Re: floatValue of nil NSNumber unreliable (was Re: Bug or feature?)
Re: floatValue of nil NSNumber unreliable (was Re: Bug or feature?)
- Subject: Re: floatValue of nil NSNumber unreliable (was Re: Bug or feature?)
- From: Daniel Jalkut <email@hidden>
- Date: Tue, 14 Feb 2006 13:48:45 -0500
Yeah - it's confusing. I am assuming that "usually" means "for all
messages whose return value goes into general purpose register $r3".
I'm looking at the Objective-C PowerPC msgSend implementation, and it
looks like it relies entirely on "leaving $r3 alone when nil" to get
the expected behavior of returning nil as the result.
However, the bold note in the documentation that describes "identical
behavior" in Intel as of Xcode 2.2 seems to contradict the caveats of
the paragraph that follows.
It seems like there has been a general wisdom that float-returning
methods should not be relied upon when nil-messaging. All signs
point to "YES" for continuing to follow that wisdom.
Daniel
On Feb 14, 2006, at 1:40 PM, Martin Hairer wrote:
Of course, one may wonder what the word "usually" means in this
context. It seems
to me that this sentence could just as well have been replaced by
something like
"on PowerPC based machines the return value is undefined", which
would have been
less misleading. Regards,
_______________________________________________
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