Hello,
I am porting a Classic application to Carbon and am experiencing a crash
from using CFSTR() that I cannot understand. The program always crashes for
a colleague of mine on three different machines running 10.4, 10.3, and
10.2. From what I can tell from the crash log, it crashes inside of CFSTR()
for him, but it never crashes for me.
In my code that initializes CoreMIDI, I was originally doing the following.
OSStatus InitCoreMidiDriver()
{
OSStatus err;
CFStringRef pname, strtemp;
err = MIDIClientCreate(CFSTR("Bol Processor"), CMNotifyCallback, NULL,
&CMClient);
To pinpoint the error, I tried this instead:
Alert1("1: At the top of InitCoreMidiDriver(); about to create client
name string ...");
strtemp = CFSTR("Bol Processor");
if (strtemp == NULL) {
Alert1("2: Err. CFSTR() returned NULL while creating client name
string.");
return -1;
}
Alert1("3: Created client name string; about to call MIDIClientCreate()
...");
err = MIDIClientCreate(strtemp, CMNotifyCallback, NULL, &CMClient);
My colleague reports that he only sees the first alert message before the
crash, so I am assuming that CFSTR() does not finish before the crash. Here
is an excerpt from a sample crash log:
Date/Time: 2007-03-08 00:25:21 +0100
OS Version: 10.3.9 (Build 7W98)
Report Version: 2
Exception: EXC_BAD_ACCESS (0x0001)
Codes: KERN_INVALID_ADDRESS (0x0001) at 0x436f7265
Thread 0 Crashed:
0 net.sourceforge.bolprocessor 0x00003aac _start + 0x100
1 net.sourceforge.bolprocessor 0x0007f808 InitCoreMidiDriver + 0x30
2 net.sourceforge.bolprocessor 0x000046c8 RegisterMidiDrivers + 0x78
3 net.sourceforge.bolprocessor 0x00003c68 main + 0x90
4 net.sourceforge.bolprocessor 0x00003b28 _start + 0x17c
5 net.sourceforge.bolprocessor 0x000039a8 start + 0x30
PPC Thread State:
srr0: 0x00003aac srr1: 0x0000d030 vrsave: 0x00000000
cr: 0x42000222 xer: 0x00000004 lr: 0x00003a84 ctr: 0x90001a44
r0: 0x00000000 r1: 0xbffffb50 r2: 0xa00016b4 r3: 0x00000000
r4: 0x00117500 r5: 0x0011755c r6: 0x8fe505fc r7: 0x8fe50784
r8: 0x00000000 r9: 0x436f7265 r10: 0x00000000 r11: 0x436f7265
r12: 0x90001a44 r13: 0x00000000 r14: 0x00000000 r15: 0x00000000
r16: 0x00000000 r17: 0x00000000 r18: 0x00000000 r19: 0x00000000
r20: 0x00000000 r21: 0x00000000 r22: 0x00000000 r23: 0x00000000
r24: 0x00000000 r25: 0x00000000 r26: 0x00000000 r27: 0x0000000c
r28: 0x00101df8 r29: 0x00000001 r30: 0x000f6f08 r31: 0x000039c0
The return address in InitCoreMidiDriver (0x0007f808, offset 0x30)
corresponds to the next instruction after the call to
___CFStringMakeConstantString which is what I am guessing the macro CFSTR()
resolves to. You can see that the stack trace though indicates that the
program seems to be executing in _start again! (And I just noticed that the
"address" 0x436f7265 is "Core" in ASCII).
The program is a Mach-O executable being built by CodeWarrior 8.3 on OS X
10.2.8. The program is 100% C and it links to CoreMIDI, CoreAudio, Carbon,
and the System frameworks; crt1.o; and the MSL libraries. In the
CodeWarrior settings, I have "Reuse Strings", "Make Strings Read-Only", and
"Linker Pools Strings" checked.
I know this is a lot of information, but I am completely mystified by this
bug and since I cannot reproduce it myself, it is all the more confusing. I
hope that someone here will have some suggestions about what could be wrong.
Thanks in advance!!
Anthony Kozar
anthonykozar AT sbcglobal DOT net
http://anthonykozar.net/
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Carbon-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/carbon-dev/email@hidden
This email sent to email@hidden
