[sane-devel] Re: Sane-devel digest, Vol 1 #107 - 8 msgs

[Henning Meier-Geinitz]

Hi,

On Thu, Oct 31, 2002 at 11:19:57PM +1100, Glenn Barry wrote:
> > Thanks to everyone that answered my query about the saned daemon.
> 
> To Clarify, my Indy only has an 8 bit graphics card, i.e 256 colours,
> not 8 bit as one would normally assume for each channel, 8 bits
> total.

I know, I've used such a device for quite some time until the hard
disc stopped working. 

> I have verified that the 16 bit scanning works fine running xsane
> locally, but not however over the network, I get all sorts of amazing
> colours something like the rainbow effects of oil on water. I don't
> scan a lot of film and negatives, which is best with 16 bit colour
> depth, but I will have to use the Indy in future when doing so, until
> the bug can be fixed.

Is the Indy big-endian? It may be a problem with swapping bytes.
According to the SANE-standard, it's the job of the net backend to
swap bytes if necessary. Maybe this code doesn't work in the Windows
version.

Can you try to connect to the Indy with a Linux (or other Unix)
computer?

> BTW if I don't use the -d switch then XSane on Win32 is unable to
> find a device, I am however not troubled by this.

The -d switch of saned? Sorry, but using this switch just can't work,
if you use inetd.

With earlier versions of SANE, there were bugs in the saned code that
resulted in race conditions. Sometimes saned crashed in inetd but not
in debug mode. But these bugs should have been fixed in 1.0.8.
maybe there is still another bug, but currently you seem to be the
only one experiencing this problem.

If you put "localhost" in net.conf on the Indy and run saned by inetd,
is the scanner found by scanimage? Over the "net" I mean:
net:localhost:umax:/dev/something?

Bye,
  Henning