January 2007 - Darwin-dev - lists.apple.com

Re: Need compile-time date string in C++ file in XCode ...
by Steve Checkoway 06 Jan '07

06 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Jan 5, 2007, at 5:07 PM, Jay Reynolds Freeman wrote: __DATE__ and __TIME__ will do for fallback, but I would like to be able to use some of the formatting options of "date", so I wonder if anyone has any other ideas ... ? -- Steve Checkoway _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com Since you want to do this in Xcode, the xcode-users list might be more appropriate. smime.p7s

1 0

Re: Need compile-time date string in C++ file in XCode ...
by Jay Reynolds Freeman 06 Jan '07

06 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com __DATE__ and __TIME__ will do for fallback, but I would like to be able to use some of the formatting options of "date", so I wonder if anyone has any other ideas ... ? Thanks much to Steve Checkoway for reminding me about the built-in macros. -- Jay Reynolds Freeman --------------------- Jay_Reynolds_Freeman(a)mac.com http://web.mac.com/jay_reynolds_freeman (personal web site) On Jan 5, 2007, at 4:52 PM, Steve Checkoway wrote: On Jan 5, 2007, at 4:44 PM, Jay Reynolds Freeman wrote: I would like to give my program access to a string that shows the date it was compiled. If I build in the Terminal shell, I can use something like -- Steve Checkoway _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… See <http://gcc.gnu.org/onlinedocs/cpp/Standard-Predefined- Macros.html>. In particular, __DATE__ and __TIME__. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Need compile-time date string in C++ file in XCode ...
by Steve Checkoway 06 Jan '07

06 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Jan 5, 2007, at 4:44 PM, Jay Reynolds Freeman wrote: I would like to give my program access to a string that shows the date it was compiled. If I build in the Terminal shell, I can use something like -- Steve Checkoway _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com See <http://gcc.gnu.org/onlinedocs/cpp/Standard-Predefined- Macros.html>. In particular, __DATE__ and __TIME__. smime.p7s

1 0

Need compile-time date string in C++ file in XCode ...
by Jay Reynolds Freeman 06 Jan '07

06 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com I would like to give my program access to a string that shows the date it was compiled. If I build in the Terminal shell, I can use something like g++ -D DATE=`date` ... to create a preprocessor definition that contains a date string, and use it via something like char releaseDate[] = DATE; That works the way I expect it to when compiling in Terminal; when I (eg) printf releaseDate, I get (eg) Fri Jan 5 16:14:51 PST 2007, and yes, that works even if there are no double-quotes around "DATE". I would like to make the same kind of string available to a Cocoa application built in XCode. (XCode 2.4.1, Macbook, OS X 10.4.8). So I open up the inspector for my build target and try adding something to the "preprocessor macros" setting in the "Build" section of the inspector. What I put in is the text DATE=`date` separated by blank spaces from other macros that are there. It doesn't work. what appears where I was expecting, e.g., "Fri Jan 5 16:14:51 PST 2007" is instead the literal text (in a ".i" file -- it won't compile) `date` That is, the program acts as if the build process doesn't know to access the Unix "date" utility. I am probably making a simple mistake somewhere; can anyone advise me? I suppose I could always arrange for a compile-time script to create a C++ file containing the appropriate text, but I don't know how to make sure that that script runs before the C++ stuff compiles, so the program gets the newest version number. Thanks for putting up with a beginner ... -- Jay Reynolds Freeman --------------------- Jay_Reynolds_Freeman(a)mac.com http://web.mac.com/jay_reynolds_freeman (personal web site) _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kevent Kernel Panic by Unprivileged User
by Curtis Jones 06 Jan '07

06 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=IrARx1h01lLUUJ6OliVvM5/fe2U1HHC3AEgG2UJuH9BreidTVDY4hlG34HO3Se4fkJYzTkqGJ/5bdC1eHJdoyzkeWm4+Z/ZEDCHSZ2ZplxefD//76GYkozrdF9ltp/OCjT1U57kXwQMWxr8kaH84y9e0bXTXzbEfsKasUs5rsRQ= This looks somewhat (but not entirely) related: http://lists.apple.com/archives/Darwin-kernel/2005/Jul/msg00115.html -- Curtis Jones curtis.jones(a)gmail.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Kevent Kernel Panic by Unprivileged User
by Curtis Jones 06 Jan '07

