State is cloned in two places:
syncState: https://github.com/phoenixframework/phoenix/blob/master/assets/js/phoenix.js#L1282
syncDiff: https://github.com/phoenixframework/phoenix/blob/master/assets/js/phoenix.js#L1323

We have a presence channels with couple of thousand of users. There are users joining and leaving quite often.

Here's a 10 second performance test in chrome:
![Screen Shot 2

Well, we need to check all the projects and remove outdated ones.

We declare this in the readme:

However, keep in mind that we don't accept mammoth's shit. Only active and interesting projects with good documentation are added. Dead and abandoned projects will be removed.

But, sadly at this moment, this project looks like a mammoth shit itself.

We really need to fix it. I think

Steps to reproduce

Run asdf. In help output a line is printed for the asdf env command:

asdf env <command> [executable]      Prints or runs an executable under a command environment

I would assume <command> is an arbitrary command. But from what I can tell needs to be a plugin name or a shim name? I'm also not sure how this differs from executing the shim directly.

FY

Environment

• Credo version: 0.8.5
• Erlang/Elixir version: 1.5.0
• Operating system: Windows 10 (build 15063)

Special characters (arrows, edge) display as garble

I've noticed this error when running dialyzer over a project that depends on Absinthe:

:0:unknown_type Unknown type: Result.Object.t/0.

I think it comes from this reference to Result.Object.t() in Absinthe.Blueprint.Execution. I think it'd just be a

Hello 🙂

I am currently following the tutorial Getting started with Guardian and within the section Create Implementation Module is a broke link to the implementation module docs.
The correct one would be `https://github.com/ueberauth/guardian/blob/master/guides/introduction/

First thanks for the great project and the great documentation! 👏 💃

The docs for migrations describe a pretty good approach. However, to the best of my search ability the docs seem not to mention how to do rollbacks of migrations which to me is an integral part of any deployment that I wouldn't want to launch without.

If there is a means to do this then I apologize and would someone please point to it or document it for easier use. The erlang NIF guide shows the load funtion as:

int (*load)(ErlNifEnv* env, void** priv_data, ERL_NIF_TERM load_info)

I am unable to find a way to do that with rustler but it is something I need as my NIF does some setup when the module loads and I want to be able to access

Using v3.1, tuples do not encode:

iex(1)> Poison.encode({1, 2})
{:error, {:invalid, {1, 2}}}

I think encoding to an array would be more robust. If that is a no-go, documenting the inability to encode tuples would be an improvement.

Doing the following:

    file = "something.xml"
    HTTPoison.post(
      "https://example.com",
      {:multipart, [{:file, file, {"form-data", [name: "\"foobar\"", filename: "\"#{Path.basename(file)}\""]}, [{"Content-Type", "application/xml"}]}]},
      [{"Content-Type", "application/xml"}],
      []
    )

Is it possible not to send the content-length in the multipart req

We should either show good Elixir style in all koans or add a koan at the end that shows good elixir style.

This is just a reminder to double-check what we have.

https://github.com/niftyn8/elixir_style_guide

A tutorial that goes through steps from start to finish on setting up elixirscript in a phoenix application. Aimed at beginners.

Just helped a user on Slack through this:
you cant create rootfs overlay files for the appdata partition in Nerves. This is because it gets formatted on first boot before being mounted. I think this behavior is fine, but we need to document it.

The Overview section shows example of Article factory:

  def article_factory do
    %MyApp.Article{
      title: "Use ExMachina!",
      # associations are inserted when you call `insert`
      author: build(:user),
    }
  end

However, in Ecto Associations section there is following snippet:

def article_factory do
  %Article{
    title: "Use ExM

SendGrid api has the following properties for emails:

https://sendgrid.com/docs/API_Reference/Web_API_v3/Mail/index.html

See the disposition and content_id properties under attachments.

These turn out to be pretty important when you're sending .ics files (calendar invites). By default SendGrid populates the disposition value with "attachment" if it isempty, but Microsoft Outlook requi

Steps to reproduce

iex> t1 = "2019-04-03T21:07:45Z" |> Timex.parse!("{ISO:Extended:Z}") |> Timex.to_datetime("Europe/Berlin")
#DateTime<2019-04-03 23:07:45+02:00 CEST Europe/Berlin>
iex> Timex.day(t1)
93
iex> t2 = "2019-04-03T22:07:45Z" |> Timex.parse!("{ISO:Extended:Z}") |> Timex.to_datetime("Europe/Berlin")
#DateTime<2019-04-04 00:07:45+02:00 CEST Europe/Berlin>
iex> Timex.

Hello,

Could you add the pop-up documentation of a function/variable when we press Ctrl + Q? For now it shows No documentation found. even if I used @doc on the function.

Thanks

It looks like the new parse_document's output is not the same as the deprecated parse. By using the example provided in the documentation:

html = """
  <html>
  <body>
    <div class="example"></div>
  </body>
  </html>
"""
Floki.parse_document(html)
{:ok,
 [
   {"html", [],
    [
      {"head", [], []},
      {"body", [],
       [
         "\n",
         "    ",
     

For checkbox inputs of associations with unique pluralization requirements, e.g. specialty/specialties, ExAdmin incorrectly pluralizes the <association>_ids param key: specialtys.

In https://github.com/smpallen99/ex_admin/blob/master/lib/ex_admin/param_associations.ex#L34, if you replace

new_key =
  String.replace_suffix(key_as_string, "_ids", "s")
  |> String.to_atom()

I am getting random {:error, :closed} from :hackeny.send_body/2 and Tesla.client crashes with following stack trace.

 {:EXIT, #PID<0.23655.28>, {{:badmatch, {:error, :closed}}, [{
Tesla.Adapter.Hackney, :"-request_stream/5-fun-0-", 3, [file: 'lib/tesla/adapter/hackney.ex', line: 74]}, 
{Enum, :"-reduce/3-anonymous-2-", 3, [file: 'lib/enum.ex', line: 3015]}, 
{Stream, :do_transform_e

FYI
Readme says:

def deps do
  [{:coherence, "~> 0.6"}]
end

However, in https://hex.pm/packages/coherence it only has version 0.5.2. Users get this error:

➜  path-to-project git:(master) ✗ mix deps.get
** (Mix) No matching version for coherence ~> 0.6 (from: mix.exs) in registry

The latest version is: 0.5.2

See screenshot. Thank you!

![screen s

Precheck

I didn't find any existing issues (open or closed) that cover this.

I think my Dialyzer and Erlex are up-to-date but I could be wrong – please let me know.

Environment

$ elixir --version
Erlang/OTP 21 [erts-10.2.2] [source] [64-bit] [smp:4:4] [ds:4:4:10] [async-threads:1] [hipe]

Elixir 1.8.0 (compiled with Erlang/OTP 20)
$ cat mix.lock | grep dialyxir
  "dia

Turns out this is a whitelist that defaults to []. We just got bit pretty hard by this behavior being hidden, so it's worth calling out in the docs. Effectively we had a bunch of plugs all acting on the presence of query params and were very confused why they never seemed to actually trip. This whitelist behavior was very unexpected.

Seems to have been added quite a while back in https://git