02990: joemacr, joemacra: [debug] Access Violation

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
02990	Crash/Freeze	Critical (emulator)	Always	Feb 28, 2009, 11:32	Feb 25, 2010, 21:37

Tester	Firewave	View Status	Public	Platform	MAME (Self-compiled)
Assigned To		Resolution	Unable to reproduce	OS
Status [?]	Closed			Driver	dataeast/simpl156.cpp
Version	0.129u5	Fixed in Version		Build	Debug
		Fixed in Git Commit		Github Pull Request #

Summary	02990: joemacr, joemacra: [debug] Access Violation
Description	----------------------------------------------------- Exception at EIP=004FBB9A (deco16_tilemap_34_combine_draw+0x022c): ACCESS VIOLATION While attempting to read memory at 146F4112 ----------------------------------------------------- EAX=146F4112 EBX=00000068 ECX=0000BAFC EDX=00000000 ESI=00009481 EDI=0000BAFC EBP=0022FAA8 ESP=0022FA40
Steps To Reproduce
Additional Information
Github Commit

Flags	64-bit specific, Debug build specific
Regression Version	0.129u5
Affected Sets / Systems	joemacr, joemacra

Attached Files

No.03854 Fujix Administrator Feb 28, 2009, 16:31	No repro in a normal build, I added the debug flag.
No.03856 Phil Bennett Developer Feb 28, 2009, 17:21	Doesn't repro in a 32-bit debug build. 64-bit only?
No.03857 Tafoid Administrator Feb 28, 2009, 17:27	Must be - my freshly made debug doesn't crash either, 32-bit
No.03862 Firewave Senior Tester Mar 1, 2009, 13:43	I wasn't able to reproduce it outside of a regtest run. Also no problems on linux and in valgrind (it's a 32-bit system though).
No.03863 couriersud Developer Mar 1, 2009, 13:55	Unable to reproduce on linux 64 bit.
No.05760 etabeta Developer Feb 25, 2010, 12:36	the weirdest thing of all is that simpl156.c does not use deco16_tilemap_34_combine_draw therefore, I cannot see how this driver could produce the reported error. However, in next MAME update, the deco 16bit video chips have been converted to be devices, hence I suggest to give a last try to reproduce this. if it is still impossible to reproduce it, we should close the report.
No.05762 Haze Senior Tester Feb 25, 2010, 15:50	these simply crash error messages almost never give the actual crash function, that's why only a proper backtrace is useful.
No.05767 Tafoid Administrator Feb 25, 2010, 21:37	Current testing of SVN (pre-0.136u4) in 64-bit debug build in Windows is not causing a crash. Closing.