gscan2pdf-help Mailing List for gscan2pdf

Addendum:

If I select to scan only the BACK side of the page, the scanner scans the page, but scantpage does nothing with it - it's not displayed and the mouse pointer turns into a wheel, as if waiting for something, and the program remains unresponsive. The only thing to do at this point is close it.

I wonder WHAT broke when I installed the 3.0.5 version.

Installed xsane, to see if there's something wrong with the Sane backend, but it works well - scans duplex, scans the front of the pages, the back, no problem.

Thanks





On Tuesday, April 21, 2026 at 04:59:52 PM EDT, Charles Manson via gscan2pdf-help <gsc...@li...> wrote: 





Hi,

I was able to uninstall ver. 3.0.3 and installed 3.0.5.

It has the same problems as 3.0.3:

- OCR is still performed regardless of the setting (I have it off).

- Still can't delete user-defined tools (edit/preferences). Running from the terminal, the following error appears when I try:
TypeError: PreferencesDialog._add_user_defined_tool_entry.<locals>.delete_udt() takes 0 positional arguments but 1 was given

- File's date is still 4 hours behind when saving a scanned document.

- There are a number of buttons with the wrong icon showing on the toolbar (the icon that shows is the circle with a diagonal line).


In addition to the above:

- Can't scan in duplex mode. The back side of the pages doesn't seem to be scanned.

- After all sheets are scanned, the following shows up in terminal:
ERROR:frontend.image_sane:Error running process 'scan_page': Document feeder out of documents

- After having installed 3.0.5, neither 3.0.3 nor 3.0.4 are able to scan duplex. I had 3.0.3 and it was able to do duplex without issues before I installed 3.0.5.

I'm attaching the log to this email.

This is beyond frustrating.


Other observations:

- Tried installing from the repository, but that installs 3.0.0 (Linux Mint 21).

- Tried installing the deb package and it complains aboout ocrmypdf and doesn't install. Message: Error: Dependency is not satisfiable: ocrmypdf (>= 15.4.3).
(I have ver. 13.4.0+dfsg-1)


Thanks!

_______________________________________________
gscan2pdf-help mailing list
gsc...@li...
https://lists.sourceforge.net/lists/listinfo/gscan2pdf-help

 Hi,

I was able to uninstall ver. 3.0.3 and installed 3.0.5.

It has the same problems as 3.0.3:

- OCR is still performed regardless of the setting (I have it off).

- Still can't delete user-defined tools (edit/preferences). Running from the terminal, the following error appears when I try:
TypeError: PreferencesDialog._add_user_defined_tool_entry.<locals>.delete_udt() takes 0 positional arguments but 1 was given

- File's date is still 4 hours behind when saving a scanned document.

- There are a number of buttons with the wrong icon showing on the toolbar (the icon that shows is the circle with a diagonal line).


In addition to the above:

- Can't scan in duplex mode. The back side of the pages doesn't seem to be scanned.

- After all sheets are scanned, the following shows up in terminal:
ERROR:frontend.image_sane:Error running process 'scan_page': Document feeder out of documents

- After having installed 3.0.5, neither 3.0.3 nor 3.0.4 are able to scan duplex. I had 3.0.3 and it was able to do duplex without issues before I installed 3.0.5.

I'm attaching the log to this email.

This is beyond frustrating.


Other observations:

- Tried installing from the repository, but that installs 3.0.0 (Linux Mint 21).

- Tried installing the deb package and it complains aboout ocrmypdf and doesn't install. Message: Error: Dependency is not satisfiable: ocrmypdf (>= 15.4.3).
(I have ver. 13.4.0+dfsg-1)


Thanks!

Changes compared to v3.0.4

* Refactor hOCR parsing into thread to improve UI responsiveness
* Batch insertions of words in hidden text layer to improve UI
   responsiveness

Available on Github:

https://github.com/carygravel/scantpaper/releases/tag/v3.0.5

Or via the PPA:

https://launchpad.net/~jeffreyratcliffe/+archive/ubuntu/ppa

Hi Richard,

On 15/04/2026 00:44, Richard Lyons wrote:
>   1 - The program loaded successfully, and located my scanner (Epson
>       all-in-one L4260). Once unmaximised, I had to widen the window to
>       see the very wide toolbar. This would be a disadvantage in normal
>       use as I generally work in gscan2pdf with the three windows beside
>       each other, the main window in portrait format, the scan window and
>       the save dialog.

