02322: All Drivers using 8080 CPU: 8080 Core problem with score/credit counts and other variables

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
02322	Core	Major	N/A	Sep 22, 2008, 22:48	Sep 23, 2008, 16:47

Tester	user413	View Status	Public	Platform	MAME (Official Binary)
Assigned To		Resolution	Fixed	OS	Windows XP/Vista 32-bit
Status [?]	Resolved			Driver
Version	0.127u4	Fixed in Version	0.127u5	Build	Normal
		Fixed in Git Commit		Github Pull Request #

Summary	02322: All Drivers using 8080 CPU: 8080 Core problem with score/credit counts and other variables
Description	See pictures below. Invaders: Credits increase by 33 and not decrease by 1 when you start a game. First picture shows before you start (with 1 credit), second picture shows when you start. edit Bug now confirmed, turns out it's an unfortunately trobulematic problem with changes to the 8080 core in .127u3. This has also effected all games on warpwarp.c (ex. Warp & Warp, Gee Bee, Bomb Bee, etc.) Several pictures of romset warpwarp have been added below to show the effections due to the changes. The before pictures are the first three pictures, and the after pictures are the nine other pictures (I'm only mentioning the pictures that show Warp & Warp, not the Space Invaders pictures). I noticed in Space Invaders that the TI SN76477 CSG sound is downpitched by three notes lower than older versions of MAME, that also seems to have been affected by the 8080 core change. The score also acts up in warpwarp.c from the 8080 core change, see the Warp & Warp images. Notice: The Warp Warp images are marked with "before" and "after" names on their file name then the number of the image uploaded (ex. before1, before2, after1, after2, etc.) The after4 image shows the green bars in one of it's "Warp" colors, it's fortunately not a bug at all, I was supposed to take an image where the bars say "Warp" but I accidentaly took the image at the wrong time! I originally intended to take the bars in their "Warp" frame, but I took the bars in their non-green/non-Warp forms! Oops, worry me for taking the image at the wrong time!
Steps To Reproduce
Additional Information	Visible effects noted in many drivers including: 8080bw.c mw8080bw.c n8080.c warpwarp.c and many not listed..
Github Commit

Flags
Regression Version	0.127u3
Affected Sets / Systems	All Drivers using 8080 CPU

Attached Files	invadersinsert.png (925 bytes) Sep 22, 2008, 22:49
	invadersstart.png (820 bytes) Sep 22, 2008, 22:49
	before1.png (1,672 bytes) Sep 23, 2008, 00:12
	before2.png (1,934 bytes) Sep 23, 2008, 00:13
	before3.png (2,060 bytes) Sep 23, 2008, 00:13
	after1.png (2,038 bytes) Sep 23, 2008, 00:13
	after2.png (2,049 bytes) Sep 23, 2008, 00:14
	after3.png (2,075 bytes) Sep 23, 2008, 00:14
	after4.png (2,080 bytes) Sep 23, 2008, 00:15
	after5.png (2,173 bytes) Sep 23, 2008, 00:15
	after6.png (2,552 bytes) Sep 23, 2008, 00:22
	after7.png (2,300 bytes) Sep 23, 2008, 00:22
	after8.png (2,305 bytes) Sep 23, 2008, 00:22
	after9.png (1,792 bytes) Sep 23, 2008, 00:23

No.02535 Tafoid Administrator Sep 22, 2008, 23:11	Seems to be a greater problem with changes to the 8080 emulation core in 0.127u3 (looks like a MESS import). Anything prior to that is fine. Confirmed and added regression version. There is another bug I'm readying which involves WARPWARP.c (pretty much every game in there) with score oddness and also credit behavior that is described here.
No.02536 Tafoid Administrator Sep 23, 2008, 00:47	Gnomick, Sure - this bug started as an invaders only - but the problem is visible in most sets in most drivers using an 8080 CPU. DO NOT EVER re-edit or DELETE items that any Moderator or Developer has taken time to add or correct. I will once again correct the fields to reflect the proper issue. As a Moderator here, it's my job to make sure the bug descriptions and all parts of the bug are correct and impart the correct information for the developers. Since this is a widespread bug and multiple drivers are effected by probably a simple CPU core change - we will keep this in one bug which you created. Also, no more snapshots are needed at this time - please do not keep adding those. The point was made with the invaders snapshots already and no more are needed. I'll be once again changing this description again to have the bug cover all issues with the 8080 CPU core run games. While we here appreciate your enthusiasm in reporting bugs, you need to be able to respond and comply with Moderator requests as well as follow site guidelines. Your inability to even recognize the direction that Fujix and I are giving you is making it harder and harder to let you continue to post new bug reports or respond to other bug reports. Slow down and follow the rules, please.
No.02537 user413 Sep 23, 2008, 00:51	OK! Will do!
No.02540 micko Developer Sep 23, 2008, 05:46	Hi, since I did that change in u3 I have done some testing myself and found what part of update was a problem. in i8085.c there are lines : /* For 8080 NF flag is not used but bit is always set */ if(!I.cputype ) { I.AF.b.l = I.AF.b.l \| NF; } Please try to remove them recompile and try your tests. I have tried invaders and now they work ok. Hope this removal should fix all issues you have found.
No.02541 Gyrovision Tester Sep 23, 2008, 07:19	Just to add, mlander and lrescue - fuel numbers turn to random garbage as you use the thruster and the zero in "CREDIT 0" is garbled in both games.
No.02542 Haze Senior Tester Sep 23, 2008, 08:46	I'm somewhat concerned over what source of information was used to make these (obviously broken) changes in the first place. Doesn't look very reliable.
No.02543 micko Developer Sep 23, 2008, 10:21	Haze: It seams that this bit is always set only on russian clone version of processor, since on those machines it returns always this bit set. So perhaps I should add new sub processor type for this one.
No.02544 micko Developer Sep 23, 2008, 11:27	Gyrovision : I can confirm that after removing code I have mentioned, those two drivers work fine. Will send patch to Aarron for this.
No.02554 Tafoid Administrator Sep 23, 2008, 16:47	This change corrects all issues I noted when testing - I'll mark as resolved for next version. Thanks for the quick fix, Micko