September 2009 - Darwin-dev

obcj4/CFLite on Windows
by Brent Fulgham 16 Sep '09

16 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:date:message-id:subject :from:to:content-type:content-transfer-encoding; bh=ArxH/jlTDF5CXoxBKaZ0E/4oZ1NPoZCUiPfotRw8x9E=; b=P9bkjBQgtKFk92UU3Y4I69dYadao/U7PkCaIDp1eRjwWFgDUw+o5IuX0W0MYE9CteS ZRRa/V9CDn2FiX0OsUnJX3STh6LgK7FGnSH0VUJP5nZ3eYgGpLshQfdQFWcDxcBrZg5I n37u1Hj+KeVTyc1tFc3dmS8p/gUDQvJe9qVZU= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type :content-transfer-encoding; b=RGxqauf4yFydNQYJaJlRKoRcFPwqamuu02NOnWEJ7C/OzGWmH5rg9nBMAYgjv4E6qw hX8di7Su3KQ/gG9HPoUR2ttSC+8s4tT051/FKH7Rpv7W+/11y1C8vS+p/t17DBx5gjfb qR67/qIBjq41AmJCPrKrk5xQV1ZIz4zcaDs60= Hi again, To recap: 1. I am attempting to get CFLite up and running on Windows using the 10.6 source base. 2. I have successfully built objc4 for use with CFLite. 3. I have gotten CFLite (550) building on Windows. 4. CFLite (550) crashes on startup. Originally I thought this was an issue with the objc4 library, but I was able to confirm that I see the same crash when running with the objc.dll that ships with Safari on Windows, and was likewise able to use my own locally-built objc.dll with Safari without problems. Therefore I assume my problem is in my CFLite build, and more specifically in the initialization logic. During program initialization (specifically CoreFoundation bundle registration) CFlite attempts to convert a pathname into an array of CFStrings containing the path components. Down a few layers in this code, it fails: objc_debug.dll!objc_msgSend(objc_object * a=0x6a363ff8, objc_selector * b=0x0012ea8c, ...) Line 123 C CFLite_Debug.dll!CFRetain(const void * cf=0x101e2768) Line 535 + 0x4c bytes C++ CFLite_Debug.dll!__CFStringCreateImmutableFunnel3(const __CFAllocator * alloc=0x101e1060, const void * bytes=0x00c6ee10, long numBytes=0, unsigned int encoding=1280, unsigned char possiblyExternalFormat=0, unsigned char tryToReduceUnicode=0, unsigned char hasLengthByte=0, unsigned char hasNullByte=0, unsigned char noCopy=0, const __CFAllocator * contentsDeallocator=0x101e1060, unsigned int converterFlags=0) Line 1224 + 0xb bytes C++ CFLite_Debug.dll!CFStringCreateWithSubstring(const __CFAllocator * alloc=0x101e1060, const __CFString * str=0x00c6edc8, CFRange range={...}) Line 1558 + 0x36 bytes C++ CFLite_Debug.dll!CFStringCreateArrayBySeparatingStrings(const __CFAllocator * alloc=0x101e1060, const __CFString * string=0x00c6edc8, const __CFString * separatorString=0x00c62158) Line 3869 + 0x3a bytes C++ > CFLite_Debug.dll!WindowsPathToURLComponents(const __CFString * path=0x00c6edc8, const __CFAllocator * alloc=0x101e1060, unsigned char isDir='') Line 3485 + 0x1b bytes C++ CFLite_Debug.dll!WindowsPathToURLPath(const __CFString * path=0x00c6edc8, const __CFAllocator * alloc=0x101e1060, unsigned char isDir='') Line 3521 + 0x11 bytes C++ CFLite_Debug.dll!_convertToURLRepresentation(__CFURL * url=0x00c61f90) Line 3620 + 0x14 bytes C++ CFLite_Debug.dll!CFURLGetString(const __CFURL * url=0x00c61f90) Line 2383 + 0x9 bytes C++ CFLite_Debug.dll!_CFBundleURLLooksLikeBundleVersion(const __CFURL * url=0x00c61f90, unsigned char * version=0x0012f377) Line 1842 + 0x17 bytes C++ CFLite_Debug.dll!_CFBundleCreate(const __CFAllocator * allocator=0x101e1060, const __CFURL * bundleURL=0x00c621d8, unsigned char alreadyLocked=0, unsigned char doFinalProcessing='') Line 1229 + 0x13 bytes C++ CFLite_Debug.dll!CFBundleCreate(const __CFAllocator * allocator=0x101e1060, const __CFURL * bundleURL=0x00c621d8) Line 1334 + 0x11 bytes C++ CFLite_Debug.dll!RegisterCoreFoundationBundle() Line 1203 + 0x13 bytes C++ CFLite_Debug.dll!DllMain(HINSTANCE__ * hInstance=0x10000000, unsigned long dwReason=1, void * pReserved=0x0012fd24) Line 1219 + 0x5 bytes C++ CFLite_Debug.dll!__DllMainCRTStartup(void * hDllHandle=0x10000000, unsigned long dwReason=0, void * lpreserved=0x00000000) Line 498 + 0x8 bytes C CFLite_Debug.dll!_DllMainCRTStartup(void * hDllHandle=0x76ef9083, unsigned long dwReason=269031836, void * lpreserved=0x10000000) Line 462 + 0x11 bytes C The failure happens when attempting to call CFRetain on a static empty string CoreFoundation object: CFString.c: Line 1224: return CFRetain(kCFEmptyString); // Quick exit; won't catch all empty strings, but most Where kCFEmptyString is declared as: CFString.c: Line 152: CONST_STRING_DECL(kCFEmptyString, "") This is the first attempt to call objc_msgSend in the program, and it fails in the assembly logic where it is trying to perform a cache lookup for the selector ("retain" in this case). I'm wondering if perhaps the selector information is never getting populated. I put a breakpoint in _objc_init_image and have not hit this during startup, so I'm guessing that some kind of initialization of the objc runtime is not happening. However, I don't see any explicit calls to this initialization logic elsewhere, and the CFLite in Darwin seems to work properly. Any suggestions for where this initialization should be happening? Thanks, -Brent _______________________________________________ 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

