Re: [hpoj-devel] Sane owner?

[<pa...@rc...>]

Obi-Wan wrote:
> OK, I duplicated it.  It happens after I successfully scan one photo.
> When I want to scan another photo at a different orientation (horz vs
> vert), I hit the magnifier-undo, then the magnifier-minus (zoom out),
> then hit "acquire preview".  It scans virtually the entire image, then
> locks up xsane.  After a few minutes, the PSC makes a noise, then prints
> the LCD message "Power system failure.  Turn power off then on again."
> Still no response from xsane.  No cron jobs this time.
So are you saying that this particular procedure causes the problem every
time?  Also, what version of xsane are you currently running?

> The stack trace
> isn't very useful, since it wasn't compiled with -g.
ptal-mlcd normally is compiled with -g.  Did you take that out of the
Makefile?

> Power cycling the PSC seems to be the only thing necessary to fix
> the problem.  I didn't have to restart xsane or ptal-mlcd, and it
> started working again.
> 
> Upon revival, the rulers are off again.  The numbers along the top
> are exactly double the numbers along the left.
...
> Are these ruler numbers dependent on info reported by the device?
> Is this what happens when it reports a different vertical vs horizontal
> resolution, for instance?
It would be a good idea to restart xsane if you have to power cycle the
peripheral.  When a SANE backend device is opened, the backend probes the
device for various things and sets some values in the device.  It's quite
likely that the ruler spacing is based on some of these one-time settings,
and power cycling the device pulls the rug out from under the software.

David