06 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=VJ2ftrnZ9ig0kRv950NVNBTZEAzu2aqHgsA2KWlipHYa9IWLANvOkzxOty7DSCX1y+V8ElyngWlD6r+RcFJUMH/ehgHvXzVH8AsMODbZNCJeM8gYYdiwZIDcgZCHlK52gP+TvuQUvRLwp93RWqg1R6M9dlhrB/XHc6f3oJr2oeo= While it definitely wasn't my goal to find a way to panic my machine, I've done so anyway. I'm wanting to use kqueue/kevent to track fork() activity. I've taken my code and stripped it down to the minimum necessary to cause a panic. If I replace NOTE_TRACK with any other set of NOTE_* values, the program will work fine. Using NOTE_TRACK by itself or with any other NOTE_* values causes the panic. I've been crashing a Mac mini core duo, fyi. I've been running the program like this: # ./kevent & # ls which causes bash to fork, which ultimately causes the panic. gdb shows the following stacktrace: #0 Debugger (message=0x3c9540 "panic") at /SourceCache/xnu/xnu-792.13.8/osfmk/i386/AT386/model_dep.c:770 #1 0x00128d1f in panic (str=0x3cf1f4 "Unresolved kernel trap (CPU %d, Type %d=%s), registers:\nCR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\nEAX: 0x%08x, EBX: 0x%08x, ECX: 0x%08x, EDX: 0x%08x\nCR2: 0x%08x, EBP: 0x%08x, ESI: 0x%08x, EDI"...) at /SourceCache/xnu/xnu-792.13.8/osfmk/kern/debug.c:202 #2 0x001a3135 in kernel_trap (state=0xb733e30) at /SourceCache/xnu/xnu-792.13.8/osfmk/i386/trap.c:630 #3 0x0019a8d4 in trap_from_kernel () #4 0x003334f0 in kevent_register (kq=0x20c8c64, kev=0xb733ee8, p=0x0) at /SourceCache/xnu/xnu-792.13.8/bsd/kern/kern_event.c:1187 #5 0x00333a76 in filt_proc (kn=0x20c9dc8, hint=1073742038) at /SourceCache/xnu/xnu-792.13.8/bsd/kern/kern_event.c:520 #6 0x00332bd7 in knote (list=0x25ddb6c, hint=1073742038) at /SourceCache/xnu/xnu-792.13.8/bsd/kern/kern_event.c:1762 #7 0x0033a652 in fork (p=0x25dd9c4, uap=0x2634158, retval=0x263419c) at /SourceCache/xnu/xnu-792.13.8/bsd/kern/kern_fork.c:364 #8 0x00378337 in unix_syscall (state=0x26654cc) at /SourceCache/xnu/xnu-792.13.8/bsd/dev/i386/systemcalls.c:196 #9 0x0019acae in lo_unix_scall () I just wanted to run this by anyone who might have helpful input before I post a bug report. -- Curtis Jones curtis.jones(a)gmail.com #import <stdio.h> #import <stdlib.h> #import <sys/fcntl.h> #import <sys/event.h> #import <sys/types.h> #import <sys/time.h> #import <unistd.h> int main (int argc, char *argv[]) { int kqfd = 0; if (0 > (kqfd = kqueue())) { printf("%s().. failed to kqueue\n", __PRETTY_FUNCTION__); return EXIT_FAILURE; } printf("%s().. kqueue success, kqfd = %d\n", __PRETTY_FUNCTION__, kqfd); { int pid = getppid(); struct kevent kev; EV_SET(&kev, pid, EVFILT_PROC, EV_ADD, NOTE_TRACK, 0, NULL); if (0 != kevent(kqfd, &kev, 1, NULL, 0, NULL)) { printf("%s().. failed to kevent for pid = %d\n", __PRETTY_FUNCTION__, pid); return EXIT_FAILURE; } printf("%s().. kevent success for pid = %d\n", __PRETTY_FUNCTION__, pid); } while (1) sleep(1); close(kqfd); return EXIT_SUCCESS; } _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