Open Source Release: Mac OS X 10.6.1, iPhone 3.1, libdispatch
by William Siegrist 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com http://opensource.apple.com/ Thank you, William Siegrist Darwin Team _______________________________________________ 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… The Darwin Team is proud to announce the release of the open source components in Mac OS X 10.6.1 (Darwin 10.1) and iPhone 3.1. They can be found at the usual location below. Also, we have added libdispatch to the Darwin releases for completeness since libdispatch was released on Mac OS Forge. This email sent to site_archiver(a)lists.apple.com

1 0

Re: fsaclctl: MIA in Snow Leopard
by Nathan Gray 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi Terry, On Tue, Sep 15, 2009 at 2:21 PM, Terry Lambert <tlambert(a)apple.com> wrote: > On Sep 15, 2009, at 1:58 PM, Nathan Gray wrote: >> >> Hi folks, >> >> My Backup Bouncer metadata test suite used to use fsaclctl to enable >> ACLs on its test volumes but that program has vanished from snow >> leopard. Can anybody tell me what the procedure for >> enabling/disabling ACLs is now? I've searched for ACL/acl in the man >> pages for hdiutil, mount_hfs, diskutil, etc. without luck. > > It is enabled by default, and has been since Leopard. Ah, OK. It turns out my chmod +a ... was failing for some other reason and I assumed it was related to the failure to find fsaclctl. Thanks! > The call to allow it to be enabled/disabled on the fly is to do an fsctl on > the volume with the HFS_SETACLSTATE operand. Note that you will need to run > as root for this to be effective. I probably don't need it after all, but that's good to know. Thanks, -n8 -- http://n8gray.org _______________________________________________ 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: carbon throwing objc exceptions
by Nathan Herring 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AQHKNX3ygeYV2+Z12k+e/p33wOh/FJDXDfCAgAElHtCAAHuIgP//i4Kg Thread-topic: carbon throwing objc exceptions One could imagine there being some flag set in new code that could perturb the operation of ObjC[++] unwinding s.t. the ObjC runtime creates thunks between old-style and new-style code that do the marshalling between libUnwind exceptions and setjmp/longjmp exceptions. Then you only really care about the cases where you used to have a callback written in C or some other not-unwindable language where you were quietly unwinding before and now you can't -- to which I argue that you could never have counted on that before, and depending on whether or not you own that code, was probably fragile anyway. (Like in OS n, it was a pure callback, and in os n+1, it created side effects that needed cleanup, and now you've broken it.) -----Original Message----- From: Greg Parker [mailto:gparker@apple.com] Sent: Tuesday, September 15, 2009 3:16 PM To: Nathan Herring Cc: Joel Reymont; Shawn Erickson; darwin dev Subject: Re: carbon throwing objc exceptions On Sep 15, 2009, at 2:58 PM, Nathan Herring wrote: > I have separately made the suggestion, via rdar://7225898, that > Apple move ObjC[++] to use libUnwind, so that callbacks of any kind > would see this as a (possibly foreign) exception, and would either > (1) terminate due to un-unwindability (rather than quietly unwind > stack that cannot be unwound), (2) allow for standard exception > handling, including catching and rethrowing for finally-like behavior. That's unfortunately not binary-compatible with existing 32-bit Mac binaries. Such code does not include the unwind tables that libunwind needs to do its work, so exceptions that work today would almost always terminate as un-unwindable. iPhone and 64-bit Mac do have C++-compatible Objective-C exceptions courtesy of libunwind. Objective-C exceptions run destructors and `catch (...)` clauses, and C++ exceptions unlock `@synchronized` and run `@catch (...)` and `@finally` clauses. (But C++ and Objective-C exceptions are implemented with setjmp/ longjmp on iPhone OS, so you still get the same clumsy machine code that started this whole thread.) -- Greg Parker gparker(a)apple.com Runtime Wrangler _______________________________________________ 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: carbon throwing objc exceptions
