February 2006 - Darwin-dev - lists.apple.com

Re: Exception catching problem since 10.4.3 using gcc 3.3 - IOKit. same kind of problem : crash after upgrading to QuickTime 704
by Ananda Tallur 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi Martin, I wanted to thank you once again for this further insight. Exception: EXC_BAD_ACCESS (0x0001) Codes: KERN_INVALID_ADDRESS (0x0001) at 0xfffffff4 The problem occurs when : -> compiling with gcc3.3, -> linking against QuickTime, directly or indirectly -> linking using flat_namespace mode Problem can be solved : -> by adding option -single_module at link time Hope this can help someone. Regards Anand On 4 janv. 06, at 01:50, Martin Costabel wrote: Martin Costabel wrote: [] __Unwind_GetIP __Unwind_GetLanguageSpecificData __Unwind_GetRegionStart These come from the module eh_personality.o from libstdc++.a. -- Martin _______________________________________________ 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… Also, this -single_module option has also enabled us recently to solve the same kind problem, but this time, due to QuickTime 7.0.4 upgrade. QuickTime 7.0.4 is dynamically linked against several gcc4 runtime dynamic libraries, as show the following command : otool -L /System/Library/Frameworks/QuickTime.framework/Versions/A/ QuickTime Here is the result : QuickTime: /System/Library/Frameworks/QuickTime.framework/Versions/A/ QuickTime (compatibility version 1.0.0, current version 63.0.0) /System/Library/Frameworks/CoreServices.framework/Versions/A/ CoreServices (compatibility version 1.0.0, current version 18.0.0) /System/Library/Frameworks/Carbon.framework/Versions/A/ Carbon (compatibility version 2.0.0, current version 128.0.0) /System/Library/Frameworks/CoreFoundation.framework/Versions/ A/CoreFoundation (compatibility version 150.0.0, current version 368.26.0) /System/Library/Frameworks/CoreAudio.framework/Versions/A/ CoreAudio (compatibility version 1.0.0, current version 1.0.0) /System/Library/Frameworks/AudioToolbox.framework/Versions/A/ AudioToolbox (compatibility version 1.0.0, current version 1.0.0) /System/Library/Frameworks/AudioUnit.framework/Versions/A/ AudioUnit (compatibility version 1.0.0, current version 1.0.0) /System/Library/Frameworks/Accelerate.framework/Versions/A/ Accelerate (compatibility version 1.0.0, current version 4.0.0) /System/Library/Frameworks/ApplicationServices.framework/ Versions/A/ApplicationServices (compatibility version 1.0.0, current version 22.0.0) /System/Library/Frameworks/Security.framework/Versions/A/ Security (compatibility version 1.0.0, current version 24989.0.0) /System/Library/Frameworks/IOKit.framework/Versions/A/IOKit (compatibility version 1.0.0, current version 275.0.0) /System/Library/Frameworks/SystemConfiguration.framework/ Versions/A/SystemConfiguration (compatibility version 1.0.0, current version 1.0.0) /usr/lib/libstdc++.6.dylib (compatibility version 7.0.0, current version 7.4.0) /usr/lib/libgcc_s.1.dylib (compatibility version 1.0.0, current version 1.0.0) /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 88.1.2) As you can see, QuickTime is now linked dynamically against : /usr/ lib/libstdc++.6.dylib, and /usr/lib/libgcc_s.1.dylib, which are components of C++ runtime library coming with gcc4. This time, the problem caused by this mix up was a crash (KERN_INVALID_ACCESS), when using QString objects (Trolltech Qt strings) : Below is an excerpt of the backtrace we had : Thread 0 Crashed: 0 libstdc++.6.dylib 0x955733bc std::basic_ostream<char, std::char_traits<char> >::sentry::sentry[in-charge] (std::basic_ostream<char, std::char_traits<char> >&) + 40 1 libstdc++.6.dylib 0x95575a98 std::basic_ostream<char, std::char_traits<char> >& std::operator<< <std::char_traits<char> > (std::basic_ostream<char, std::char_traits<char> >&, char const*) + 40 It clearly show that gcc4 libstdc++.6.dylib is being used, whereas the application is compiled with gcc3.3. libstdc++.6.dylib gets loaded because of QuickTime (7.0.4), and QuickTime is loaded because of Qt (Trolltech). time and DYLD_PRINT_LIBRARIES and DYLD_PRINT_BINDINGS at run time. What I observed makes me believe that the crucial variable is __Unwind_RaiseException To be complete for the record: There are a couple of other symbols behaving in the same way: This email sent to site_archiver(a)lists.apple.com

