On 2019-12-20 04:37, Dr Rainer Woitok wrote:
Bruce and All,
On Tuesday, 2019-12-17 07:07:14 +1000, Bruce Schultz wrote:
...
You could try rebuilding the backends with the following command:
EGIT_OVERRIDE_COMMIT_SANE_PROJECT_BACKENDS=6b7052c4 ACCEPT_KEYWORDS="**"
emerge -av1 =media-gfx/sane-backends-9999
That will install git commit 6b7052c4, which should get you pretty close
to the same version of the backend that you originally reported working
on Ubuntu, which was version 1.0.28+git20190831-xenial0 from the PPA.
Doesn't work either. But that is again 1.0.28, a version I possibly
missed by hair's breadth [1]:
$ scanimage -V
scanimage (sane-backends) 1.0.28-348-g6b7052c4-dirty; backend version 1.0.28
Just for reference, the version string "1.0.28-348-g6b7052c4-dirty" a
git version reference which can be broken down as follows
- 1.0.28 - that is the last git tag referenced before the current
commit (in this case it's the 1.0.28 release)
- 348 - the number of commits since the tagged commit
- g6b7052c4 - 'g' for git; '6b7052c4' is the short version of the
commit hash
- dirty - means that the source tree is modified from the git commit;
in this case, it should be the build patches applied by the ebuild
$ scanimage '--batch=.%03d0.png' --batch-increment 1 --batch-print
--batch-start 1 --device pixma:MX880_new-host --source 'Automatic Document
Feeder' --format png --mode Color --resolution 300 -l 3 -t 0 -x 210 -y 297
Scanning infinity pages, incrementing by 1, numbering from 1
Scanning page 1
Scanned page 1. (scanner status = 5)
Scanning page 2
scanimage: sane_read: Document feeder out of documents
Scanned page 2. (scanner status = 7)
Batch terminated, 2 pages scanned
$
[1]: I've again checked my old Ubuntu installation logs and in particul-
ar compared them to the time stamps of my scanned "*.pdf" files.
Under Ubuntu the last multi-page scan not using the flatbed appar-
ently was on 2019-08-08.
I installed sane-backends 0.1.28+git20190810-xenial0 on 2019-08-17
due to a routine upgrade. Even though there are some "*.pdf" files
scanned thereafter, these were flatbed scans. So I honestly cannot
tell whether or not this version was able to scan from the ADF und-
er Ubuntu. And the sane-backends 1.0.28+git20190831-xenial0 inst-
alled via just another routine upgrade was possibly never ever used
at all.
The version installed when I did my last successful ADF scans under
Ubuntu then has to be 0.1.27+git20190731-xenial0, which I installed
on 2019-08-01.
In that case, the 1.0.28 release was tagged on 31 Jul 2019, so I imagine
that the version you had installed there was effective a pre-release
version of 1.0.28. You've already tested that version on Gentoo with no
success, but it is interesting that your initial tests reported scanner
status 9 (I/O error) versus the later behaviour where it fails with
scanner status 7 (out of documents); I have no idea what that means,
however!
Back on 15-Dec, Olaf's provided the following advice:
> I know about the existence of environment variable "SANE_DEBUG_BJNP" but
> have no idea what value to set it to to provide reasonable information.
> Could someone please guide me?
Just set it to 127 if you don't know. The developers should be able to
slice and dice the result as needed.
I'd suggest that from here, your best way forward is probably to
re-install back to the latest git version and capture the debug logs.
Also, you might want to consider Louis's earlier suggestion of adding
'auto_detection=no' to the config file.
Regards
Bruce
