[vorbis-dev] use of 'long' in vorbis structures

[Alejandro G. Belluscio]

Hello David,

Wednesday, May 21, 2003, 9:06:44 PM, you wrote:
>>> Since on every other modern sane platform out there
>>> sizeof(long)=sizeof(int) anyway wouldn't this be a fairly clean
>>> change?

>> This is not true. For instance, on most (I'd suspect 'all', but I
>> don't know) 64 bit systems, a long is 64 bits, but an int remains
>> 32 bit.

David> Not on versions of Irix I used a few years ago, mostly to avoid
David> breaking a bunch of old code. I don't have any alphas or sparcs
David> to test on right now.

David> I would argue that that makes it all the more reason to change
David> the public interface, so it's well-defined instead of
David> platform-dependent. Also, I'd suspect that those platforms
David> didn't have 64-bit integer dividers either, just like the SGI,
David> which means they'd incur a software call as well.

 Actually, Linux at 64bit, Solaris 64bit, HP-UX 64bit, xBSD at 64
bits, Ultrix, Windows Whatever 64 bits, and I'm taking a wild guess
and adding OSX when they do the 64bit transition (with the arrival of
the PPC970 on the fourth quarter, thou OSX may take longer) use LP64
model. That means long and pointers are 64 bits, integers are 32 bits.
So I'm most intrigued for you to define "modern sane platform" where
64bits actually matters (like 64 bits architectures).

Alejandro

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.