Scan from ADF doesn't know when to stop

[Cody A.W. Somerville]

If I select Scan "All pages from feeder", it'll scan all the pages in the feeder and then loop trying to scan another page that doesn't exist (ie. it doesn't know when to stop).

I don't have this problem with xsane.

Ubuntu: 9.04 (Karmic)
Printer: HP OfficeJet J6480 All-in-One.
Simple Scan: 0.8.1-3

simple-scan.log:
** (simple-scan:23274): DEBUG: sane_read (212)
** (simple-scan:23274): DEBUG: sane_read (212)
** (simple-scan:23274): DEBUG: sane_read (212)
** (simple-scan:23274): DEBUG: sane_read (212)
** (simple-scan:23274): DEBUG: sane_start (page=5, pass=0)
** (simple-scan:23274): DEBUG: sane_get_devices ()
** (simple-scan:23274): DEBUG: Saving device cache
** (simple-scan:23274): DEBUG: Storing device 'hpaio:/net/Officejet_J6400_series?ip=192.168.1.175' in cache
** (simple-scan:23274): DEBUG: Stopping scan thread
** (simple-scan:23274): DEBUG: sane_exit ()

[Robert Ancell (robert-ancell)]

Could you please try the bzr version and attach the full log of simple-scan.log? I have added additional debugging that will hopefully indicate if/when SANE indicated the last page had been reached.

[Cody A.W. Somerville (cody-somerville)]

Please find requested log attached.

[Robert Ancell (robert-ancell)]

bzr trunk updated to automatically choose the ADF source

[Robert Ancell (robert-ancell)]

Cody, where did you get your scanner driver? It doesn't appear to be a standard SANE driver. Also, does it now work in Simple Scan >= 0.9? It seems to have a boolean option "batch-scan" which may need to be set to make it go into batch mode.

[Robert Hrovat (robi-hipnos)]

I've hoticed the same problem using HP 5590 scanner with ADF. It scans first and second while displaying page 1. After scanning third page, it displays it as no. 2. Using latest version from ppa archive.

[Robert Ancell (robert-ancell)]

Robert, can you attach a log of this occurring? (~/.cache/simple-scan/simple-scan.log)

[Robert Hrovat (robi-hipnos)]

File attached. I tried to scan 4 pages and simple-scan showed only first and last.

[Robert Ancell (robert-ancell)]

Robert, did you load 4 pieces of paper or two? SANE has only reported two pages were scanned.

But the interesting thing is SANE reports you have two ADF sources "ADF" and "ADF Duplex". When you scan with xsane do you select ADF Duplex? Does that scan both sides of the paper?

[Robert Ancell (robert-ancell)]

(Simple Scan chose ADF as the source)

[Robert Hrovat (robi-hipnos)]

I loaded 4 pages, just like in GScan2PDF which I use on daily basis.

And yes, I selected ADF as source. This scanner does not load from ADF even if it has something loaded and ADF is not selected.

Dos this program use scanimage or scanadf?

[Robert Ancell (robert-ancell)]

No, it uses the SANE API directly. Could you run the following from the command line and attach the output:
$ ltrace scanadf 2>&1 | grep sane_

[Robert Hrovat (robi-hipnos)]

Attaching ltrace output.

I checked simple-scan.log and I've noticed that simple-scan doesn't get anything from scanner while log outputs lines
DEBUG: sane_read (107) -> (SANE_STATUS_GOOD, 107)

There's plenty of them, but during that scanner is still reading pages.

[Ronald McCollam (fader)]

I can reproduce this issue on a Brother scanner, but not on an HP scanner.

The Brother scanner:
Brother MFC-7440N
Connected over network (ethernet)

The HP scanner:
HP ScanJet 6300C (with attached ScanJet ADF document feeder)
Connected via USB

The HP scanner behaves as expected, stopping after the last page (or not returning any data if no pages are loaded). The Brother scanner continues to scan indefinitely, or returns an endless supply of empty pages if no pages are loaded.

I will attach debug logs from both scanners.

[Robert Ancell (robert-ancell)]

Robert - I think you're problem is fixed by the change in bug 542870, please test with bzr trunk.

[Robert Ancell (robert-ancell)]

This bug has got a bit out of control with the number of issues being tracked... I'm going to close it for clarity; if you can reproduce any problem that was not resolved in 0.9.10 please open a new bug. Thanks for all the great feedback!