RE: gscan2pdf 1.3.2 from the sources under OpenSuse Tumbleweed
Problem: scanning works, but empty blank pages when saving or printing
I had many problems to get a version running under OpenSuse, but now I could install something from https://software.opensuse.org. I could now also compile the latest version from SF.
gscan2pdf --version reports gscan2pdf 1.3.2
It does scan multiple pages from my older BROTHER MFC-7820N and I do see the multiple thumbnails and the last page is shown fully in the viewing panel.
when trying to select another page (e.g. the first page) from the thumbnails in the left column -> nothing happens. Still the last page is shown in the right window.
When trying to save or to print all pages, or a single page: empty page/s !
WARN - GLib-GObject-WARNING **: The property GtkDialog:has-separator is deprecated and should
n't be used anymore. It will be removed in a future version. at ./gscan2pdf line 4742.
(several lines like the following:)
WARN - unhandled exception in callback:
Image file '/tmp/gscan2pdf-BYmN/qllMRPI4B7.pnm' contains no data at ./gscan2pdf line 47
21.
*** ignoring at ./gscan2pdf line 4742.
Perhaps I overlooked a simple stupid thing?
Pls. assist. Let me know if you need further data.
Can you post the complete log file, please?
see http://dpaste.com/1WSC36S
Please also confirm the contents of lines 4742 and 4721 in bin/gscan2pdf.
see http://i.imgur.com/nlgB8nf.png
Can you try again without the unpaper (clean up) step?
The unpaper output seems to be corrupt.
@Jeffrey, sorry, I do not know what you mean. I cannot find an option in the GUI to deactivate "unpaper" or "clean up".
pls. elaborate.
On the scan dialog, in the "Post-processing" frame, there are two rotate options, and then "Clean up images", which seems to be activated at the moment. Try deactivating and then scan.
Bravo, it now works!!!!
What can I help to fix the (now deactivated, but certainly very useful) options ?
pls. let me know if you want me to test something.
What does
unpaper --version
give from the command line?
Please attach a test input image which produces corrupt output when processed with the Tools/Clean Up menu.
unpaper --version
6.1
Please attach a test input image which produces corrupt output when processed with the Tools/Clean Up menu
All my scanned images are resulting in empty pages. So I don't understand what you mean exactly by "test input image".
UPDATED
All my scanned images are resulting in empty pages, when I use the "unpaper" enhancement option which appears to be buggy as mentioned by you.
So I don't understand what you mean exactly by "test input image".
On 24 June 2015 at 21:17, Wikinaut wikinaut@users.sf.net wrote:
So unpaper 6.1 is corrupting all output. Or something else is going on.
Please make a test scan (with clean-up turned off), save it, and run
unpaper over it on the command line. Then post the input image and the
output from unpaper
uuuh, this was quick.
I saved an image scanned by gscan2pdf (true gray / 300dpi), using lossless-compression (I always use png when saving, I am working in the field of image compression and exactly know the difference between lossy and lossless compression).
See attachment: the result from this step.
Then I applied unpaper to <thisfile>:
May be a problem with libav on my Opensuse Tumbleweed.
gscan2pdf as such is working, unpaper not.
See https://github.com/Flameeyes/unpaper/blob/master/doc/file-formats.md for pixel and file formats.
I tried RGB, TIF, PNG, GIF: unpaper always stops with unsupported pixel format or unsupported format (TIF).
Because gscan2pdf works (when unpaper is deactivated), you may consider to flag this issue as "solved" now.
Workaround for the problem: deactivate unpaper (as mentioned by you).
unpaper only works with pnm, so gscan2pdf automatically converts any image to pnm before passing it to gscan2pdf.
I've compiled unpaper 6.1, and also see your problem. unpaper 0.4.2 works fine.
I will make sure that gscan2pdf traps errors like this better in the next release.
Let me know, if I can help, e.g. in testing. The program as such is great, I like it!
I've now added code to trap errors like this. You'll see it in the next release.
@Jeffrey,
just as a follow-up, no answer required:
running now unpaper version 6.2, gscan2pdf appears to work with it, I am very happy.