2000-11-12 15:56:05 +00:00
|
|
|
Last update: Sun Nov 12 16:47:01 CET 2000
|
1999-08-09 18:06:01 +00:00
|
|
|
|
|
|
|
- Compiling SANE aborts with error "virtual memory exhausted" on Slackware
|
|
|
|
linux with gcc. What goes wrong?
|
|
|
|
|
|
|
|
The optimization set by "-O2" makes problems, try which one of the
|
|
|
|
following calls work:
|
|
|
|
|
|
|
|
make CFLAGS="-g -Wall -O1"
|
|
|
|
make CFLAGS="-g -Wall -O"
|
|
|
|
make CFLAGS="-g -Wall"
|
|
|
|
|
|
|
|
- Avoiding damage on Mustek flatbed scanners
|
2000-08-12 15:11:46 +00:00
|
|
|
Most Mustek flatbed scanners have no protection against exceeding
|
1999-08-09 18:06:01 +00:00
|
|
|
the physical scan area height. That is, if a scan is attempted with
|
|
|
|
a height that exceeds the height of the scan surface, the scanner
|
|
|
|
begins making loud noises and the scan mechanism may be damaged.
|
|
|
|
Thus, if you hear such a noise, IMMEDIATELY turn off the scanner.
|
|
|
|
|
|
|
|
Normally, the Mustek backend will ensure that the maximum scan
|
|
|
|
height is not exceeded. However, if your scanner model has not been
|
|
|
|
tested yet, this safety-guard may not work. In such a case,
|
|
|
|
you may want to use scanimage's -y option to gradually determine the
|
|
|
|
exact height of the scan area (making sure to turn off the scanner as
|
|
|
|
soon as it starts making noises). Once you know the proper height,
|
|
|
|
mail the following information to sane-devel@mostang.com:
|
|
|
|
|
2000-08-12 15:11:46 +00:00
|
|
|
- scanner name (labels on the front and back of the scanner)
|
|
|
|
- debug logfile
|
1999-08-09 18:06:01 +00:00
|
|
|
|
2000-08-12 15:11:46 +00:00
|
|
|
To get the debug logfile enter the following:
|
1999-08-09 18:06:01 +00:00
|
|
|
|
2000-08-12 15:11:46 +00:00
|
|
|
SANE_DEBUG_MUSTEK=5 scanimage -L 2>logfile.txt
|
1999-08-09 18:06:01 +00:00
|
|
|
|
|
|
|
The current list of Mustek scanners that are known to work
|
|
|
|
properly can be found in man-page sane-mustek(5).
|
|
|
|
|
|
|
|
- My Linux box used to scan just fine but now it suddenly won't work anymore,
|
|
|
|
what's up?
|
|
|
|
If you increased SG_BIG_BUFF when building SANE (as is recommended
|
|
|
|
by sane-scsi(5) for performance reasons), be careful when upgrading
|
|
|
|
the kernel as that will typically install a new version of
|
|
|
|
/usr/include/scsi/sg.h, with the old, smaller value. This has
|
|
|
|
the effect that SANE _thinks_ the SCSI buffer is big, but since
|
|
|
|
the Linux kernel uses a smaller value, scanning will fail with
|
|
|
|
"out of memory" errors. To fix this problem, increase SG_BIG_BUFF
|
|
|
|
to the old value, rebuild the kernel, then reboot the machine
|
|
|
|
Newer versions of SANE (>1.01) tries to read the current
|
|
|
|
SG_BIG_BUFF from /proc/sys/kernel/sg-big-buff which is
|
|
|
|
available with newer linux kernels (>= v2.2).
|
|
|
|
|
|
|
|
- Missing /usr/include/scsi on Linux systems
|
|
|
|
Older Linux distributions are missing the /usr/include/scsi directory.
|
|
|
|
In such a case, it is necessary to copy the relevant files from
|
|
|
|
the kernel distribution. Normally, the command:
|
|
|
|
|
|
|
|
cp -a /usr/src/linux/include/scsi /usr/include
|
|
|
|
|
|
|
|
should fix this problem.
|
2001-04-22 20:01:30 +00:00
|
|
|
|
|
|
|
- Security problems with pnm
|
|
|
|
If the pnm backend is installed and saned is used to allow users on
|
|
|
|
remote computers to scan on the local machine, pnm files can be read by
|
|
|
|
the remote user. This is limited to the files saned can access (usually
|
|
|
|
it's running as user "sane"). All pnm files can be read if saned runs
|
|
|
|
as root which isn't recommended anyway. The pnm backend is disabled
|
|
|
|
by default. If you want to use it, enable it with configure (see
|
|
|
|
configure --help for details). Be sure that only trusted users can
|
|
|
|
access the pnm backend over saned.
|