Today I started getting a heap corruption and crash in localspl:
dank@snazzy:~/wine-git/dlls/localspl/tests$ ../../../tools/runtest -P wine -M localspl.dll -T ../../.. -p localspl_test.exe.so localmon.c err:heap:HEAP_ValidateInUseArena Heap 0x110000: bad size 001232a0 for in-use arena 0x123128 err:module:find_forwarded_export module not found for forward 'winspool.drv.EnumPortsW' used by L"C:\windows\system32\spoolss.dll" localmon.c:839: Tests skipped: got 2 localmon.c:982: Tests skipped: got 2 fixme:localspl:localmon_XcvDataPort command not supported: L"" fixme:localspl:localmon_XcvDataPort command not supported: L"monitorui" localmon: 72 tests executed (0 marked as todo, 0 failures), 2 skipped. wine: Unhandled page fault on read access to 0x04455379 at address 0x7ef978f1 (thread 0012), starting debugger... Backtrace: =>1 0x7ef978f1 process_detach+0x5b(bForceDetach=0x1, lpReserved=0x1) [/home/dank/wine-git/dlls/ntdll/loader.c:1074] in ntdll (0x0032fe78) 2 0x7ef9aa77 LdrShutdownProcess+0x6f() [/home/dank/wine-git/dlls/ntdll/loader.c:2144] in ntdll (0x0032fea8) 3 0x7ee35c79 ExitProcess+0x17(status=0x0) [/home/dank/wine-git/dlls/kernel32/process.c:2102] in kernel32 (0x0032fec8) 4 0x7ec8d4b7 in localspl_test (+0xd4b7) (0x0032fef8) 5 0x7ee31e5b start_process+0xe4(arg=0x0) [/home/dank/wine-git/dlls/kernel32/process.c:904] in kernel32 (0x0032ffe8) 0x7ef978f1 process_detach+0x5b [/home/dank/wine-git/dlls/ntdll/loader.c:1074] in ntdll: movl 0x34(%eax),%eax 1074 if ( !(mod->Flags & LDR_PROCESS_ATTACHED) )
I'm not alone; test.winehq.org shows others are getting it today, too, but not yesterday.
Anyone know what's up with that?
On Di, 2008-06-24 at 10:25 -0700, Dan Kegel wrote:
Today I started getting a heap corruption and crash in localspl: ... Anyone know what's up with that?
That was introduced by Huw (added a circular dependance).
Changelog: Revert "winspool.drv: Make sure the printer driver registry entry exists.
On Wed, Jun 25, 2008 at 12:17:36AM +0200, Detlef Riekenberg wrote:
On Di, 2008-06-24 at 10:25 -0700, Dan Kegel wrote:
Today I started getting a heap corruption and crash in localspl: ... Anyone know what's up with that?
That was introduced by Huw (added a circular dependance).
Changelog: Revert "winspool.drv: Make sure the printer driver registry entry exists.
But won't we still get the same problem when a new printer (and its driver) are added? In which case we should fix that rather than hiding the issue.
Huw.
On Wed, Jun 25, 2008 at 11:28:35AM +0100, Huw Davies wrote:
On Wed, Jun 25, 2008 at 12:17:36AM +0200, Detlef Riekenberg wrote:
On Di, 2008-06-24 at 10:25 -0700, Dan Kegel wrote:
Today I started getting a heap corruption and crash in localspl: ... Anyone know what's up with that?
That was introduced by Huw (added a circular dependance).
Changelog: Revert "winspool.drv: Make sure the printer driver registry entry exists.
But won't we still get the same problem when a new printer (and its driver) are added? In which case we should fix that rather than hiding the issue.
Indeed, if you run the localspl test with a yet undiscovered cups printer you get the same problem. I've sent a hack to wine-patches that ensures that winspool.drv is loaded before spoolss in the test. That should be removed once the circular dependency is fixed.
No real app will load localspl before winspool.drv so this'll not hit us in the real world.
Huw.
-
Dan Kegel -
Detlef Riekenberg -
Huw Davies