Re: Question: What is the impact of changing .cpp AudioUnitEffect source to .mm
Re: Question: What is the impact of changing .cpp AudioUnitEffect source to .mm
- Subject: Re: Question: What is the impact of changing .cpp AudioUnitEffect source to .mm
- From: Kyle Sluder <email@hidden>
- Date: Sun, 20 Jun 2010 23:11:18 -0700
On Sun, Jun 20, 2010 at 10:02 PM, Iain Houston <email@hidden> wrote:
> Kyle - how can we evidence this?
Detailed technical analysis, including precisely why and when
objc_msgSend will block, has already been posted to this thread.
> objc_msgSend is supposed not to block nonatomic methods.
There is no such thing as a "nonatomic method." Objective-C does have
"nonatomic properties," which imply absolutely nothing about
non-blocking behavior. A setter for a nonatomic property might do
nontrivial work that requires a lock, but still can't guarantee that
all threads will see atomic behavior. Likewise, it's possible to
implement an atomic method without taking any locks at all.
Furthermore, the lock we're talking about blocking on happens at
low-level method dispatch, so nothing about properties, which are a
concept built on top of regular methods, could possibly avoid this
lock.
> The answer is relevant here. e.g. to anyone attempting to distinguish what aspects of the program
> structure of the aurioTouch sample are historical and what's currently still essential technically
> to ensure no hiccups in obj-c deployment of remote io unit.
Um, huh? What does "obj-c deployment of remote io unit" mean?
--Kyle Sluder
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden