
For example, I have several independent tabs and one of the tabs called
"Rating" and It has some
On Wednesday, November 30, 2016 at 2:41:21 AM UTC, sergii.k...@zalando.de
wrote:
>
> Even more, it is exactly like described here
> https://gist.github.com/evancz/2b2ba366cae1887fe621 by the author of Elm
>
This is a gem. Not seen it before but there are some really useful ideas in
there.
The best parts of it seem to be its compiler (performance and dead branch
trimming of compiled code). Otherwise, it looks like a similar machination
to Angular 1. (HTML-based templates, mustaches, two-way binding,
directives).
Component-orientedness = object-orientedness. The same refactoring
This does sound like a useful addition. Unlike most of the functions that
people suggest adding to core, I don't think Cmd.andThen can be written by
an end user (or library author).
On Wed, Nov 30, 2016 at 7:19 AM, Mark Hamburg wrote:
> A bigger issue going forward is
“those will go from a series of No’s to a series of Yes’s” would also make
sense.
> On 30 Nov 2016, at 21:42, Will White wrote:
>
> This is what it comes down to for me:
>
> Richard F talking about Maybes: https://youtu.be/IcgmSRJHu_8?t=9m21s
>
> He says "those will
Yeah. And considering that the high-level web socket API is held up as the
standard of where things are going, it would seem the absence of exposed
chainable tasks is likely only to get worse.
Another way to look at this as I think about it is: Why Cmd msg as opposed
to Task Never msg as a result
We've been building an app that communicates with the server via web
sockets. The server associates account information with the socket so we
don't have to send it repeatedly. We've recently been thinking about how to
handle login to multiple accounts. The proposal that floated up was to
simply
On Tuesday, November 29, 2016 at 4:44:01 PM UTC, Rupert Smith wrote:
>
> This write up here is proving very helpful in understanding the new
> elm-lang/http. Has before and after examples showing how code is changed
> between 0.17 and 0.18:
>
>
As a relative newcomer to Elm, the impression I have is that Cmd is
"special" and really part of the underlying runtime world. Maybe this is a
misconception, but the impenetrability of Cmd gives it that sense. At the
same time, Task is accessible, well supported, and put forward as the
proper way
Not only me asking this questions.
https://groups.google.com/forum/#!msg/elm-discuss/u3J2eSThkjw/4FT3p50uBAAJ
and http://blog.jenkster.com/2016/04/how-i-structure-elm-apps.html
On of the interesting solutions https://github.com/folkertdev/outmessage
another
27 matches
Mail list logo