Succ-U-Bus

[dmitri]

I didn't have a look at the code but it seems strange to me that sending the whole firmware sysex on its own works fine in MCT but sending it after sending 'erase_sysex' doesn't. Surely the size of the firmware sysex is the same in both case so why you don't need to break it up in the first case and then suddenly you have to break it up in the second case.

[elrules]

Well, the key then is in the erase sysex message. Something happens after the sending/reception of this message. Is the MIDI connection reset or broken momentarily? Is the power of the unit interrupted or something? I don't know. Something must happen. That message is the only difference between bootloaders.

[dmitri]

Well, the key then is in the erase sysex message. Something happens after the sending/reception of this message. Is the MIDI connection reset or broken momentarily? Is the power of the unit interrupted or something? I don't know. Something must happen. That message is the only difference between bootloaders.

It doesn't have anything to do with the bootloader! It happens even if you use virtual MIDI ports.

[Ken Forgettable]

Surely the size of the firmware sysex is the same in both case so why you don't need to break it up in the first case and then suddenly you have to break it up in the second case.

MidiPort has no hook on the OS callback handler so it can't know when the data has been read by the OS. This means it doesn't know when it can refill its own buffer safely. I also think the OS callback scheduler timing is a little conservative so individual messages (buffered by MidiPort) can get lost (overwritten) before the callback handler knows anything about them.

We'll see

[elrules]

So, in the end, do you want me to modify the MCT to send the firmware in parts?

Well, you have now the source code... have you tested it yourself?

About the errors while updating... are always talking only about Windows version (using MidiPort)? or doesn't it work even with the Linux version?

[dmitri]

So, in the end, do you want me to modify the MCT to send the firmware in parts?

Well, you have now the source code... have you tested it yourself?

About the errors while updating... are always talking only about Windows version (using MidiPort)? or doesn't it work even with the Linux version?

I didn't try it on Linux. Why won't you try it yourself? You don't even need to have a MegaDrum module connected. Select any MIDI port for output and you'll get the error. If you select a virtual MIDI port as output and then with MIDI-OX (or anything similar on Linux) monitor what is sent to this port you will see that 'erase_sysex' comes through but not a single byte of the firmware sysex does.

[Ken Forgettable]

So, in the end, do you want me to modify the MCT to send the firmware in parts?

Well, you have now the source code... have you tested it yourself?

About the errors while updating... are always talking only about Windows version (using MidiPort)? or doesn't it work even with the Linux version?

I didn't try it on Linux. Why won't you try it yourself? You don't even need to have a MegaDrum module connected. Select any MIDI port for output and you'll get the error. If you select a virtual MIDI port as output and then with MIDI-OX (or anything similar on Linux) monitor what is sent to this port you will see that 'erase_sysex' comes through but not a single byte of the firmware sysex does.

I think the code I posted fixes these problems for windows using MidiPort!

[elrules]

That would be great news!!

I would test it myself but I don´t have a secure bootloader. In fact, as I moved again to amsterdam, I don't even have a finished megadrum hehehe

[dmitri]

That would be great news!!

I would test it myself but I don´t have a secure bootloader. In fact, as I moved again to amsterdam, I don't even have a finished megadrum hehehe

You don't need MegaDrum to test it: http://www.megadrum.info/forums/viewtop ... 9577#p9577

[elrules]

I wanted to see if it works on a real device because for me, all this time, with virtual ports I have been able to send the firmware correctly.