Do not encode data, keep as buffer

[StefanScherer]

I'm using the ipp module to download print documents. This is one of the seldom cases where the module fills obj.data. Unfortunately the data was corrupted and eg. a PDF file couldn't be downloaded correctly.

Keeping the data in a buffer solves the problem. Instead of encode the data with the encoding for the IPP header I just slice the buffer at the position after the IPP header.

[williamkapke]

Interesting.

This change will break someone that was depending on it returning the decoded string though. We will need to do a major version bump.

The relevant definitions from the spec:

data = OCTET-STRING
BYTE = %x00-FF
OCTET-STRING = *BYTE

^{https://tools.ietf.org/html/rfc2910#section-3.2}

So, hmm. I guess that looks like a Buffer. 🤔

Does it throw an error when it tries to decode?

[michaelrsweet]

@williamkapke FWIW, the more recent/relevant reference is in RFC 8010 now: https://tools.ietf.org/html/rfc8010#section-3.2

:)

[williamkapke]

@michaelrsweet ah right! Thanks- it's worth a lot!

[StefanScherer]

No error is thrown during the decode. The data (eg. PDF) was corrupt.