1 0

Re: intel source code release
by Michael Bartosh 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Feb 21, 2006, at 10:03 PM, Mark F. Murphy wrote: -mb _______________________________________________ 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… An all too familiar pattern in the world of Apple changes (what I posted on earlier in this thread). When Apple changes something, fear rears its ugly head. I've seen it over and over in my 22 years of Mac involvement. I'm glad something has happened which can end (at least for now) this current FUD thread. Just as often, what you perceive as fear becomes the reagent that allows Apple to be reactionary to a particular occurrence. This email sent to site_archiver(a)lists.apple.com

1 0

Problem of FTP on same machine
by Gaurav Kumar 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com Organization: hi I have two partitions of MAC-OSx 10.4 one with development tools another without development tools. I want to do ftp of data on same machine. If i do ftp on partition with development tool it works properly. But problem is with ftp on partition without development tools. In this when i log on with user name and password i must directly go to the users home folder. Instead of that i log into users folder. Then i have to give cd Username to reach the users folder.I want to know the real problem behind this. If anybody want more information do mail me. Thanks Gaurav -- Gaurav Kumar <gaurav.kumar(a)spsoftindia.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

Re: intel source code release
by Dave Schroeder 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com I see today a much more populated source tree for x86. Sorry; I had extremely good reason to post what and when I did. It wasn't just idle speculation. - Dave _______________________________________________ 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… Thank you to everyone responsible. Indeed, I also would like to pass along my thanks, since I was one of the people to comment on this with my concern before. An all too familiar pattern in the world of Apple changes (what I posted on earlier in this thread). When Apple changes something, fear rears its ugly head. I've seen it over and over in my 22 years of Mac involvement. I'm glad something has happened which can end (at least for now) this current FUD thread. Additionally, we have been directly told that a specific announcement on the general situation with Darwin and open source is still forthcoming. This email sent to site_archiver(a)lists.apple.com

1 0

Re: intel source code release
by Graham J Lee 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 22 Feb 2006, at 03:25, Peter O'Gorman wrote: Peter O'Gorman wrote: I see today a much more populated source tree for x86. Strongly seconded :-) _______________________________________________ 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 I was amazed to find that the gas sources had been split out of cctools, so they could be provided in accordance with the GPL, but no other part of cctools was made available. So I never did get an answer to my question. Thank you to everyone responsible. smime.p7s

1 0

Re: KERN_PROCNAME
by Eric Albert 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Feb 21, 2006, at 9:09 PM, Rob Braun wrote: I'm going to go wildly out on a limb here and suppose the headers from xcode 2.2.1 include some headers from 10.4.5 x86 xnu, which has not been released in source form. Or do we need to go and rummage around in our xcode 2.2.1 systems and try to extract the headers we need and try to cobble together some set of headers that might (possibly, maybe, probably not) work? -Eric _______________________________________________ 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… I've been going down the pre-doomed path of trying to build Libc 391.4.1 from 8G1454 (10.4.5 x86), and come across a discrepancy with xnu-792.6.61 (10.4.5 ppc) and my existing xnu 792.6.61 based 10.4.5 ppc system. Specifically, <sys/sysctl.h> in the source is missing KERN_PROCNAME, whereas <sys/sysctl.h> on my installed 10.4.5 ppc w/xcode 2.2.1 system has it. Poking around, no ppc sources from 10.3.9-10.4.5 have that macro defined, yet there it is, on my system. That's (mostly) correct. The Xcode 2.2.1 release matches the 10.4.4 (not 10.4.5) release for Intel. Where headers might be different between 10.4.4/Intel and 10.4.4/PowerPC, Xcode 2.2.1 has the Intel versions. Presumably you could build the PowerPC xnu, figure out which headers it installs, and copy those headers from an Xcode 2.2.1 install.... This email sent to site_archiver(a)lists.apple.com

