[sane-devel] Bug#597922: Additional Scanner Logs

Olaf Meeuwissen olaf.meeuwissen at avasys.jp
Mon Dec 6 00:44:34 UTC 2010

Hash: SHA1

On 2010-12-05 06:38, Soren Stoutner wrote:
> On Thursday, December 02, 2010 05:37:36 pm Olaf Meeuwissen wrote:
>> On 2010-12-02 20:17, Julien BLACHE wrote:
>> For some reason the device is saying that it received a cancel request
>> (did someone push a cancel button on the device?).  Protocol-wise, the
>> backend is supposed to send a CAN in that case before anything else.
> No cancel request was made, either by pushing a button on the device or by 
> clicking cancel in the scan application.
> I'm attaching three other logs that may be helpful to you.

Next time, please compress the logs.

Unfortunately, in your attempts to trim the logs a bit you chomped the
interesting part.  However, Julien's detective work showed that before
commit 63a8dd4ebc9535b23d7a922cb0e7a9a682edd955 the backend completely
ignored the error code at the end of each image data block.  Obviously,
that'll work just fine if there are no errors so you get a good scan
with the 1.0.20 version.

> scangui.txt
> This is a log using KDE's Image Scanning Application (scangui) with 
> libsane_1.0.21-7.  This scan fails, similar to the one with GIMP whose log was 
> attached to a previous email.  However, when failing there is no err dialog as 
> there was with GIMP and the scanner locks up so that no other scanning can 
> even start until the scanner is power cycled.  (GIMP did not lock up the 
> scanner.)

Not quite sure why the scanner locks up but I guess that has to do with
the way the frontend handles SANE_STATUS_CANCELLED.  Looks like scangui
just calls sane_close whereas with the GIMP I see sane_control_option
calls in the logs.

> Let me know if there is any other debugging output that would be helpful to 
> you.

More logs are probably not useful at this point.  The backend does not
do the right thing protocol-wise and I suggested a fix in a previous
comment.  Once that fix is added, you will still be out of luck because
your scans will be aborted by the backend as soon as it sees a cancel
request (the last byte of the image data block matches 0x10).  I have no
clue as to why it sends that information.  If all is okay you should be
seeing 0x00 there.  As I mentioned before, I have no information on the
0x02 part of that byte here at hand but I'll see if I can dig something up.

For the time being, I suggest you stick with sane-backends 1.0.20 or try
the epkowa backend that comes with Image Scan! for Linux.  See


Hope this helps,
- -- 
Olaf Meeuwissen, LPIC-2           FLOSS Engineer -- AVASYS CORPORATION
FSF Associate Member #1962               Help support software freedom
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/


More information about the sane-devel mailing list