Re: [Bitpim-devel] libusb / Darwin

[Steven P. <n9...@n9...>]

On Dec 15, 2003, at 4:46 PM, Roger Binns wrote:
>>   * I get an exception when importing calendar information.
>>    NotTerminatedException: The value should have been terminated and
>> wasn't
>
> The 4400/6000 can never make their minds up if a string is always
> NULL terminated.  I usually found that it was always null terminated
> if entered on the phone, and not if you used other software to
> enter the details.  Anyway I have now committed a fix so that
> exceptions won't be raised if the description field is not null
> terminated.

  At least this set of data was all entered on the phone. If you can 
deal with it both ways, I think it's best to make the software tolerate 
it.

>>   * When trying to import wallpaper into bitpim, I get a segfault:
>>
> I have no idea what that is about.  The gauge is in the status bar
> at the bottom.

  Again, this only happened if I had the protocol log running. 
Strange... Maybe it's some internal timing issue or threading conflict.

>> However, despite a "completion" of
>> the import, nothing shows up in bitpim for wallpapers.
>
> I am still working on the new code (the damn 6000 has 4 different 
> locations
> to get the wallpaper on so new code code to be written).  I haven't
> looked into too deeply why yet, but if you import them twice it usually
> works.  Also on Windows and Linux you have to resize the main window
> otherwise they aren't drawn.  Again I haven't looked too deeply to
> see why.

  I knew you were busy on this code, so I was reporting only as a "this 
is where things are here", not as a serious bug report. It's looking 
good, though. I found that doing two imports back-to-back got the 
wallpaper in the system, and I did have to resize the main window to 
draw them.

> The ringtone code is disabled.  It will be using the new wallpaper code
> since they work in almost exactly the same way.

  Ah. :-)   That explains it.

  Now, time for me to give some thought to how to attack the libusb 
issue.... Currently the way I have things working is to disable the 
MacOS driver. This means you have no serial interface on 
/dev/cu.usbmodemXXX. This means you cannot use the phone for network 
PPP connections/etc... Probably not a good thing.

  I can get (I believe) a connection to the device with the MacOS driver 
loaded, but only as a "interface" connection, I can not connect to or 
touch the device. This would require quite a bit of change, from an 
initial overview, of how things are handled. In fact, since you connect 
by specifying the vendor/product/interface information, it might even 
have to be an alternate port specification for the Mac in addition to 
usb::. I don't know.... I have to think it through, and get a test 
implementation of the interface-only connection working.

-. ----. -.-- - -.--
Steve Palm - n9...@n9...
-. ----. -.-- - -.--