Option to Dump RabbitMQ Headers and/or Other Properties

[sshaw]

In this case I reckon it'd be better to dump the payload to a YAML file with two documents, one with the properties, one with the payload:

---
{
  "priority": 0,
  "timestamp":  1455915599,
  "headers": {
    "foo": "bar"
  }

---
{
   "a": 123,
   "b": 456
}

Thoughts?

[dubek]

@sshaw I like the idea of dumping (optionally?) the message headers. What do you mean by "other properties"?

However, I don't think the choice for YAML here is the right one, because:

1. The message body may be binary.
2. As in your example, you already mix two serialization formats: JSON and YAML.

I suggest instead that this new optional mode will save the headers in a JSON file next to each saved message payload; so the program will produce:

/tmp/msg-0000                    # That's the payload, possibly binary
/tmp/msg-0000-amqp-headers.json  # Message 0's headers serialized as JSON
/tmp/msg-0001
/tmp/msg-0001-amqp-headers.json
...

What do you think?

[sshaw]

@sshaw I like the idea of dumping (optionally?) the message headers. What do you mean by "other properties"?

I just meant the AMQP message attributes, or properties, or whatever the official name is.

I suggest instead that this new optional mode will save the headers in a JSON file next to each saved message payload; so the program will produce:

That makes sense, but I would use whatever the AMQP specs uses in the filename instead of "headers" (though maybe they do use "headers" 😄).

[dubek]

Scanning the AMQP docs it seems that message properties is the right term. So:

/tmp/msg-0000
/tmp/msg-0000-properties.json
/tmp/msg-0001
/tmp/msg-0001-properties.json
...

@sshaw do you want to try and craft this into a pull request?

[sshaw]

@sshaw do you want to try and craft this into a pull

Sure! I just (almost) finished my first Go program (though I'm not necessarily itching for more Go 😏). Give me a few days and I'll put something together...