07228: srdarwin: Game freeze at the end of 1st stage.

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
07228	Crash/Freeze	Critical (emulation)	Always	Feb 11, 2019, 10:42	Feb 13, 2019, 23:09

Tester	VasiliyFamiliya	View Status	Public	Platform	MAME (Official Binary)
Assigned To		Resolution	Open	OS	Windows 10 (64-bit)
Status [?]	Confirmed			Driver	dataeast/dec8.cpp
Version	0.206	Fixed in Version		Build	64-bit
		Fixed in Git Commit		Github Pull Request #

Summary	07228: srdarwin: Game freeze at the end of 1st stage.
Description	At the place of 1st stage boss encounter, the D letters fills each odd tile string, and game stopls, althought the music is still continues to play.
Steps To Reproduce	Playthrogh the 1st stage till the boss without of any damage taking (using the state saving).
Additional Information
Github Commit

Flags
Regression Version	0.206
Affected Sets / Systems	srdarwin

Attached Files	0000.png (5,280 bytes) Feb 11, 2019, 10:42 Uploaded by VasiliyFamiliya Screenshot
Attached Files	savestate.zip (3,874 bytes) Feb 11, 2019, 10:43 Uploaded by VasiliyFamiliya Save of the state before bug appearance.

No.16125 MetalGod Senior Tester Feb 11, 2019, 22:02 edited on: Feb 12, 2019, 00:37	I've done some testing and found out that the regression is mame 0.206 The bug is not present in mame 0.205 and previous versions.
No.16133 Haze Senior Tester Feb 12, 2019, 15:43 edited on: Feb 12, 2019, 15:44	yeah this is very sensitive to MCU timing / comms, the slightest thing it doesn't like and it will randomly break on any one of the bosses when it wants data. I had to play the thing extensively when the MCU was added to make sure it never failed, I guess somebody altered the code. Tafoid, can you track down the exact revision and point whoever broke it at this report? bit annoying to see this regress after the time put into making sure it didn't crash.
No.16135 Osso Moderator Feb 12, 2019, 19:43	Couldn't get the attached save state to work with a current GIT build. I got to the boss with the invincibility cheat and didn't get the problem. Would it be possible for the original bug reporter to test this on a current GIT build (without cheats, of course)?
No.16143 Haze Senior Tester Feb 13, 2019, 23:09 edited on: Feb 13, 2019, 23:10	it didn't always occur even when it was broken before, but it occurred often enough to be game-breaking usually occurring at least once in a run. I ended up running a stress test sending random inputs to loop the game through over and overagain to make sure it no longer crashed and ran it for some hours last time. I guess it just frustrates me because a lot of work was put into making it work before, and that was just disregarded for what 'seems' better from a coding point of view, even if it doesn't work.