Feature/add formatter

[Kociamber]

I needed to be able to use "external" formatter for the message and metadata before sending them to Greylog. I made a short mod in order to be able to use formatter "callback" and I thought other ppl might find it useful too.

[jschniper]

I was going to try to merge this with the other outstanding pull request for supporting an option for an inline format. Is there anyway you could send me a sample formatter so I can build out a test case to make sure everything works as expected.

[Kociamber]

Sure, I'm pasting short code below. Would you like me to add tests to my PR as well? I can do it quickly. Cheers!

defmodule LogFormatter do
  @moduledoc """
  Provides a set of functions for logged message transformation.
  """

  @doc """
  Main function of the formatter.
  """
  @spec format(atom(), list(), tuple(), list()) :: {atom(), list(),
tuple(), list()}
  def format(level, message, timestamp, metadata) do
    metadata = add_us_precision_timestamp_to_metadata(metadata)
    {level, message, timestamp, metadata}
  end

  # helper functions
  defp add_us_precision_timestamp_to_metadata(metadata) do
    Keyword.merge(metadata, timestamp_us: :os.system_time(:micro_seconds))
  end
end

EDIT

Was carried away and added the test ;)

[jschniper]

Quick question: what was the reason for including the arity in the configuration? From the Logger docs it looks as though it's always called with four arguments

[jschniper]

Fundamentally the code is the same but I reworked a few pieces and removed the arity part of the tuple since it didn't seem necessary. I haven't pushed a new build to hex but you should be able to run the code in master right now with little change. Let me know if you run into any issues.

[Kociamber]

Hi! Thanks :) Indeed, arity wasn't in the end necessary at all so that's fine.

czw., 2 maj 2019 o 19:59 Joshua Schniper notifications@github.com napisał(a):

Fundamentally the code is the same but I reworked a few pieces and removed the arity part of the tuple since it didn't seem necessary. I haven't pushed a new build to hex but you should be able to run the code in master right now with little change. Let me know if you run into any issues.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/jschniper/gelf_logger/pull/22#issuecomment-488771031, or mute the thread https://github.com/notifications/unsubscribe-auth/ACTENIB6E43AERYSY3HXJVTPTMTXJANCNFSM4HJKUMBQ .

[Kociamber]

Hi, would you be able to push it to hex today? Cheers, Rafał

czw., 2 maj 2019 o 20:37 Rafał Kociszewski rafal.kociszewski@gmail.com napisał(a):

Hi! Thanks :) Indeed, arity wasn't in the end necessary at all so that's fine.

czw., 2 maj 2019 o 19:59 Joshua Schniper notifications@github.com napisał(a):

Fundamentally the code is the same but I reworked a few pieces and removed the arity part of the tuple since it didn't seem necessary. I haven't pushed a new build to hex but you should be able to run the code in master right now with little change. Let me know if you run into any issues.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/jschniper/gelf_logger/pull/22#issuecomment-488771031, or mute the thread https://github.com/notifications/unsubscribe-auth/ACTENIB6E43AERYSY3HXJVTPTMTXJANCNFSM4HJKUMBQ .

[jschniper]

Just pushed it. Let me know if you run into any problems.