Re: Weakly-Linked Symbols
Re: Weakly-Linked Symbols
- Subject: Re: Weakly-Linked Symbols
- From: Eric Gouriou <email@hidden>
- Date: Tue, 28 Oct 2008 22:43:54 -0700
On Oct 28, 2008, at 9:39 PM, Dave Keck wrote:
Hey everyone,
This is the most appropriate list I could find for this question
pertaining to building weakly-linked libraries and frameworks.
Here's a quote from
http://developer.apple.com/documentation/MacOSX/Conceptual/BPFrameworks/Concepts/WeakLinking.html
"Note: When checking for the existence of a symbol, you must
explicitly compare it to NULL or nil in your code. You cannot use the
negation operator ( ! ) to negate the address of the symbol."
I was simply wondering why this is - what's different between this:
if ( !mySymbol )
and this:
if ( mySymbol == NULL )
to check whether the symbol exists or not? My understanding of C seems
to suggest that comparing something with NULL would be precisely the
same thing as comparing it with 0 (overlooking the (void *) cast).
I can only guess that this advice was meant to avoid a compiler
optimisation that
would defeat the check "if (!mySymbol)" but not "if (mySymbol == NULL)".
Unfortunately current gcc releases end up assuming that no symbol
can be NULL
either and also optimize away the "if (mySymbol == NULL)" check. This
makes
weak-linking pretty hard to use these days.
If you are in control of the weak-linked framework, the current
advice is to use
__attribute__((weak_import)) on its exported functions. Naively I'd
recommend
only doing so in cases where the library is indeed weak-linked and
avoiding
the attribute being present in normal cases, however I don't know for
sure
that it will make the code any worse if always present. Caveat
programmor.
Eric
Thanks for any insight!
David
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden