On 8/4/06, Tinnus <[EMAIL PROTECTED]> wrote:
> 4) Garnet developers and users will be able to run *well written*
> applications on ALP.
That's what I fear. It's not well writen applications--it's what they
consider well written.
we have already been through one public transition as developers;
migrating from 68k to ARM. the same thing was said about well
written applications - and, it was proven, that unless you started
poking at hardware directly; your apps ran on the new architecture.
(i do recall a number of small PACE bugs in earlier debug releases)
the same thing will be true for the switch from Garnet to ALP.
granted; they'll probably still be ARM based - the framework is
still in place to allow a smooth transition.
if your garnet application is 68k - chances are it'll work fine. if you
make some native calls; you need to ensure that you are using
the PACE mechanism correctly. if not - it will definately crash and
burn on ALP. this means, if you make assumptions about specific
registers; and the data linked from them - your screwed.
native code on Garnet was a big patch in - the original plan was
just to migrate smoothly to a new architecture (ARM); but, developers
wanted to be able to write native code. the PACE interface and
PceNativeCall were put together quickly - as a temporary interface.
unless you needed native code; you probably never use it.
i definately see us having to re-compile our palmos code for ALP.
what i am hoping for; is that the ALP framework still supports the
standard API set so that all that truely happens is a conversion to
a new ABI format.
i've at least abstracted my r9 hacking techniques to using stub
functions - it even allows me to compile a 68k version *eek* :P
--
// Aaron Ardiri
--
For information on using the PalmSource Developer Forums, or to unsubscribe,
please see http://www.palmos.com/dev/support/forums/