The above is interesting. For me, the basic UI of scantpaper is 
identical to gscan2pdf. Would you mind giving me a couple of screenshots 
to show me the differences you see?

>   2 - I opened the 'scan document' window with ctl-G as in gscan2pdf. It
>       had, I think, 5 tabs. The right-hand tab showed 0, 0, 210, 297 for
>       A4 by default (escellent!) but as I usually do on gscan2pdf I
>       clicked on 'custom' and changed it to A4. The window immediately
>       shed three of the tabs leaving me with only 'postprocessing' and
>       'page options'. So I could not select colour/greyscale/line, or the
>       dpi. Also the scan button was greyed out.

This is also interesting, because the logic of the scanner handling in 
scantpaper should also be identical to gscan2pdf. The behaviour you are 
describing suggests that your scanner reloads all options when you 
change the page size, and that the reload somehow times out or fails.

Please start scantpaper from the command line with the --log=log option, 
reproduce the problem, quit, and post me the log file, which scantpaper 
should have compressed with .xz

Regards

Jeff

Hi Jeff,

I thought I would give a quick test to see whether scantpaper is already
usable on my (debian 13) box. It turns out not to be. My comments may be
of limited value, but herre they are:

 1 - The program loaded successfully, and located my scanner (Epson
     all-in-one L4260). Once unmaximised, I had to widen the window to
     see the very wide toolbar. This would be a disadvantage in normal
     use as I generally work in gscan2pdf with the three windows beside
     each other, the main window in portrait format, the scan window and
     the save dialog.

 2 - I opened the 'scan document' window with ctl-G as in gscan2pdf. It
     had, I think, 5 tabs. The right-hand tab showed 0, 0, 210, 297 for
     A4 by default (escellent!) but as I usually do on gscan2pdf I
     clicked on 'custom' and changed it to A4. The window immediately
     shed three of the tabs leaving me with only 'postprocessing' and
     'page options'. So I could not select colour/greyscale/line, or the
     dpi. Also the scan button was greyed out.

I closed the scan document window and tried to reopen it with ctl-G, but
nothing happened.  I could not do anything further as a result. 

It makes me realize what a huge mission rewriting this software must be.

I take my hat off to you! ...and look forward to future releases...

Cordially,

richard


-- 
Richard Lyons
Brasil: rua Constante Jardim 35, Santa Teresa, 20240-230, Rio de Janeiro
   whatsap e tel: +55 (21) 97437-2244
Romania: strada Morii 41, Miercurea Nirajului, 547410, Mures
   tel: +40 771 604 557

Changes compared to v3.0.3

* Fix error running OCR from dialog.
   Closes #59 (OCR Fails: `int` object has no attribute `get_size`)
* Eliminate extra tesseract call when saving PDFs.
* Fix message 'Some pages have not been saved' when all pages have been
   saved. Closes #61
* Fix delete button in Edit/Preferences/User-defined tools
* Wrap post-save tool combobox in a scrollbar in case it is too long

Available on Github:

https://github.com/carygravel/scantpaper/releases/tag/v3.0.4

Or via the PPA:

https://launchpad.net/~jeffreyratcliffe/+archive/ubuntu/ppa

Hi Charles,

On 05/04/2026 20:41, Charles Manson via gscan2pdf-help wrote:
> Then I had a dependency issue - pip install scantpaper-x.x.x-py3-none-any.whl doesn't install dependencies, despite the instructions stating otherwise - so I went through the list on git and had whatever was missing installed with synaptic package manager.
> 
> Once that finally worked, typing scantpaper on the terminal wouldn't do anything, so I spent about another half hour looking for it and found it in ~/.local/bin and was able to run it.

pip installs the Python dependencies, but not system libraries. I've 
added more notes to the README to make it clear what is necessary, plus 
your point about where the executable ends up. Thanks for the feedback.

> Scantpaper does OCR even when that option is NOT selected (I don't want OCR).

I've fixed that in the upcoming release.

> The save dialog is extremely wide. This is due to one of the entries in Preferences / General Options / User-defined tools is quite long. The save dialog's minimum width is tied to the length of the longest user-defined tool command.

I've fixed that by wrapping it in a scrollbar.

> User-defined tools can't be deleted once added. You can edit and delete their content, but the "Delete" buttons that appear to their right in the preferences dialog don't do anything.

