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

[Henning Meier-Geinitz]

Hi,

On Fri, Nov 25, 2005 at 10:02:03AM +0000, Nick Collingridge wrote:
> Thanks for the reply, but I¹m afraid that I¹m going to have to give up on
> getting SANE to work on MacOS X. It seems like it¹s going to take a while to
> get it going, and unlike most people who have the time to fiddle around with
> linux configuration problems (seemingly!) I need to get some work done! The

Well, it seems to work on Linux but not on Mac OS X. So it looks like
Mac OS X users have more time to fiddle around with configuration,
otherwise the system would work out of the box :-)

> latest problem is that the scanner is now reporting itself, apparently
> incorrectly, as vendor 0x04f9, product 0x011b instead of vendor 0x055f,

0x04f9 is Brother. Do you have a Brother USB printer?

> product 0x0210. This problem occurred when I first connected the scanner
> then rectified itself, but it has now recurred.

If that's not the case, I guess something is completely broken with
your USB system.

> I have successfully run the debug export which has produced reams of output,
> but unless you really want it I¹m not sure that it will be of great value to
> you!

I wouldn't have asked for the output if I didn't want to read. it.
Please send it to me personally (not the list).

> As an example of the linux-specific nature of the man pages, here¹s a
> snippet from the ³USB configuration tips² (sane-usb (5)), which is referred
> to a number of times in other man pages:

As already mentioned, man pages like sane-scsi and sane-usb describe
low-level problems of the interfaces. Obviously, they must go into
details of operating systems as that's the whole idea of these manual
pages.

>        Libusb  can  only access your scanner if it's not claimed by the kernel
>        scanner driver. If you want to use libusb,  unload  the  kernel driver

This is true in principle for all operating systems. At least it
should be because communication with the same device by different
application at the same time may cause lots of trouble. 

It's at least valid for Linux, OpenBSD, NetBSD and FreeBSD as there
exist kernel scanner drivers.

>        (e.g. rmmod scanner under Linux) or disable the driver when compiling a
>        new kernel.

One example is given (for Linux), the other way (compiling) is the
generic way for all other unixes using scanner drivers.

> For Linux, your kernel needs support for the USB filesystem
>        (usbfs). For kernels older than 2.4.19, replace "usbfs" with
> "usbdevfs"
>        because the name has changed. This filesystem must be  mounted.
> That's
>        done  automatically  at  boot  time, if /etc/fstab contains a line
> like
>        this:

Obviously, this problem only effects Linux. Therefore it has been
mentioned, that it only refers to Linux. If there were general usage
problems with Mac OS X, they would be mentioned here. At least if one
of the MacOS X users cared to send a patch.

> Incidentally the man page for the GT68XX backend says:
> 
>        Currently scanning seems  to  only  work  reliably  under  Linux. With
>        FreeBSD,  NetBSD and OpenBSD scanning works only once. Then the scanner
>        isn't detected anymore and has to be replugged.

Actually, it doesn't say that anymore in the development version. The
reason is that meanwhile the *BSD bugs have been fixed and I have
several positive reports from other operating systems.

> I totally applaud the work of you and the rest of the people working on
> SANE, but my impression is that it needs a lot of work yet to get it to a
> stage where it can be used by anyone other than geeks with time to waste.

I can't comment on what needs to be done on Mac OS X because I don't
use this operating system. For Linux and the *BSDs there are only two
things to be done: place the firmware on the file system and adjust
permissions. For most backends even a firmware is not necessary.

> And I say that as a pretty technically competent person ­ I¹ve never had to
> just give up on something like this before. I wish you all the best of luck
> with getting this valuable project to a usable state!

In Usenet I would have answered to such a comment with a *PLONK*

> For now I¹m moving the scanner to a Windows box where the Mustek drivers
> will enable me to perform the scans that I need. It pains me to do so, but I
> suppose that¹s why I have it sitting in the corner!

Do whatever you want. My impression is that you just want to complain.
If you had used your time for looking at the debug logs or at least
sending them to me instead of ranting maybe the reason for your
problem may have already been discovered. I'm rather sure that it's
something related to MacOS X or libusb anyway but you never know.

Bye,
  Henning