Re: [Tzdist] Next step

[Paul Eggert <[e-mail]>]

Cyrus Daboo wrote:
> I think the tricky part about tzfile is dealing with the endian and 32/64-bit
> differences. I am assuming we would need to serve up all possible variants of
> those if we used the binary format as is.

That's not a problem, because there's only one binary format. The tzfile format 
is well-defined to be bigendian on all platforms, and to be independent of 
native word size. The reference code to read and write tzfile format is designed 
to work even on unusual machines where words are neither 32 nor 64 bits wide 
(e.g., Unisys mainframes). So there's not much need for a JSON tool to convert 
to "native" format.