I've fixed the delete button

> Output files' time is off. The files' timestamp is 4 hours behind. Saved files around 6:30 pm NY time and they show up with correct Created and Accessed times (6:30 pm), but Modified time is 2:30 pm. Work-around: currently using this as a post-save user-defined action:
> 
> touch -m --no-create %i
In Edit/Preferences/General Options, does unsetting "set access and 
modification times to metadata date" achieve what you want?

Regards

Jeff

Hi Walter,

On 04/04/2026 21:29, Walter Ahrend via gscan2pdf-help wrote:
> And the post processing regarding page turning does not work either, instead of turning pages by 90 degr. it does -90 degr. (270 degr.)

I tested the post-processing rotation and concluded it was broken. I've 
now fixed this in the repo.

I'm sure that a mathematician would argue that anticlockwise angles are 
positive. However I admit that the UI would be better if it was clearer.

I've added an extra label to make it obvious.

Regards

Jeff

Hi Jeff,

Took me quite a bit to get it to finally install. First I had an issue with pdfminer.six when trying to get pip installed. The solution was to remove pdfminer.six and reinstall it.

Without the above step, python wouldn't even build the package.

Then I had a dependency issue - pip install scantpaper-x.x.x-py3-none-any.whl doesn't install dependencies, despite the instructions stating otherwise - so I went through the list on git and had whatever was missing installed with synaptic package manager.

Once that finally worked, typing scantpaper on the terminal wouldn't do anything, so I spent about another half hour looking for it and found it in ~/.local/bin and was able to run it.

Total time by someone who's only played around with python (YEARS ago) and was just following instructions blindly and googling solutions: 2 hours.

The program works nicely except for a few things.


Bugs:

Scantpaper does OCR even when that option is NOT selected (I don't want OCR).

The save dialog is extremely wide. This is due to one of the entries in Preferences / General Options / User-defined tools is quite long. The save dialog's minimum width is tied to the length of the longest user-defined tool command.

User-defined tools can't be deleted once added. You can edit and delete their content, but the "Delete" buttons that appear to their right in the preferences dialog don't do anything.

Output files' time is off. The files' timestamp is 4 hours behind. Saved files around 6:30 pm NY time and they show up with correct Created and Accessed times (6:30 pm), but Modified time is 2:30 pm. Work-around: currently using this as a post-save user-defined action:

touch -m --no-create %i

(One side effect of the touch command is that the files no longer have a Creation date - only modified and accessed.)

And interestingly, now gscan2pdf, which until yesterday saved files with correct timestamps, now is saving them with today's date, but 12:00 am as the time. I wonder what I broke, because I dread having to go into TimeShift and having to reboot, since I have so many things running, not to mention that if I did that, scantpaper will likely break for me.

Some of the icons on the toolbar are missing. See attachment.

Also, if you or anyone has any ideas on how to make the icons darker, I'm all ears. (This might be something I did years ago, but messing with screen colors is so tragically complicated in Mint that I never bothered to try. This affects scan2pdf, Thunderbird, and scantpaper.)

Let me know if you'd like me to run specific tests and send you the log.

Thanks and enjoy what's left of the weekend!

Hi fa...@ya...,

I see you've tried to send a message to the list, which it didn't 
forward due to its size.

I assume the problem is the log file. Please either attach it to an 
issue on Github: https://github.com/carygravel/scantpaper/issues or send 
it to my email address, rather than the list.

Regards

Jeff

Hi Walter,

Thanks for the report.

On 04/04/2026 21:29, Walter Ahrend via gscan2pdf-help wrote:
> this ADF-problem with the last page not being scanned also appears with my EPSON GT-20000 scanner under UBUNTU 24.04
> 
> Looks like the ADF reports an empty paper feed after he pulls in the last page and then scantpaper stops without scanning the page and not even throwing out the last page from the ADF…

Please start scantpaper from the the cli with the --log=log option, 
reproduce the problem, quit and post the log file, and I'll take a look.

> And the post processing regarding page turning does not work either, instead of turning pages by 90 degr. it does -90 degr. (270 degr.)

I'll test that. Thanks.

Regards

Jeff

Thanks for the report.

