View Source Mash (mash v0.3.1)
Mash is a dead simple tool for defining and running CI pipelines in Elixir.
Dependencies
- Elixir
- Mix
- The unix
script
utility
Installation
Add mash to your dependencies like so:
def deps do
[
{:mash, "~> 0.1.2"}
]
end
Configuring pipelines
To configure your job pipeline, create a .mash.exs
file at the root of your project containing a module which implements a jobs/0
function:
defmodule MashConfig do
import Mash.Helpers
def jobs do
[
%{
name: :test,
run: mix("test")
},
%{
name: :credo,
run: mix("credo", ["--all"])
},
%{
name: :function_example,
run: fn _io_pid ->
IO.puts("Hurray!")
end
}
]
end
end
Each entry in the jobs list must, at a minimum, have two keys:
:name
-- the name of the job which must be a unique atom:run
-- a function which defines the work for the job to do. We recommend using the helper functions from Mash.Helpers to define your run function, as they ensure that your function handles setup and teardown correctly for many common scenarios, e.g.:Mash.Helpers.mix(task_name, args)
-- execute a mix task with the giventask_name
andargs
.Mash.Helpers.shell(command, args)
-- run an executable/script namedcommand
withargs
.- Or you can pass a function with arity 1.
- The function should return
:ok
or{:ok, value}
if the job succeeds, or{:error, error}
if the job fails. - The argument passed to your function will be the PID of the IO device which is capturing IO. Since we make this PID the "group leader" for your job automatically, you can probably ignore this argument, but you may need to for interfacing with certain libraries that require you to pass an io device PID directly.
- The function should return
Optionally, a job may also contain a key needs
whose value is an array of other jobs that must pass before this job will be run.
See the docs for Mash.Job for more details, including conditional execution.
Running
mix mash
Full Documentation
Full documentation is available on HexDocs.