[sane-devel] Problem with GT68XX and MacOS X REVISED

[Henning Meier-Geinitz]

Hi,

please reply also to the sane-dvel mailing list, not only to me.

On Tue, Nov 22, 2005 at 07:10:21PM +0000, Nick Collingridge wrote:
> > There are some issues with the A3 USB in 1.0.16 which have been fixed
> > in the current development version of sane-backends. But as far as I
> > know, none of these issues were related to detection of the scanner.
> > Trying a current development snapshot may be woth a try anyway.
> 
> I can find what I think is a source version of this, but not a compiled
> binary.

Yes. Only source code is provided by the SANE project (and some links
to binaries, but people tend to create binaries only for stable
versions).

> >> I have installed all the above items and made sure that there is a properly
> >> configured (according to the instructions, including an override and a
> >> vendor and model number) GT68xx.conf file in /usr/local/etc/ on my boot
> >> drive. 
> > 
> > This scanner does not need an override. Vendor and product ids are
> > already in gt68xx.conf.
> 
> Sorry - attempting belt and braces when other solutions didn't work. Still
> confused on this front by how the backend will know which scanner name to
> use unless the conf file tells it, though...

The USB ids and respective scanner names are hard-coded. This override
crap is only needed if scanners have the same USB ids but different
hardware. In theory, this should never happen but Mustek and others
did it. So in these cases the backend has to be told which scanner is
actually connected. For most scanners (including yours) that's not
necessary.

> > Please send the output of
> > export SANE_DEBUG_GT68XX=255 (or Mac OS X equivalent)
> > export SANE_DEBUG_SANEI_USB=255
> > scanimage -L
> 
> I've performed the two export commands but I don't know where the output
> files get put - I've tried searching for them but I don't know what they're
> called even! If you can give me a clue of where the files might be I'll
> happily send them to you.

The log messages will be pronted to stderr so you should see them
on the terminal window. export is for sh/bash, if you use another
shell you may need setenv instead. To save the output to a file, use
scanimage -L 2>logfile.txt

> > You are not expected to power-cycle the scanner each time :-)
> > This doesn't happen on Linux or any of the BSDs and I also haven't
> > gotten a report about such behaviour on Mac OS X as far as I remember.
> 
> Maybe not, but this is the only way to get the scanner recognised
> successfully.

I just wanted to confirm that this is not intended or normal behaviour
of the gt68xx backend.

> >> I am afraid that all the man pages and the FAQ for TWAIN SANE are not very
> >> complete in their description of what needs to be done to get everything
> >> working properly, and also don¹t really help very well when things don¹t
> >> work.
> > 
> > man sane contains some trouble-shooting notes.
> 
> But it's all very vague and written very much from the persective of use
> with Linux - lost of mentions of unloading kernel drivers, mounting file

The sane manpage does not mention kernel drivers at all.

> systems and suchlike that are very much on the technical end of the scale
> and just don't help with a MacOS X problem!

Some MacOS X specific information is in README.darwin but it's
intended more for building SANE than using.

> I know that this is an
> open-source project and all done by volunteers, so I'm not really
> complaining - just pointing out that if you do get a problem installing on
> OS X, the Linux-focused nature of the man pages is a bit of a hindrance to
> getting things going by yourself!

Please send patches for Mac OS X specific tips, especially for man
sane-usb and man sane-scsi as these are really OS-specific.

> I am pretty technically competent with MacOS X and Windows but even
> so there is a shortfall between my knowledge coupled with what I can
> work out from the documentation and where I need to be to get this
> going.

Usually special questions/trouble like yours is best discussed on the
mailing list.

> I believe that it was SANE-backends 1.0.14 and libusb 0.1.8

Maybe a problem with libusb? IIRC there used to exist several issues
with libusb on Mac OS X but I thought they were fixed recently.

Bye,
  Henning