On 04/04/2026 20:54, daltrey wrote:
> 1. When scanning from flatbed, scantpaper repeatedly scans the page and 
> I can't stop the program without killing it. This does not happen if I 
> run the scanner via skanlite, for instance.
> 
> 2. Also, a problem I have with gscan2pdf on this scanner as well -- If 
> set to color, I will have a b/w scan. BUT, if I switch to grayscale and 
> then immediately switch back to color, I will get a color scan.
> 
> 3. It seems that the last page in an ADF scan never appears in 
> scantpaper. It does not seem to be a miss-feed problem or anything, as I 
> will get a series of pages and the last page feeds through the scanner 
> properly, but does not appear on the screen.

Please start scantpaper from the the cli with the --log=log option, 
reproduce the problem, quit, and post the log file. I'll take a look.

Regards

Jeff

Jeff,

this ADF-problem with the last page not being scanned also appears with my EPSON GT-20000 scanner under UBUNTU 24.04

Looks like the ADF reports an empty paper feed after he pulls in the last page and then scantpaper stops without scanning the page and not even throwing out the last page from the ADF…

And the post processing regarding page turning does not work either, instead of turning pages by 90 degr. it does -90 degr. (270 degr.)

All of this works well with gscan2pdf…

BR,
Walter

> Am 04.04.2026 um 21:14 schrieb daltrey <da...@da...>:
> 
> Thank you.
> 
> Report from OpenSUSE Tumbleweed > Epson ET-4800
> 
> 1. I appreciate that I can unzip the tar.gz file and run it from the directory. (I did not have to install any further dependencies beyond my last report re v3.0.1.)
> 
> 2. setup.py works, but as installed, several icons are missing (such as "scanner" icon). There are placeholders and the alt-text on hover. Scantpaper works, though.  If I run from the scantpaper directory where I extracted the .tar.gz file (python3 scantpaper/app.py), then the icons are available. After setup.py install, /user/local/bin/scantpaper then works to run scantpaper (or just "scantpaper", since /user/local/bin is in the path).
> 
> 3. The .desktop file doesn't work for me, might be a permissions issue. I can make my own .desktop file and it works fine.
> 
> 4. scantpaper can't find language files (but I don't need or want them anyway). It may be looking for the general linux language files? I removed those with bleachbit to save space.
> 
> Possibly scanner specific:
> 
> 1. When scanning from flatbed, scantpaper repeatedly scans the page and I can't stop the program without killing it. This does not happen if I run the scanner via skanlite, for instance.
> 
> 2. Also, a problem I have with gscan2pdf on this scanner as well -- If set to color, I will have a b/w scan. BUT, if I switch to grayscale and then immediately switch back to color, I will get a color scan.
> 
> 3. It seems that the last page in an ADF scan never appears in scantpaper. It does not seem to be a miss-feed problem or anything, as I will get a series of pages and the last page feeds through the scanner properly, but does not appear on the screen.
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
>> On 4/4/2026 7:08 AM, Jeff wrote:
>> Changes compared to v3.0.2
>> * + `Error during device I/O` to errors triggering "Error opening the
>>   last device used" dialog
>> * Question caching device list including libusb devices
>> * Fix race conditions when new actions triggered before previous
>>   finished. Closes #55 (Save error)
>> Available on Github:
>> https://github.com/carygravel/scantpaper/releases/tag/v3.0.3
>> Or via the PPA:
>> https://launchpad.net/~jeffreyratcliffe/+archive/ubuntu/ppa
>> _______________________________________________
>> gscan2pdf-help mailing list
>> gsc...@li...
>> https://lists.sourceforge.net/lists/listinfo/gscan2pdf-help
> 
> 
> 
> _______________________________________________
> gscan2pdf-help mailing list
> gsc...@li...
> https://lists.sourceforge.net/lists/listinfo/gscan2pdf-help

Thank you.

Report from OpenSUSE Tumbleweed > Epson ET-4800

1. I appreciate that I can unzip the tar.gz file and run it from the 
directory. (I did not have to install any further dependencies beyond my 
last report re v3.0.1.)

2. setup.py works, but as installed, several icons are missing (such as 
"scanner" icon). There are placeholders and the alt-text on hover. 
Scantpaper works, though.  If I run from the scantpaper directory where 
I extracted the .tar.gz file (python3 scantpaper/app.py), then the icons 
are available. After setup.py install, /user/local/bin/scantpaper then 
works to run scantpaper (or just "scantpaper", since /user/local/bin is 
in the path).

3. The .desktop file doesn't work for me, might be a permissions issue. 
I can make my own .desktop file and it works fine.