1 0

KERN_PROCNAME
by Rob Braun 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mutt/1.4.1i I've been going down the pre-doomed path of trying to build Libc 391.4.1 from 8G1454 (10.4.5 x86), and come across a discrepancy with xnu-792.6.61 (10.4.5 ppc) and my existing xnu 792.6.61 based 10.4.5 ppc system. Specifically, <sys/sysctl.h> in the source is missing KERN_PROCNAME, whereas <sys/sysctl.h> on my installed 10.4.5 ppc w/xcode 2.2.1 system has it. Poking around, no ppc sources from 10.3.9-10.4.5 have that macro defined, yet there it is, on my system. I know xcode updates ship headers that are not necessarily in sync with libraries and other binaries that are shipped in software updates. Is this just a case of xcode updates and system updates being out of sync and I'm doomed, or am I missing something? I'm going to go wildly out on a limb here and suppose the headers from xcode 2.2.1 include some headers from 10.4.5 x86 xnu, which has not been released in source form. If this is the case, can we get xnu's headers released on darwinsource, since they've already been released in xcode 2.2.1? Or do we need to go and rummage around in our xcode 2.2.1 systems and try to extract the headers we need and try to cobble together some set of headers that might (possibly, maybe, probably not) work? For those curious as to how I came across this, I've been trying to build Libc (obviously). Libc requires xnu headers, and there is no xnu source released for 10.4.5 x86. So, I only need the headers to build Libc for the moment, so I tried substituting 10.4.5 ppc's xnu's headers into the build environment. Clearly a sketchy move, but the correct move of using the corresponding kernel's headers is not an option available to me. Rob _______________________________________________ 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: intel source code release
by Andrew Pinski 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com -- Pinski _______________________________________________ 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… On Feb 22, 2006, at 12:03 AM, Mark F. Murphy wrote: I'm glad something has happened which can end (at least for now) this current FUD thread. First there was no FUD in it. My comments about the compiler are true based on experiences I had at Apple interning. This email sent to site_archiver(a)lists.apple.com

1 0

Re: intel source code release
by Mark F. Murphy 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Feb 21, 2006, at 9:25 PM, Peter O'Gorman wrote: I see today a much more populated source tree for x86. mark -- --------------------------------------------------------------------------- Mark F. Murphy, Director Software Development <mailto:markm@tyrell.com> Tyrell Software Corp <http://www.tyrell.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… At 9:32 PM -0600 2/21/06, Dave Schroeder wrote: Thank you to everyone responsible. Indeed, I also would like to pass along my thanks, since I was one of the people to comment on this with my concern before. An all too familiar pattern in the world of Apple changes (what I posted on earlier in this thread). When Apple changes something, fear rears its ugly head. I've seen it over and over in my 22 years of Mac involvement. I'm glad something has happened which can end (at least for now) this current FUD thread. This email sent to site_archiver(a)lists.apple.com

1 0

Re: intel source code release
by oplus 22 Feb '06

22 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Feb 21, 2006, at 10:32 PM, Dave Schroeder wrote: On Feb 21, 2006, at 9:25 PM, Peter O'Gorman wrote: Peter O'Gorman wrote: I see today a much more populated source tree for x86. Maybe apologies would go well with those from some. _______________________________________________ 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… I was amazed to find that the gas sources had been split out of cctools, so they could be provided in accordance with the GPL, but no other part of cctools was made available. So I never did get an answer to my question. Thank you to everyone responsible. Indeed, I also would like to pass along my thanks, since I was one of the people to comment on this with my concern before. - Dave This email sent to site_archiver(a)lists.apple.com

1 0