[sane-devel] Epson2 backend and ADF

Olaf Meeuwissen olaf.meeuwissen at avasys.jp
Mon Jul 29 23:35:30 UTC 2013

Michael Koltan writes:

> Hi there,
> I'm using an Epson GT-10000+ Scanner with an ADF; after upgrading to the 
> latest debian version the epkowa backend I used stopped working,

Which Debian versions?  From squeeze to wheezy?
What iscan and iscan-data package versions do you have installed?

In what way did it stop working?  Were there any indications as to why
it stopped working?  What have you tried to get it to work again?

> [snip]  The problem 
> occurres when scanning duplex: The backside of the last page does not 
> get scanned because the scanner reports that the adf is out of paper 
> after scanning the front side.
> This in itself is not a big problem: I just have to track while scanning 
> duplex whether I am on a front page or a back page and ignore the out of 
> paper status when the next page to scan is a back page.

Looks perfectly fine to me, as long as the device provides the expected
image data.

> My question however is: Does anyone of you know if all epson scanners 
> with adf behave this way? Then I may write a general fix for this 
> problem. Or is this a bug in the GT-10000+ firmware and other scanners 
> report out of paper after scanning the last back page in duplex mode 
> (which would make a lot more sense)? Then I would have to mark this as a 
> specifig GT-10000+ fix. Or is there an additional status flag that 
> offers information if there is still an unscanned backpage inside the 
> scanner?

I just checked the generic protocol specification and there might be a
flag you could use for this.  Have a look at command.h of the *epkowa*
backend.  You're looking for FSG_PAGE_END and FSI_CAP_PED.  The former
is a bit mask for the "error" byte at the end of each image data block
you could check.  The latter is a capability flag that says whether
doing so makes any sense.

FWIW, the epkowa backend already checks this and happily continues
fetching data until the last block has been acquired.  Whether this
actually pertains to the problem you observe I'm not 100% sure of,

Hope this helps,
Olaf Meeuwissen, LPIC-2           FLOSS Engineer -- AVASYS CORPORATION
FSF Associate Member #1962               Help support software freedom

More information about the sane-devel mailing list