On 1/5/12 Jan 5 -3:19 PM, Frank Goenninger wrote:
Am 05.01.2012 um 22:13 schrieb Luís Oliveira:
Hello John,
git blame says you sent us a patch to add long-long on Allegro on 64-bit platforms. Robert says that isn't quite working. Do you have any recollection of any limitations of this long-long support or something like that?
Meanwhile, I'll revert the patch: https://github.com/cffi/cffi/commit/80a06643361af383cadcbb7232ba3d73b828a96d unless you have some objection since I don't have access to a 64-bit ACL.
On Thu, Jan 5, 2012 at 8:18 PM, Robert Goldman rpgoldman@sift.info wrote:
I am pretty certain that *no* version of Allegro supports :long :long. I have verified this on 64-bit ACL for both Mac OS X and Linux. So I would strongly urge you to record that ACL has no-long-long.
Hm - why not check back with Franz? I had the impression ACL 64bit *does* support long long. Any specific tests I could run that would convince all of us ? (I do have 64bit ACL on Windows).
The manual says not. If you look at the list of primitive types at this page:
http://www.franz.com/support/documentation/8.2/doc/ftype.htm#primitive-types...
Here's a snippet from their foreign types grammar:
primitive-type := :fixnum :int :long :short :char :void :unsigned-int :unsigned-long :unsigned-short :unsigned-char :float :double :nat :unsigned-nat
It should be pretty easy for you to test this on your version --- an unmodified version of the CFFI tests will crash on ACL for Mac OS X and Linux (64 bit). So if you try to run the tests, you should see the same behavior, unless Windows offers :long :long where the other two platforms don't.
Looks like the only 8 byte quantity is :double for ACL.
Cheers,
r