4. scantpaper can't find language files (but I don't need or want them 
anyway). It may be looking for the general linux language files? I 
removed those with bleachbit to save space.

Possibly scanner specific:

1. When scanning from flatbed, scantpaper repeatedly scans the page and 
I can't stop the program without killing it. This does not happen if I 
run the scanner via skanlite, for instance.

2. Also, a problem I have with gscan2pdf on this scanner as well -- If 
set to color, I will have a b/w scan. BUT, if I switch to grayscale and 
then immediately switch back to color, I will get a color scan.

3. It seems that the last page in an ADF scan never appears in 
scantpaper. It does not seem to be a miss-feed problem or anything, as I 
will get a series of pages and the last page feeds through the scanner 
properly, but does not appear on the screen.










On 4/4/2026 7:08 AM, Jeff wrote:
> Changes compared to v3.0.2
> 
> * + `Error during device I/O` to errors triggering "Error opening the
>    last device used" dialog
> * Question caching device list including libusb devices
> * Fix race conditions when new actions triggered before previous
>    finished. Closes #55 (Save error)
> 
> Available on Github:
> 
> https://github.com/carygravel/scantpaper/releases/tag/v3.0.3
> 
> Or via the PPA:
> 
> https://launchpad.net/~jeffreyratcliffe/+archive/ubuntu/ppa
> 
> 
> 
> _______________________________________________
> gscan2pdf-help mailing list
> gsc...@li...
> https://lists.sourceforge.net/lists/listinfo/gscan2pdf-help

Hi Jeff,

I will give it a try and get back to you.

Charlie Millar

On 4/4/26 8:10 AM, Jeff wrote:
> Hi Charles,
> 
> On 04/04/2026 05:27, Charles Manson via gscan2pdf-help wrote:
>> I've been using gscan2pdf for years and am currently using version 
>> 2.13.2 under linux mint 21 with a Fujitsu fi-6140z scanner.
>>
>> I've noticed that when I scan a page and save it using G4 compression, 
>> the overhead inside the pdf is excessive (see link at bottom).
>>
>> The pdf is 152.8 KB, while the only image it contains is only 83.7 KB.
> 
> We've been working on scantpaper, a successor to gscan2pdf, for the last 
> few years, which I think would also solve your file size problem.
> 
> Because of this, I'm not planning to work on gscan2pdf unless there is 
> really good reason.
> 
> Unfortunately, as the first releases were this year, and Mint 21 is 
> based on Ubuntu Focal (20.04), which isn't supported by the Ubuntu's PPA 
> or Github any more, it isn't going to be easy to create Debian packages 
> that will work for you.
> 
> Your best option is to download the wheel file from https://github.com/ 
> carygravel/scantpaper/actions/runs/23977881944/artifacts/6270515847 and 
> install it with
> 
>   pip install scantpaper-3.0.3-py3-none-any.whl
> 
> and then start the application with
> 
>   scantpaper
> 
> or to get some logging output:
> 
>   scantpaper --log=log
> 
> Please give it a try and let us know how you get on!
> 
> Regards
> 
> Jeff
> 
> 
> 
> _______________________________________________
> gscan2pdf-help mailing list
> gsc...@li...
> https://lists.sourceforge.net/lists/listinfo/gscan2pdf-help

Changes compared to v3.0.2

* + `Error during device I/O` to errors triggering "Error opening the
   last device used" dialog
* Question caching device list including libusb devices
* Fix race conditions when new actions triggered before previous
   finished. Closes #55 (Save error)

Available on Github:

https://github.com/carygravel/scantpaper/releases/tag/v3.0.3

Or via the PPA:

https://launchpad.net/~jeffreyratcliffe/+archive/ubuntu/ppa

 Hi,
Ah! There's something new! You should indicate this in https://gscan2pdf.sourceforge.net/
After a few excruciating hours going back and forth with chatgpt and my own attempts (because its responses, for some reason had small, stupid mistakes - i% instead of %i, etc.), I came up with a solution using the data-save hook.
Add this in the hook and save files as tiff g4:
sh -c 'in="$1"; convert "$in" -compress Group4 /tmp/tmp-g4.tif && tiff2pdf -o "${in%.*}.pdf" /tmp/tmp-g4.tif && rm -f /tmp/tmp-g4.tif "$in" && xdg-open "${in%.*}.pdf"' -- %i
You will need to have imagemagick installed. It takes the resulting tiff file, saves it in /tmp with G4 compression (I'm not sure this step is needed now that I've slept), converts that tiff to pdf, deletes the tiff saved by gscan2pdf, and opens the pdf in the default viewer.
I'll give gscantpdf a shot next week and report back if I run into any issues. 

