Re: Crash in MusicPlayerStart - OS X bug?
Re: Crash in MusicPlayerStart - OS X bug?
- Subject: Re: Crash in MusicPlayerStart - OS X bug?
- From: Stephen Davis <email@hidden>
- Date: Tue, 9 Jan 2007 14:26:25 -0800
On Jan 9, 2007, at 1:33 PM, Simone Tellini wrote:
Il giorno 09/gen/07, alle ore 21:49, Daniel Jalkut ha scritto:
So that's the good news :) The bad news is obviously you're seeing
a crash.
actually, the bad news is that I can't see a crash on any of my
macs (MBP/Mini G4), while a single user does ;-)
By the way, my MusicSequence is full of events inserted with
MusicTrackNewUserEvent(), which I used to track the playing
position and update the UI accordingly (to do it, the callback
invokes performSelectorOnMainThread, to avoid problems).
But I would bet it's somehow related to your own code. If you
haven't run with the usual memory debugging tricks like
MallocScribble and guardmalloc, now might be a good time.
Things get a lot more interesting when I turn on Guard Malloc: now
I can see a crash as well, totally different from the one reported
by the user, but still it leads me to point the finger at
CoreAudio. In fact, the crash origins in the depths of a thread
which is entirely in system code:
Program received signal: "EXC_BAD_INSTRUCTION".
Did you set the MALLOC_ALTIVEC_SIZE=1 in your environment before
launching your app? By default, Guard Malloc does not vector-align
allocations (no, I don't know why) which violates the Mac OS X
contract that all allocated pointers are vector-aligned. All sorts
of weird things will occur unless you do this...
hth,
stephen
_______________________________________________
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