Red Hat Bugzilla – Bug 36577
Agfa 1236s SCSI + aha152x do scanning in a very strange way.
Last modified: 2007-04-18 12:32:43 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.76 [fr] (X11; U; Linux 2.4.3 i686)
I tried to make a _color_ scan (grayscale worked) with an aha152x and SCSI
AGFA 1236s and output picture seems to be dissassembled..
Steps to Reproduce:
1. modprobe aha152x
2. modprobe sg
3. be sure to have permission on /dev/sg*
4. xsane &
5. Choose scanner -> /dev/sga
6. maybe acquire preview...
7. Scan in color mode..
8. ee output.pnm & (where output.pnm is the pic from the scanner)
Actual Results: You may see what happened by looking at this url :
http://confiote.ath.cx/~anvil/pwetpwet.pnm <-- the bad color scan
http://confiote.ath.cx/~anvil/pwetpwet2.pnm <-- The grayscale scan
Here the color scan is "different".
Expected Results: The color scan should be like the grayscale one.. with
Are you able to try out a new snapscan backend? Take a look at
http://snapscan.sourceforge.net/ to find out about it.
Yeah man.. that rox.. =]
I think you should have a look at
I explain what i've done :
1/ get sane-backends-1.0.4.tar.gz
2/ get snapscan-20010411.tar.gz
3/ tar zxvf sane-backends-1.0.4.tar.gz
4/ ln -s sane-backends-1.0.4 sane-1.0.4
5/ tar zcvf sane-backends-1.0.4.tar.gz sane-backends-1.0.4 sane-1.0.4
6/ I modified slightly the sane.spec from 1.0.3.src.rpm (from SRPMS seawolf iso)
(minor changes.. do diff sane-1.0.3.spec sane.from.me.spec to see)
7/ Put snapscan-20010411.tar.gz and the new sane-backends-1.0.4.tar.gz into
8/ rpm -ba sane.spec
9/ Build was very successful so rpm -Uvh sane-1.0.4-dams.i386.rpm
10/ Differences in xsane :
-> preview can be made in colour mode.
-> colour scanning is great. you could see the result at
http://confiote.ath.cx/~anvil/coincoin.pnm (and compare with old colour scan..;)
it seems to really work now
Thanks a lot for the hint.
I've put packages for pre-1.0.5 in rawhide, if you'd like to test them out.
The 1.0.5 release is quite soon.
I've made some scan with this packages and it seems to work good.
Thank you =]
For me, the bug is definitly resolved.