procworker

Documentation

Overview

Package procworker implements a worker that reads a subprocess specification task, executes it, and puts results into an outbox.

Basically, it runs what you ask it to run, and pushes results where you want them.

Because of this, YOU SHOULD NEVER USE THIS. Really. Just don't. It's super dangerous - anyone that has access to push tasks into your queue can make you run arbitrary things as your process user. That's horrible and bad and scary, even in a controlled environment.

Containers do not make this better, at least not better enough.

Index

• Constants
• func Run(ctx context.Context, t *entroq.Task) ([]entroq.ModifyArg, error)
• type SubprocessInput
  • func (p *SubprocessInput) AsOutput() *SubprocessOutput
  • func (p *SubprocessInput) JSON() []byte
• type SubprocessOutput
  • func (p *SubprocessOutput) JSON() []byte
• type TeeWriter
  • func NewTeeWriter(writers ...io.Writer) *TeeWriter
  • func (w *TeeWriter) Write(data []byte) (int, error)

Constants

const MaxStdOutSize = 1024 * 1024

Functions

func Run

func Run(ctx context.Context, t *entroq.Task) ([]entroq.ModifyArg, error)

Run is a function that can be passed to a worker's Run method. The worker will call this for each task it wishes to process. It pulls subprocess call information from its input task, including which outbox to write results to. If no outbox is specified, the input task's queue name is suffixed with "/done" to produce one.

Types

type SubprocessInput

type SubprocessInput struct {
	Cmd    []string `json:"cmd"`
	Dir    string   `json:"dir"`
	Env    []string `json:"env"`
	Outbox string   `json:"outbox"`
	Errbox string   `json:"errbox"`

	// Directory names for stdout and stderr. If empty, outputs as string into
	// Stdout and Stderr fields. If specified, creates a random file
	// name.stdout/stderr file.
	Outdir string `json:"outdir"`
	Errdir string `json:"errdir"`

	// Err is the most recent error, in case this is moved to a failure queue
	// before a process can start.
	Err string `json:"err"`
}

SubprocessInput contains information about the subprocess to run. Task values should be JSON-encoded versions of this structure.

func (*SubprocessInput) AsOutput

func (p *SubprocessInput) AsOutput() *SubprocessOutput

AsOutput creates an output structure from the input, for filling in results.

func (*SubprocessInput) JSON

func (p *SubprocessInput) JSON() []byte

JSON creates JSON from the input.

type SubprocessOutput

type SubprocessOutput struct {
	Cmd    []string `json:"cmd"`
	Dir    string   `json:"dir"`
	Env    []string `json:"env"`
	Err    string   `json:"err"`
	Outdir string   `json:"outdir"`
	Errdir string   `json:"errdir"`
	Stdout string   `json:"stdout"`
	Stderr string   `json:"stderr"`

	// If Outfile and Errfile were specified in the input, this contains
	// the final full pathnames of those files.
	Outfile string `json:"outfile"`
	Errfile string `json:"errfile"`
}

SubprocessOutput contains results from the subprocess that was run. Task values pushed to the outbox will contain this as a JSON-encoded string.

func (*SubprocessOutput) JSON

func (p *SubprocessOutput) JSON() []byte

JSON creates JSON from the input.

type TeeWriter

type TeeWriter struct {
	// contains filtered or unexported fields
}

TeeWriter writes to two writers what is written to it.

func NewTeeWriter

func NewTeeWriter(writers ...io.Writer) *TeeWriter

NewTeeWriter crates a new TeeWriter with the given writers to fan input out to.

func (*TeeWriter) Write

func (w *TeeWriter) Write(data []byte) (int, error)

Write forwards the data written to it to the two writers within.