Encoding Arrays with File Paths

[Richard Gaskin]

Bob Sneidar wrote:

> I played around with this last night, and determined that the length does
> not have to be sent, the receiving process can simply keep reading for n
> characters, appending the data to a low level file in a repeat loop until
> the it variable is empty. I tried this and the entire payload was received.
>
> I think the advantage to this method is that I do not have to worry about
> file size, as any file can be transferred using this method, whereas
> otherwise a file can theoretically be large enough to overwhelm the sender
> or receiver.


I hope Mark Waddinghham will correct me if I'm wrong on this, but I believe the looped chunking you're scripting is effectively what happens automatically when you write the whole data.

I would be reluctant to rely on a handful of tests alone, and it never hurts to wear both belt and suspenders. I'm assuming HTTP 1.1 added a requirement for specifying payload length for good reason.

If I understand the protocol well enough, you're already sending metadata (for the file name), yes?  What's a few more bytes in that header to add payload size?


This excercise raises a question:  rather than invent another protocol, why not use HTTP? Saves dev time, eliminates the need to write and maintain documentation for a custom protocol, leverages existing robust tooling, allows for integration with other packages as customer needs evolve, and stakeholders often get to buy-in faster where open standards are employed.

--
Richard Gaskin
FourthWorld.com