[sane-devel] Erratic problems with saned/xsane
Wed, 11 Dec 2002 13:40:02 -0300
Don't think I've forgotten about the problem. It's just that we had
a few more pressing issues. Another reason is that for some reason
the erratic behaviour has almost disappeared. The 'invalid parameter'
messages seem to appear only when I'm not around ;-)
I did several tests yesterday and all seemed correct - change modes,
resolutions, etc. Still, the other day, one of the volunteers phoned
me to say they had to reboot the machine (Windoze wise ;-) to get
the scanner working. I'll have another go at testing in a few days.
On Sun, 8 Dec 2002 00:28:28 +0100
Henning Meier-Geinitz <firstname.lastname@example.org> wrote:
> On Sat, Dec 07, 2002 at 12:14:37PM -0300, John Coppens wrote:
> > > I have never tried that but you could play with permissions of the
> > > epson.conf file and/or libsane-epson.so.version. Setting them so only
> > > saned can load the backend may work.
> > Ok. I'll try that this afternoon - the PC is at the Foundation, so I
> > can't do it right now. Could the double access (xsane -> parallel port,
> > and saned -> parallel port) have something to do with the erratic
> > behaviour?
> I don't think so because you don't have access to the parport as
> normal user anywway.
> But to avoid disturbance from other backends, comment out everything
> but net and epson in dll.conf.
> > > Another way would be to link xsane to the net backend directly,
> > > without using libsane-dll.
> > Don't know how to do that... But I'll have a look. It would be easier
> > to use afterwards, I think.
> For testing: LD_PRELOAD=/usr/local/lib/sane/libsane-net.so xsane &
> Otherwise when you run make in the xsane distribution, look for the
> final linking command
> > For example: Xsane comes up in Binary Mode (1bit), I try to change to
> > color or gray and get the 'invalid parameters' error. Yes, I can still
> > give the order to scan, for example. And even though the error was
> > signalled, the menu now shows Color or Gray. The scanner scans Binary.
> > Sometimes, it's the scan command that fails with the same message. It's
> > as if the backend doesn't relay the capabilities reliably to the xsane
> > frontend. Could this be a reliability issue for the comm thru the parallel
> > port? Or are these exchanges checksummed & tested?
> I don't think so because you would see the errors wwith xsane as root
> in this case, too.
> Either it's really a saned/net issue or a problem of the epson backend
> that only occurs in conncetion with saned/net. The latter thing has
> occured some times, saned/net is a bug-finder :-)
> > > > All versions are the last ones - fresly downloaded. Kernel is 2.4.18,
> > > > and the scanner is an Epson ActionScanner II (GT-5000M). All points
> > > > to unreliable comm between xsane and saned...
> > >
> > > The communication flow is like this:
> > >
> > > xsane <- linking -> libsane-dll <- dynamic loading -> libsane-net <- network
> > > network -> saned <- linking -> libsane-dll <- dynamic loading -> libsane-epson
> > No wonder someone gets confused ;-)
> > BTW, I'm not scanning over a network. I use 'localhost' here.
> That's network, too :-) net connects to saned over tcp. From the SANE
> point of view it doesn't matter if it's localhost.
> > Will do all that - but I guess it will be long. Shall I send it
> > personal?
> If they are longer than 10 kb (compressed), send them to me
> Sane-devel mailing list