Add CommandRunner utility to @actions/exec

[nikolai-laevskii]

Provides more feature-rich alternative for getExecOutput in order to to provide shortcuts for the most common use-cases.

• encourages writing clean and readable code by allowing to chain different effects related to command executions
• provides shorthands for the most common use-cases such as: failing action on command fail, throwing error on command fail, producing a specific log when command failed to execute
• provides instruments to distinguish between different types of errors such as error during command execution (exitcode !== 0) and error to execute command itself
• extended functionality allows to set different behavior for any amount of arbitrary types of errors or outputs (specific exitcode, output matching certain pattern, empty output, etc.)
• can be tailored to specific needs on an individual project level or easily adjusted on a library level without breaking the backwards compatability
• command arguments or options can be overwritten for a specific run which allows to create easily readable and extensible functions which correspond to different commands

const {stdout} = await exec.createCommandRunner(arbitraryCommand, [], {cwd})
  .onExitCode('> 0', 'throw') // default message will be autogenerated 
  .onEmptyOutput('throw')
  .on(['stderr', 'execerr'], 'throw') // can separately cover all other errors
  .run()

or even

const {stdout} = await exec.createCommandRunner(arbitraryCommand, [], {cwd})
  .on('!ok', 'throw') // will automatically figure out default message
  .run()