by Greg Parker 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com -- Greg Parker gparker(a)apple.com Runtime Wrangler _______________________________________________ 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 Sep 15, 2009, at 2:58 PM, Nathan Herring wrote: I have separately made the suggestion, via rdar://7225898, that Apple move ObjC[++] to use libUnwind, so that callbacks of any kind would see this as a (possibly foreign) exception, and would either (1) terminate due to un-unwindability (rather than quietly unwind stack that cannot be unwound), (2) allow for standard exception handling, including catching and rethrowing for finally-like behavior. That's unfortunately not binary-compatible with existing 32-bit Mac binaries. Such code does not include the unwind tables that libunwind needs to do its work, so exceptions that work today would almost always terminate as un-unwindable. iPhone and 64-bit Mac do have C++-compatible Objective-C exceptions courtesy of libunwind. Objective-C exceptions run destructors and `catch (...)` clauses, and C++ exceptions unlock `@synchronized` and run `@catch (...)` and `@finally` clauses. (But C++ and Objective-C exceptions are implemented with setjmp/ longjmp on iPhone OS, so you still get the same clumsy machine code that started this whole thread.) This email sent to site_archiver(a)lists.apple.com

1 0

RE: carbon throwing objc exceptions
by Nathan Herring 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AQHKNX3ygeYV2+Z12k+e/p33wOh/FJDXDfCAgAElHtA= Thread-topic: carbon throwing objc exceptions I have separately made the suggestion, via rdar://7225898, that Apple move ObjC[++] to use libUnwind, so that callbacks of any kind would see this as a (possibly foreign) exception, and would either (1) terminate due to un-unwindability (rather than quietly unwind stack that cannot be unwound), (2) allow for standard exception handling, including catching and rethrowing for finally-like behavior. This would have some benefit towards a more unified exception model. -----Original Message----- From: darwin-dev-bounces+nathanh=microsoft.com(a)lists.apple.com [mailto:darwin-dev-bounces+nathanh=microsoft.com@lists.apple.com] On Behalf Of Joel Reymont Sent: Monday, September 14, 2009 2:25 PM To: Shawn Erickson Cc: darwin dev Subject: Re: carbon throwing objc exceptions Sean, On Sep 14, 2009, at 9:57 PM, Shawn Erickson wrote: > Why are you attempting to catch the except at all? It looks like no > attempt is made to deal with the exception and in general exceptions > in Cocoa are truly exceptional and imply a mistake by the programmer. Quoting from an existing bug [2]: "If a cocoa exception is thrown in the middle of a gecko callback, really bad things happen (like you can no longer load any pages, for example). Cocoa exceptions are basically setjmp/longjmp, and, presumably, without a handler, will jump back to the main event loop's default handler. This will unwind the stack without any stack cleanup." > I wouldn't attempt to catch the exception but instead let a top > level handler get it and then make a decision to log at that point > (look at NSExceptionHandler). This is of no use since the issue is C++ stack unwinding [1]. More he > Also CoreFoundation is not Carbon. My bad. Thanks, Joel [1] https://wiki.mozilla.org/Cocoa_In_XPCOM [2] https://bugzilla.mozilla.org/show_bug.cgi?id=163260 --- fastest mac firefox! http://wagerlabs.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/nathanh%40microsoft.com This email sent to nathanh(a)microsoft.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: fsaclctl: MIA in Snow Leopard
by Terry Lambert 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi folks, My Backup Bouncer metadata test suite used to use fsaclctl to enable ACLs on its test volumes but that program has vanished from snow leopard. Can anybody tell me what the procedure for enabling/disabling ACLs is now? I've searched for ACL/acl in the man pages for hdiutil, mount_hfs, diskutil, etc. without luck. It is enabled by default, and has been since Leopard. -- Terry _______________________________________________ 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 Sep 15, 2009, at 1:58 PM, Nathan Gray wrote: The call to allow it to be enabled/disabled on the fly is to do an fsctl on the volume with the HFS_SETACLSTATE operand. Note that you will need to run as root for this to be effective. This email sent to site_archiver(a)lists.apple.com

