Re: [potluck dish] the (potluck struct) module

[Nala Ginrut <nalaginrut@gmail.com>]

Hi Matt!
I'm finding `python pack' stuff, and I found your potluck is very cool!
Could you put it in a git repo somewhere? It would be worth to maintain
or receive patches. And maybe a better module name rather than (potluck
struc)
Thanks!

On Tue, 2016-02-16 at 05:53 -0800, Matt Wette wrote:
> First message got garbled.  Redo:
> 
> If you have used the Python struct module then this will look
> familiar.
> Otherwise, check out https://docs.python.org/2/library/struct.html
> 
> Attached are three files:
> * struct.scm: the source code 
> * struct.texi: documentation 
> * struct.test: test code
> 
> Struct Module
> =============
> 
> The '(potluck struct)' module provides procedures for packing and
> unpacking scheme data to and from bytevectors based on a format
> string.
> 
>      (use-modules (potluck struct))
> 
>      ;; pack two unsigned shorts and a double float in big endian
> order
>      (define data (pack ">2Hd" 3 22 34.0))
>      (write data) (newline)
>      ==>
>      #vu8(0 3 0 22 64 65 0 0 0 0 0 0)
> 
>      ;; verify using unpack
>      (write (unpack ">2Hd" data)) (newline)
>      ==>
>      (3 22 34.0)
> 
>  -- Scheme Procedure: pack format vals ...
>      Return a bytevector that contains encoded data from VALS, based
> on
>      the string FORMAT.
> 
>  -- Scheme Procedure: unpack format bvec
>      Return a list of scheme objects decoded from the bytevector
> BVEC,
>      based on the string FORMAT.
> 
>  -- Scheme Procedure: packed-size format
>      Return the number of bytes represented by the string FORMAT.
> 
>    The _format_ string used for PACK and UNPACK is constructed as a
> sequence of digits, representing a repeat count, and codes,
> representing
> the binary content.
> 
> The string may optionally begin with a special character that
> represents
> the endianness:
>     =        native endianness
>     <        little-endian
>     >        big-endian
>     !        network order -- i.e., big-endian
> 
> Type codes used in the format string are interpreted as follows:
>     x        blank byte
>     c        8-bit character
>     ?        boolean
>     b        signed 8-bit integer
>     B        unsigned 8-bit integer
>     h        signed 16-bit integer
>     H        unsigned 16-bit integer
>     i        signed 32-bit integer
>     I        unsigned 32-bit integer
>     l        signed 32-bit integer
>     L        unsigned 32-bit integer
>     q        signed 64-bit integer
>     Q        unsigned 64-bit integer
>     f        32-bit IEEE floating point
>     d        64-bit IEEE floating point
>     s        string
> 
>    The following issues remain to be addressed:
> string padding
>      'pack' assumes that the string length in the format is the same
> as
>      in the passed string.  Non-conformance is not trapped as an
> error.
>