It would not be compatible with the Ninja Tap but I think it is possible to do something simpler and faster to manage four joysticks.
Has we say in French: « the best is the enemy of good ». ^^
The Ninja Tap exists, the games that support it exist too.
Even if the peripheral could be better, it's not that bad.
Anyway, I wouldn't see the point of making another one from scratch.
In addition to the original Ninja Tap (I don't know how many are still in circulation) there should be new ones produced by Danjovic and an MSX Village user.
With the original driver available and the support in MSXgl I hope that more games will take advantage of this in the future.
@aoineko, Ive been browsing the messages at MSXVillage. I have seen that my detection routine works with CX5MII and more systems, where the original drivers did not worked.
That is because somewhere along the development of the DM system 2, the detection code was changed from original and the developers haven't realized the "short circuit" problem because by chance the original interface used a chip capable of surpassing the internal ports of the computer they have at the time.
if you look closer at the DM System 2 GETNIN routine, you will notice that the DE register is loaded with 0BA3Ah to keep pin 7 at "input" state. That part of code remained the same, I mean, mostly uncommented from what we believe is the original drivers.
If you change the bytes in MSXgl drivers then they will work on all MSX machines with the original schematics, and the ACT chip won't be critical anymore. Don't you think it is worth to try to make a test program with the bytes changed ans test it on real hardware?
Oh, and one more thing, I should not be producing the interfaces to sell. On the other hand, the shinobi tap, as usual to all my projects, is fully open which means that anyone can build their own ans even build it and sell to other people, as long as the credits to the original project are maintained.
Regarding compatibility it's more a software issue that can be solved with a patch. After all it was a patch to this game that motivated me to bump this sleepy Ninja Tap topic
If you change the bytes in MSXgl drivers then they will work on all MSX machines with the original schematics, and the ACT chip won't be critical anymore. Don't you think it is worth to try to make a test program with the bytes changed ans test it on real hardware?
I switched to something else (the V9990 module) while waiting for a build of the latest version of OpenMSX with the (expected) fixes.
But, if you are want, I can send you a version tonight.
But, if you are want, I can send you a version tonight.
I have tested already a patched version of your driver with the ninja tap original schematic and it worked fine even with other chip families than ACT.
The detection routine of the DM System 2 drivers has a bug, and there's no need for you to keep this bug on your MSXgl drivers, don't you agree?
Can you please ask the guys of MSX village to test the modified MSXgl driver in the cx5mII ? It should work fine on all machines too. The worst that can happen is that you'll end up with a driver that is compatible with more machines and requires no special chip family on the circuit, lol!!!
(edit: Also the fixed driver won't require a workaround on the function MSXPSG.cc of openMSX )
If you change the bytes in MSXgl drivers then they will work on all MSX machines with the original schematics, and the ACT chip won't be critical anymore. Don't you think it is worth to try to make a test program with the bytes changed ans test it on real hardware?
I switched to something else (the V9990 module) while waiting for a build of the latest version of OpenMSX with the (expected) fixes.
But, if you are want, I can send you a version tonight.
Builds are usually ready with 20 minutes after the commit on GitHub and may take a few hours on vampire's site.