1 0

fsaclctl: MIA in Snow Leopard
by Nathan Gray 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi folks, My Backup Bouncer metadata test suite used to use fsaclctl to enable ACLs on its test volumes but that program has vanished from snow leopard. Can anybody tell me what the procedure for enabling/disabling ACLs is now? I've searched for ACL/acl in the man pages for hdiutil, mount_hfs, diskutil, etc. without luck. Thanks, -n8 PS: Please CC me with replies. -- http://n8gray.org _______________________________________________ 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

Testing Objc4
by Brent Fulgham 15 Sep '09

15 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:date:message-id:subject :from:to:content-type; bh=xr0h/1WbSxRaLFwVZStbQgrjQ7E8aXXETd5TOOYMkkE=; b=pM4VJkhGQXD9bbyjzadJjJ6I5mvWfvJJ7pOqK/oDmwLuRibhK0m5S3zXWw2Ea/GV7u R8oNYTXFvtMD16pOXgY3OYuE6PghPRYW0F25H3h75LB5oeISBctGiGNhBrO1kl54M1KG zHiz6CEJwbgef45/60qJ64fc74v29iu2KCav0= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=oNifxvMU7N7qCudDHSi9W+jP3tkhlo704A878NVQWnbRKmcQmyp/Z0gpUn3b8yhKjb znAzOYQl+3s/nVW1M0Pj4XYvFJmwc1zt8TWSMAyuUMrzmeloUf6CFiSsyHv/NVulYywK yEKTX24WTJTSUQ/3l0+Q9vXtcbgbPo9v+A4KE= I managed to get both objc4 and CF-lite 550 building under Windows, but have immediately hit a crash on initialization that seems to be related to interactions with the objective C runtime. I'd like to rule out the objc4 runtime library itself, but the supplied test cases do not build under Windows in their current form and seem to use . Before I attempt to generate one myself, I was wondering if an existing (simple) test case might be available? Thanks, -Brent _______________________________________________ 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: carbon throwing objc exceptions
by Eric Schlegel 14 Sep '09

14 Sep '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Sep 14, 2009, at 2:24 PM, Joel Reymont wrote: Sean, On Sep 14, 2009, at 9:57 PM, Shawn Erickson wrote: Quoting from an existing bug [2]: -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… Why are you attempting to catch the except at all? It looks like no attempt is made to deal with the exception and in general exceptions in Cocoa are truly exceptional and imply a mistake by the programmer. "If a cocoa exception is thrown in the middle of a gecko callback, really bad things happen (like you can no longer load any pages, for example). Cocoa exceptions are basically setjmp/longjmp, and, presumably, without a handler, will jump back to the main event loop's default handler. This will unwind the stack without any stack cleanup." Generally, exceptions should be caught at the level of the callback, rather than around the lower-level functions that you're calling in CoreFoundation. As to whether the CF function you're calling could actually throw an exception, I can't say for sure, although it would be surprising. I don't think CoreFoundation throws exceptions in general. However, the source is public; you could download it and determine the answer yourself. This email sent to site_archiver(a)lists.apple.com

1 0