Thanks again!!!!
    On Saturday, April 4, 2026 at 08:11:22 AM EDT, Jeff <jf...@po...> wrote:  
 
 Hi Charles,

On 04/04/2026 05:27, Charles Manson via gscan2pdf-help wrote:
> I've been using gscan2pdf for years and am currently using version 
> 2.13.2 under linux mint 21 with a Fujitsu fi-6140z scanner.
> 
> I've noticed that when I scan a page and save it using G4 compression, 
> the overhead inside the pdf is excessive (see link at bottom).
> 
> The pdf is 152.8 KB, while the only image it contains is only 83.7 KB.

We've been working on scantpaper, a successor to gscan2pdf, for the last 
few years, which I think would also solve your file size problem.

Because of this, I'm not planning to work on gscan2pdf unless there is 
really good reason.

Unfortunately, as the first releases were this year, and Mint 21 is 
based on Ubuntu Focal (20.04), which isn't supported by the Ubuntu's PPA 
or Github any more, it isn't going to be easy to create Debian packages 
that will work for you.

Your best option is to download the wheel file from 
https://github.com/carygravel/scantpaper/actions/runs/23977881944/artifacts/6270515847 
and install it with

  pip install scantpaper-3.0.3-py3-none-any.whl

and then start the application with

  scantpaper

or to get some logging output:

  scantpaper --log=log

Please give it a try and let us know how you get on!

Regards

Jeff

_______________________________________________
gscan2pdf-help mailing list
gsc...@li...
https://lists.sourceforge.net/lists/listinfo/gscan2pdf-help
  

Hi Charles,

On 04/04/2026 05:27, Charles Manson via gscan2pdf-help wrote:
> I've been using gscan2pdf for years and am currently using version 
> 2.13.2 under linux mint 21 with a Fujitsu fi-6140z scanner.
> 
> I've noticed that when I scan a page and save it using G4 compression, 
> the overhead inside the pdf is excessive (see link at bottom).
> 
> The pdf is 152.8 KB, while the only image it contains is only 83.7 KB.

We've been working on scantpaper, a successor to gscan2pdf, for the last 
few years, which I think would also solve your file size problem.

Because of this, I'm not planning to work on gscan2pdf unless there is 
really good reason.

Unfortunately, as the first releases were this year, and Mint 21 is 
based on Ubuntu Focal (20.04), which isn't supported by the Ubuntu's PPA 
or Github any more, it isn't going to be easy to create Debian packages 
that will work for you.

Your best option is to download the wheel file from 
https://github.com/carygravel/scantpaper/actions/runs/23977881944/artifacts/6270515847 
and install it with

  pip install scantpaper-3.0.3-py3-none-any.whl

and then start the application with

  scantpaper

or to get some logging output:

  scantpaper --log=log

Please give it a try and let us know how you get on!

Regards

Jeff

Hi,
First of all, thanks for a great program and thanks for making it free.
I've been using gscan2pdf for years and am currently using version 2.13.2 under linux mint 21 with a Fujitsu fi-6140z scanner.
I've noticed that when I scan a page and save it using G4 compression, the overhead inside the pdf is excessive (see link at bottom).
The pdf is 152.8 KB, while the only image it contains is only 83.7 KB.
Here are the details I was able to gather from the pdf:
$ pdfimages -list out-gscan2pdf.pdf 
page   num  type   width height color comp bpc  enc interp  object ID x-ppi y-ppi size ratio
--------------------------------------------------------------------------------------------
   1     0 image    2488  3284  gray    1   1  ccitt  yes       13  0   300   300 83.7K 8.4%
$ pdfinfo out-gscan2pdf.pdf 
Title:           out fr gscan2pdf
Creator:         gscan2pdf v2.13.2
Producer:        PDF::Builder 3.023 [see https://github.com/PhilterPaper/Perl-PDF-Builder/blob/master/INFO/SUPPORT]
CreationDate:    Fri Apr  3 00:00:00 2026 EDT
ModDate:         Fri Apr  3 00:00:00 2026 EDT
Custom Metadata: no
Metadata Stream: no
Tagged:          no
UserProperties:  no
Suspects:        no
Form:            none
JavaScript:      no
Pages:           1
Encrypted:       no
Page size:       597.12 x 788.16 pts
Page rot:        0
File size:       156489 bytes
Optimized:       no
PDF version:     1.4

 Saving as tiff results in an 83.1 KB file, which roughly half the size of the same file saved as pdf. 

