Re: About Universal Binaries...
Re: About Universal Binaries...
- Subject: Re: About Universal Binaries...
- From: Nick Zitzmann <email@hidden>
- Date: Tue, 31 Jan 2006 21:45:38 -0700
On Jan 31, 2006, at 6:06 PM, Juan P. Pertierra wrote:
From an initial look to the online Apple transition documentation
it seems all the differences that have to be addressed span from
the difference in byte order i.e. big endian vs little endian.
That's one thing, but there's a lot more than that:
1. On PPC, if some code divides by zero, then the result returned is
NaN and the executable will continue running. On Intel, if some code
divides by zero, then the executable crashes.
2. On PPC, the scalar values returned by sending a message to a nil
argument are technically undefined, but tend to return 0 for all
values that are not 64 bits long or longer. I wouldn't rely on this
behavior on Intel builds.
3. Watch how you're using Unicode buffers. On PPC builds, -[NSString
initWithData:encoding:] with NSUnicodeStringEncoding expects the data
to be in big-endian format if there is no BOM; on Intel builds, it'll
be expected to be in little-endian format if there is no BOM.
4. If you've got any vector code for PPC, you'll (most likely) have
to rewrite it for Intel.
5. A few Carbon data structures (e.g. KeyMap) are very different on
Intel than they are on PPC.
There's probably a few more, but that's all I can think of at the
moment...
Nick Zitzmann
<http://www.chronosnet.com/>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Cocoa-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden