I really hope the two of you could patch things up and work together more in the future.
+1
As I've said, it doesn't matter ... and I guess that is a big part of the whole misunderstanding ...
Xiangfu can support it, or even get ckolivas himself to support it.
Having the hardware is necessary to do proper support.
By support I mean: advancing the code, fixing bugs and answering questions.
... just like I do with the Icarus code ... and the MMQ and BFL FPGA code ... coz I have all 3.
... and I don't do with the ztex code, coz I don't have one ...
Xiangfu will have an Avalon soon, but I'd certainly go as far as suggesting that giving ckolivas one and thus getting his support would be even better for Avalon and it's customers.
(and I will repeat again what I said ... I don't want one)
The ASIC software environment will be very interesting ...
ckolivas has never had any FPGAs so he has, to a large extent, stayed away from most of the FPGA driver code.
He has of course done most of the changes not directly part of the FPGA driver code, to the main cgminer code that all devices use.
He also did the one change to the BFL FPGA code that was the biggest performance gain since it was first written
(I had earlier done the equivalent change to the Icarus code)
As I have said quite a few times and also for a very long time, the initial choice for the FPGA code of serial-USB was problematic and in general a bad choice.
I finally got around to doing something about that at the end of last year by rewriting the serial-USB part and using direct USB, that all future USB code will be using by whoever writes the drivers (I've also converted both the MMQ and the BFL FPGA code over to it)
But regarding the GPL, as I've said from the beginning, the cgminer code already exists and according to Avalon, existed complete on the 20th Jan.
There will not be any ground breaking code in there. I can tell you up front.
ckolivas is now working on the main change to cgminer itself for ASIC, handling devices with MCU queues
I believe Xiangfu did a somewhat related change to support sending ~24 items to the Avalon all at once, but having ckolivas do it in the main code (as I wanted it done) and then having an ASIC use all the new USB code will be very interesting indeed, most so also due to the fact that ckolivas himself will be completely involved with the direct ASIC code due to having ASIC devices.
I am indeed looking forward to what happens with cgminer with this new code and both of us working on it
... and of course anyone else who fonts up and provides good quality code and ongoing support for whatever other devices we don't have the hardware for.