After this test, I opened the TIFF file using both Scandall (Windows sw that came with the scanner) and naps2. Both produced smaller pdfs than gscan2pdf.
You can check all 4 files by visiting https://drive.proton.me/urls/FAZJPVXBPM#80u9PvpK5h2k
I wonder if there's a setting I'm missing or if I should update to the latest version (2.4 on the web site), or if there's anything else I can do to ensure the pdfs created are as small as possible.
Thanks!!!!

Hi Chris,

On 28/03/2026 16:49, Chris Deuchar wrote:
> Hi, Firstly thank you for making this software freely available.
> Next, a bit of history.  I have used gscan2pdf since 2021 on a Ubuntu 
> based machine.  This worked fine until about a year ago when I started 
> getting an error on opening the program and attempting to scan.  This 
> said, "Error opening device: Error during device I/O" or similar.
> I did report this (on Github?) at the time but received no response so 
> switched to another program.  However, late in 2025, after an update, it 
> all started working again - so I was happy :-)

Apologies that your report evidently got lost.

> Scantpaper then arrived and, on opening, I get another  "Error opening 
> device: Error during device I/O" message and can proceed no further.  
> This persists right up to and including today's update.

Please start scantpaper from the command line with the --log=log option, 
reproduce the problem, quit and post me the log file, which scantpaper 
should have compressed with xz. I'll take a look.

Regards

Jeff

Hi, Firstly thank you for making this software freely available.
Next, a bit of history.  I have used gscan2pdf since 2021 on a Ubuntu based
machine.  This worked fine until about a year ago when I started getting an
error on opening the program and attempting to scan.  This said, "Error
opening device: Error during device I/O" or similar.
I did report this (on Github?) at the time but received no response so
switched to another program.  However, late in 2025, after an update, it
all started working again - so I was happy :-)

Scantpaper then arrived and, on opening, I get another  "Error opening
device: Error during device I/O" message and can proceed no further.  This
persists right up to and including today's update.

Worse however, is that gscan2pdf no longer works again - but now I can't
get beyond the scan document dialogue box.  The scan button is greyed out
and the only option is to cancel.

Interestingly, both programs, despite the error stated, *do* report the
correct scanner - ie Epson Perfection 1250.  (Yes I know its old but it
does still work and scans transparencies and slides as well as conventional
documents.)

Please advise!

Chris

Changes compared to v3.0.0

* Fix selection changed callback. Closes #45
    Thanks to Chris Mayo for the patch.
* Fix opening an encrypted PDF. Closes #46
    Thanks to Chris Mayo for the patch.
* Improve session logs. Closes #47
    Thanks to Chris Mayo for the patch.
* Fix image context menu mnemonics. Closes #49
    Thanks to Chris Mayo for the patch.
* Fix missing message when PDF has no images. Closes #50
    Thanks to Chris Mayo for the patch.
* Fix updating image resolution via property dialog.
    Closes #44 (AttributeError when changing resolution)
* Replace deprecated IconSet and IconFactory. Closes #3
    Thanks to Chris Mayo for the patch.
* Use qpdf to encrypt PDFs instead of pdftk. Closes #48
    Thanks to Chris Mayo for the patch.
* Fix messages dialog text width. Closes #52
    Thanks to Chris Mayo for the patch.
* Rename scanner icon to scan. Closes #53
    Thanks to Chris Mayo for the patch.

Available on Github:

https://github.com/carygravel/scantpaper/releases/tag/v3.0.1

Or via the PPA:

https://launchpad.net/~jeffreyratcliffe/+archive/ubuntu/ppa

Hi all,

I got Scantpaper running in my openSUSE Tumbleweed/KDE Plasma install, 
so I thought I would share the results.

Because openSUSE rpm packaging/organization is a lot different from 
Ubuntu/Debian deb systems, it's somewhat difficult to find certain 
packages, such as goocanvas.

I had a lot of dependencies to track down and install, not the least of 
which was pytest and how to run pytest on Scantpaper. (Not that 
difficult, but new to me.) Many pytests hung up on OCRmyPDF until I 
finally got that resolved, then Tesseract and iso639, but finally 
Scantpaper is working. Early pytest runs would die at 29 errors, but the 
errors pointed me towards the missing packages and cleared up when I 
installed them.

