07050: vectrex: 6522 VIA shift register issue

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
07050	Timing	Minor	Always	Jul 21, 2018, 22:53	Jul 19, 2020, 16:27

Tester	zaxxon	View Status	Public	Platform	MAME (Official Binary)
Assigned To		Resolution	Open	OS
Status [?]	Confirmed			Driver	miltonbradley/vectrex.cpp
Version	0.199	Fixed in Version		Build
		Fixed in Git Commit		Github Pull Request #

Summary	07050: vectrex: 6522 VIA shift register issue
Description	there is currently an issue with the vectrex emulation. this issue was not present in mame version 0.178 (apologies i don't know in which version of mame this issue first appeared). currently in mame the vectrex emulation has a VIA bug, which is due to the shift register being incorrectly shifted every cycle. the shift register should be shifted every two cycles (i have attached the information from the 6522 VIA datasheet showing this). this current emulation inaccuracy results in text shown on the display being displayed as 'compressed'.
Steps To Reproduce
Additional Information
Github Commit

Flags	Verified with Original
Regression Version	0.186
Affected Sets / Systems	vectrex

Attached Files	6522 VIA - Datasheet.png (17,392 bytes) Jul 21, 2018, 22:53 Uploaded by zaxxon
	MAME 0.199.png (3,599 bytes) Jul 21, 2018, 22:56 Uploaded by zaxxon Incorrect
	MAME 0.178.png (3,358 bytes) Jul 21, 2018, 22:56 Uploaded by zaxxon Correct

No.15287 Tafoid Administrator Jul 22, 2018, 00:46	First regressed April 28, 2017: https://github.com/mamedev/mame/pull/2262 https://github.com/mamedev/mame/pull/2262/files Attempted fix May 04, 2017: https://github.com/mamedev/mame/pull/2278 https://github.com/mamedev/mame/pull/2278/commits/eb0dd7370f3c0eb1c759c9808ed410598617fc15 Pretty sure this is the offender. There is a later regression but that was due to main CPU changes, not the VIA
No.17837 Robbbert Senior Tester Jul 19, 2020, 10:03	I think this is fixed? Can someone confirm?
No.17841 Tafoid Administrator Jul 19, 2020, 16:27	It still appears to be happening, just not quite as noticeable as the original before. If you use SHIFT+P to watch the intro lines scroll per frame and compare that to 0.185 and earlier, you'll see jumpiness in the scrolling where lines not 'smoothly' moving, parts of the line disappearing and reappearing next frame.