Hi, On Mon, Dec 05, 2005 at 12:00:30AM +0100, Reinhard Mayr aka Czerwinski wrote: > sorry, I did not realize that there is a difference between current CVS > and experimental.
Are you really sure that you use experimental now? :-) Some of your errors occur for me with the normal CVS, but not with experimental. Also, when you get an error, make sure you unplug and replug the scanner when you do another test. Otherwise it may still be confused by the earlier test. Which frontend do you use? Is the scanner connected to a USB2 hub? Just to be sure: xsane 0.98 does have a bug that can cause stack corruption in the backend. While I didn't see this in connection with the genesys backend, it happend with other backends. Try to use xsane 0.98b or apply the bug fix Oliver Rauch posted on this list. I use xsane 0.98b for testing. Details: > gray/8bit: > 75: ok > 150: ok > 300: ok > 600: ok > 1200: ok > 2400: Error during read: error during device I/O. I can avoid this by > enforcing backtracking through system load (e.g. by task switching) No backtracking here. No I/O errors. Image is ok but stretched in Y-direction (as already mentioned in my report). You can enable debugging this way: SANE_DEBUG_GENESYS=255 SANE_DEBUG_GENESYS_GL841=255 xsane 2>log The logs may be to big to be send to this list, however so please try to compress them or place them on a website. The limit is 50 KB. > gray/16bit: > 75: ok > 150: ok > 300: ok > 600: backtracking. horizontal lines I only get backtracking here with high load of the computer. However I can confirm the horizontal lines. Looks like something connected to backtracking at 16 bit doesn't fully work yet. I don't have these lines in 8 bit mode. This also happens with the Canon LiDE 50. Here is a sample image of such a scan with forced backtracking (4.2 MB): http://www.meier-geinitz.de/sane/tmp/lide50/stripes-16bit-gray-0600.png > 1200: backtracking. Error during read: error during device I/O No error here. I get some backtracking and one or several strange horizontal lines when the scanner does backtracking. These lines contain data from the scanned image but are horizontally moved somehow. Data corruption? > 2400: backtracking. horizontal lines - picture is unusable Same here as for 1200 dpi. > color/8bit: > Segmentation fault at any resolution Works for me. Backtracking in high resolutions, but no stripes. Please create a log file and run the frontend in gdb and do a backtrace there (bt) when the error occurs. > color/16bit: > 600: Failed to start scanner: invalid argument > other resolutions: segmentation fault Works for me. No stripes even when backtracking with the Canon LiDE 60. However, I do get some stripes with the Canon LiDE 50. So there is a backtracking problem in 16 bit mode (3.5 MB): http://www.meier-geinitz.de/sane/tmp/lide50/stripes-16bit-color-0300.png Bye, Henning
