Re: GCC 4.0 GetKeys() and BigEndianLong for Intel
Re: GCC 4.0 GetKeys() and BigEndianLong for Intel
- Subject: Re: GCC 4.0 GetKeys() and BigEndianLong for Intel
- From: Eric Albert <email@hidden>
- Date: Thu, 16 Jun 2005 03:06:28 -0700
On Jun 16, 2005, at 1:48 AM, MacArthur, Ian (SELEX) (UK) wrote:
Sorry I wasn't clear enough there. BigEndianLong is a long on
PowerPC and a struct on Intel. Here's the Intel definition, from
Endian.h:
struct BigEndianLong {
long bigEndianValue;
};
typedef struct BigEndianLong BigEndianLong;
So you want to swap theKeys[1].bigEndianValue on little-endian
systems, not theKeys[1].
Ouch!
I guess it's way to late to get it changed now, but that strikes me as
exactly the sort of thing that will be a pain.
There's no real reason *why* it has to be a struct rather than just a
long, I guess, but it forces us to use target specific code and to
know the detail of the underlying implementation. Seems a shame. If it
was just a long, we could simply call htonl() (for example) and get
the Right Thing on *any* architecture.
I agree. It'd be even better if it GetKeys always returned an array of
native-endian values, so you wouldn't have to do any swapping and you
could just use the same code across all architectures. After all,
that's what we've tried to do with most other APIs in the system.
Fortunately, since we haven't yet shipped a non-preview OS for Intel we
can still make changes like this. I'd suggest filing a bug report.
Send me the bug number when you get it back and I'll track it.
-Eric
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden