MSX trivial dropbox, the suite of applications to manage a Dropbox acount from a MSX computer, stopped working on May 2013 due to Dropbox phasing out version 0 of its API. Konamiman has just released MSX trivial dropbox 1.1, a new version of the suite that uses version 1 of the Dropbox API, so network-enabled MSX computers are now Dropbox-enabled again.

However, there is a catch. The new Dropbox API requires mandatory usage of HTTPS, which is something that currently a MSX computer can't handle. Therefore MSX trivial dropbox is no more a standalone application: it requires the help of an intermediate proxy to convert plain HTTP to HTTPS and the other way around (an example of such a proxy is stunnel). The full details are in the included user's manual.

There are good news for the owners of a DenyoNet card too: a new version of the TCP/IP UNAPI BIOS. This version corrects a bug that caused the state of a TCP connection to be reported as "closed" when it was actually "performing ARP resolution", thus often failing to properly establish the connection.

And for those owning a Panasonic FS-A1 GT in addition to a DenyoNet card, the good news are double. The DenyoNet TCP/IP UNAPI BIOS has a bug that causes the computers of this model to hang at boot time. The new BIOS version still has this bug, but a seamless workaround has been made available: an alternative version of the BIOS that does not get initialized at boot time, plus a command line tool that initializes the BIOS. So the strategy for FS-A1 GT + DenyoNet owners is to flash the alternative BIOS in the DenyoNet, and invoke the initializer tool either manually or from AUTOEXEC.BAT. This should made the card perfectly usable while the bug is being investigated.

Last but not least, new versions of the PING and TCPCON tools have been made available. The previous versions crashed the computer (again, on FS-A1 GT machines) when they were about to finish and return to the DOS prompt.

Relevant link: Konamiman's MSX page (more precisely, networking section and DenyoNet section)

Comments (9)

By konamiman

Paragon (1035)

konamiman's picture

01-02-2014, 23:14

By the way, here is a related forum topic.

Someone can tell me what the hell is wrong with the FS-A1 GT? I'm completely puzzled and I don't have one at hand to investigate... :-?

By mars2000you

Enlighted (5488)

mars2000you's picture

01-02-2014, 23:24

Maybe the presence of MSX-MIDI interface ?

By konamiman

Paragon (1035)

konamiman's picture

02-02-2014, 22:32

Well, maybe, I have no idea of what the MIDI interface does but I could do some research about it (perhaps a conflict with the ports used by DenyoNet?). Anyway I find it very strange that the hang happens at boot time, but not when the card is initialized after the system has booted.

Also, the problem does not appear with the Ethernet UNAPI BIOS, only with the TCP/IP UNAPI BIOS. The only singificant difference between them is that the TCP/IP BIOS patches the timer interrupt hook and the Ethernet BIOS does not.

Hum... any ideas?

By mars2000you

Enlighted (5488)

mars2000you's picture

02-02-2014, 23:10

By zeilemaker54

Master (244)

zeilemaker54's picture

05-02-2014, 06:49

Which hook do you use to attach your timer code ? The turbo-r uses a slightly adjusted interrupt service routine.

By konamiman

Paragon (1035)

konamiman's picture

05-02-2014, 16:02

I use H_TIMI (FD9Fh). The source code of the DeyoNet BIOS is available in my web, in case you want to take a look. :)

Anyway this seems to happen on the GT only, not on STs (I have a ST and it works fine).

By zeilemaker54

Master (244)

zeilemaker54's picture

06-02-2014, 19:50

The FS-A1 GT has a slightly longer interrupt routine, handling the interrupts coming from the MIDI hardware. But that has nothing to do with the H.TIMI part of the BIOS. How much time does it take to execute your own H.TIMI code ? H.TIMI is called 50 times a second (PAL) or 60 times a second (NTSC). You should finish within time with your H.TIMI code.
An other possibility is the use of I/O port 028H. Not sure if this is being used on a GT.
Sounds like a diffcult problem...

Reading again about the problem: it is only accuring at boottime ? Very strange indeed!

By zeilemaker54

Master (244)

zeilemaker54's picture

06-02-2014, 20:25

Looking at your code, I find that the routine GETSLTP1 does nothing with the interrupt state, but GETSLTP2 does enable interrupts and the end. I would leave the control of the interrupt state in the INIT code of the ROM and not in the GETSLTP2 routine. Looking at your init code, you better leave interrupts off while initializing.
Not sure this will fix your problem though....

By konamiman

Paragon (1035)

konamiman's picture

16-02-2014, 12:53

Thank you for the suggestions, I'll try to get some time to take a look.