Sharp ZQ-290s sync software

[Rein Klazes]

On Sun, 25 Apr 2004 02:40:49 -0400, you wrote:


> trace:comm:PurgeComm handle 0x48, flags 5
> trace:comm:PurgeComm handle 0x48, flags a
> (then a bunch of these ...)
> trace:comm:ClearCommError handle 0x48 cbInQue = 0 cbOutQue = 0
> trace:comm:ClearCommError handle 0x48 cbInQue = 0 cbOutQue = 0

cbInQue and cbOutQue are the number of characters in the in- and output
queues. Are you sure that there was no line with a number <> 0 ?


> (then)
> trace:file:ReadFile 0x48 0x419f6aa0 1 0x4074f3f4 (nil)
> trace:file:ReadFile 0x48 0x419f6aa1 1 0x4074f3f4 (nil)
> trace:file:ReadFile 0x48 0x419f6aa2 1 0x4074f3f4 (nil)
> trace:file:ReadFile 0x48 0x419f6aa3 1 0x4074f3f4 (nil)
> trace:file:ReadFile 0x48 0x419f6aa4 1 0x4074f3f4 (nil)
> trace:file:ReadFile 0x48 0x419f6aa5 1 0x4074f3f4 (nil)

Programs requests a number of characters, one at the time. Unfortunately
you cannot see if the program gets the byte. The +relay debug channel
will provide the return value of ReadFile, and perhaps the "LastError"
value. This debug log is /much/ longer the +file, so be warned.


> trace:comm:PurgeComm handle 0x48, flags a
> (then it does it again, and then)
> trace:comm:EscapeCommFunction handle 0x48, function=6
> trace:comm:EscapeCommFunction CLRDTR
> trace:comm:EscapeCommFunction handle 0x48, function=4
> trace:comm:EscapeCommFunction CLRRTS
> (and afterwards it tries to detect the device on another port...)

Anyway, you see there has been plenty of activity on COM1. 

> 
> I've tried out the device in windows, and used portmon to monitor the
> serial port. The device usually sends out a specific string every time
> it's ready to connect.

That is totally absent from your trace: no WriteFile's. 

> The program then polls the serial port until the
> signal is caught, and then does some more handshaking with the device.
> It would seem from this that the software is not reading the correct id
> string. Anything to do with "fixme:comm:SetupComm insize 3102 outsize
> 3102 unimplemented stub"? 

That message is harmless.

> Is there anyway to monitor what's being sent
> and read from the serial port under linux (like portmon for windows)?

If you compile wine from source and know a bit of C  you may want to put
some additional TRACE statements in the source of ReadFile, WriteFile.

> Could it be a dll problem (though the program should be fairly simple,
> and it should run under windows 98). I didn't see any errors on the
> "file" debug channel, though there were some warnings with 32-bit dlls
> not found.

+file generates plenty of warnings, usually as a result of finding the
location/existence of files.

> I'd really appreciate any help with this...
> Thank you, Jad

A number of programs that use serial communication are not working at
the moment, while some others do. There is a chance you will have to
wait until this problem is fixed.

Rein.
-- 
Rein Klazes
rklazes at xs4all.nl