Improved encode/decode field values

[adamsussman]

Ok, here is an updated pull for the encode/decode stuff. This was done against the 0-9-1 errata as well as looking in Rabbit's internals. There is a great deal of divergence between what Rabbit does and what the 0-9-1 spec says. Also, there is some divergence between the 0-9-1 errata and Rabbit's actual implementation.

Here are the exact changes:

For encode:

• added 'd' field, double precision float
• added conversion of longs > 64 bits into long strings
• added 'V' no-field

Other notes on encode:

• I didn't add 'f', single precision floats can live inside doubles
• I let other small integers live inside 'I' 32 bit integers (b and s)
• Rabbit does not support unsigned integers at all, except for 'b' 8 bit, but this appears to be a bug in their code as that usage does not match their 0-9-1 errata and there are no other instances of unsigned types.
• I didn't add 'x' since I wasn't sure what data type to start with there

For decode:

• Added 'b', unsigned 8 bit integer (NB: Rabbit's code is different from the errata here, there are no unsigned integers supported otherwise)
• Added 'f' single precision IEE float
• Added 'd' double precision IEEE float (turned out to be the same as xdr)
• Added 's' signed 16 bit integer
• Added 'x' binary string
• Added 'V' no-field

[majek]

Thanks for this high quality patch.

I wouldn't be myself if I could resist tweaking your code a bit. Please take a look: c11aaf2ee805e6f66b6011260632f4b1315becc2 and cfba40e3b2abff598477df5c0f5d07ad6107f627

[adamsussman]

Looks good. Only issue I see is that the 'bit_length' function is only available starting in python 2.7, which we aren't on yet :(

-adam

On Fri, Jun 15, 2012 at 8:03 AM, Marek Majkowski reply@reply.github.com wrote:

Thanks for this high quality patch.

I wouldn't be myself if I could resist tweaking your code a bit. Please take a look: c11aaf2ee805e6f66b6011260632f4b1315becc2 and cfba40e3b2abff598477df5c0f5d07ad6107f627

---

Reply to this email directly or view it on GitHub: https://github.com/majek/puka/pull/19#issuecomment-6358724

[majek]

I was not aware of that. Let me use absolute values then.