turning off malloc output
by Alexander Peyser 05 Jan '07

05 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Thunderbird 3.0a1 (Macintosh/20061213) I was looking for a way to turn of malloc debugging for a program that test for a malloc failure. The same question went unanswered in http://lists.apple.com/archives/projectbuilder-users/2002/Nov/msg00179.html: * Subject: turning off malloc warnings * From: Andrew Scheck <email@hidden> * Date: Fri, 22 Nov 2002 13:17:14 -0500 This Makefile snippet: libnomalloc_printf.dylib: nomalloc_printf.c $(CC) $(CFLAGS) -dynamiclib -o $@ $< tester: tester.c libnomalloc_printf.dylib $(CC) $(CFLAGS) -Wall \ -force_flat_namespace -bind_at_load \ -L. -lnomalloc_printf \ -o $@ $< where nomalloc_printf.c just contains: void malloc_printf(const char* format, ...) will shut malloc up. This appears to be the minimum set. Is there some way to turn the malloc definitions into weak references and avoid this complication (such as force_flat_namespace)? Make these functions into variables, and name them uniquely in each library? In general, each library should be able to override mallocs, and have theirs overridden uniquely -- see the regex library for an exercise in malloc frustration. begin:vcard fn:Alexander Peyser n:Peyser;Alexander email;internet:a.peyser@umiami.edu x-mozilla-html:TRUE version:2.1 end:vcard _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

re: Re: Cann't make connection to a graphics IOService "IOFramebuffer"
by Haarm-Pieter Duiker 03 Jan '07

03 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Organization: Duiker Research Thread-index: Accvi2jRrPvzwRYiQ+e9QFEQJ6+czA== Hello, I’m in the process of writing a simple tool that will allow users to dynamically switch between their existing monitor profiles and other luts. I’m running into the same stumbling block that the message below mentions, namely that IOServiceOpen is failing to open the IOFramebuffer service because the window server has exclusive access to it. Are there other apis that should be used for accessing the graphic card’s gamma table? Any help you can provide would be appreciated. Regards, HP Duiker Duiker Research Referenced message can be found here: http://lists.apple.com/archives/darwin-kernel/2005/Apr/msg00074.html Text: Hi Amit, I am trying to execute the code written below. Here IOServiceOpen is giving error -536870203 Can any one tell the reason for this error. There's documentation that describes how to interpret I/O Kit error codes. <http://developer.apple.com/documentation/DeviceDrivers/Conceptual/Accessing…> In this case, -536870203 is 0xe00002c5, which indicates a system of I/O Kit (0x38) and, within that, the common subsystem (0). The error value is that in <IOKit/IOReturn.h>, namely: There's documentation about this error code as well. <http://developer.apple.com/documentation/DeviceDrivers/Conceptual/Accessing…> What is it you are trying to do? Cheers, -- J-Sun Seng Networking & Communications / Hardware Developer Technical Support Apple Computer, Inc. #1 Infinite Loop M/S: 303-2T _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com At 17:14 +0530 15/4/05, Amit Gupta wrote: #define kIOReturnExclusiveAccess iokit_common_err(0x2c5)// exclusive access and // device already open The Window Server has the IOFramebufferUserClient open exclusively, thus the reason for the error you are seeing.

1 0

Re: Mapping Physical Memory to Multiple Virtual Memory Addresses
by chas williams - CONTRACTOR 03 Jan '07

03 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com In message <459AB119.2000805(a)cs.uiuc.edu>,John Criswell writes: >Is there a way in Darwin to map, within a single address space, physical >memory to multiple virtual addresses? use IOConnectMapMemory(). i have done this to map the same physical region back to back in virtual space to handle wrap around on a streaming buffer. address = 0; size = SIZEOF_PHYS_REGION; kresult = IOConnectMapMemory(connection, kMyUserClientBuffer, mach_task_self(), &address, &size, kIOMapAnywhere); address += size; kresult = IOConnectMapMemory(connection, kMyUserClientBuffer, mach_task_self(), &address, &size, 0); if you didnt care where the second mapping was located you could just pass kIOMapAnywhere instead of 0. there is likely a more efficient way to do this with mach_vm_remap() after you get the first mapping. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Mapping Physical Memory to Multiple Virtual Memory Addresses
by John Criswell 02 Jan '07

02 Jan '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Organization: University of Illinois User-agent: Thunderbird 1.5.0.9 (X11/20061206) Dear All, Is there a way in Darwin to map, within a single address space, physical memory to multiple virtual addresses? Thanks in advance. Regards, John Criswell _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0