This is fairly similar to the process I used in the early bad-old-days 
(for me) of running gscan2pdf in openSUSE/KDE. But luckily quite awhile 
ago enterprising persons incorporated gscan2pdf into openSUSE 
Tumbleweed, making life much easier. (I hope they will find out it is 
time to add Scantpaper.)

1. last pytest results -- 37 failed, 98 passed, 28 skipped, 1 xfailed, 
67 warnings.

2. goocanvas -- very problematic, ultimately it was some version of 
goocanvas 2.0 installed as 3.0?? (That was an AI assist ...) It's in the 
OSS repository, but I did not easily find it in a YAST search until I 
knew what I was looking for.

3. several dependencies were difficult to track down due to the 
difference between openSuse and Ubuntu (deb) packaging and naming 
conventions. Also, the openSuse online software lookup system was not 
working properly (or something), as the repositories referenced in the 
"install manually" option did not exist and the "automatic" option does 
not work in my browser, just opens a text page. I prefer to add the 
appropriate repository so the packages will update. I searched around 
for the referenced repository or something similar and couldn't find it.

However, the link to download the rpm package worked, so I was able to 
retrieve packages that were not in the OSS repository.

4. For OCRmyPDF I used a snap install, which required to log out and log 
back in to get snap added to the PATH -- I could add it manually without 
logging out, but Scantpaper didn't recognize the /snap/bin path until I 
logged out and back in?

5. I had to track down "community" builds of things like iso639 and 
ocrmypdf.py.

I didn't make very good notes, so this isn't a perfect walk-through.

On first run after the pytest listed at 1 above, everything worked 
right, sort of, but the scanner oddly kept rescanning the same page on 
the flatbed over and over and I had a little difficulty interrupting it 
so I could exit scantpaper by closing the window.

a separate issue: I have an anomaly with my Epson ET4800 printer/scanner 
that I also have in Gscan2pdf (I brought it up one time, but never 
followed through with Jeffrey, as I found a "workaround".)  Oddly, if 
scanner is set to "color" when the scan dialog opens, it nevertheless 
scans black and white. However, if I change the dialog to greytone or 
monochrome -- and then immediately change it back to color -- it scan in 
color just fine. Very odd.

For some reason when I used ADF I only got a scan of the first page, 
although the second page is feeding through the scanner. However, I have 
to test this on another scanner -- I don't know if it is a problem with 
scantpaper or a problem with the ET4800. (I have a couple of other MFC 
printers/scanners I can test, but they are in another room, so more 
difficult to test.)

Thanks for all you do Jeffrey! I have tested the various other available 
Linux scan programs, as well as some Windows ones, but nothing else is 
satisfactory.

In fact, on my Windows 11 computer, I nevertheless use Gscan2pdf, as it 
is running in the WSL2 gui and integrates nicely. (And, is less 
problematic than the commercial windows scan program I have.)

Barrington

Hi Martin,

On 19/03/2026 14:58, Martin Herrnkind wrote:
> The first thing I noticed was an error message when launching Scantpaper 
> from the console:
> "Error retrieving scanner options: expected string or bytes-like object, 
> got ‘NoneType’"

Please start scantpaper from the command line with the --log=log option, 
reproduce the above error, quit, and post me the log file, which 
scantpaper should have compressed with xz. I'll take a look.

> Subjectively, Scantpaper seems to me to work more slowly than GScan2pdf 
> when loading files and performing text recognition.

Thanks. Once we have shaken out the important bugs, we can take a look 
at performance issues. It is possible that using SQLite as a file store 
is enough of an overhead to be noticeable.

> When loading JPG files, I receive the following error messages for each 
> file or page (excerpt for illustrative purposes):*
> "WARNING:py.warnings:/usr/lib/python3/dist-packages/gi/ 
> _propertyhelper.py:220: Warning: value "137" of type 'gint' is invalid 
> or out of range for property 'page-number-start' of type 'gint'
>    instance.set_property(self.name, value)

Yes. I've spotted that, too. They are harmless warnings, but annoying. 
I'll get to them at some point soon.

> When using Scantpaper, I receive the following error messages (excerpt 
> for illustrative purposes):*
> Estimating resolution as 765
> ERROR:tools_menu_mixins:Can't display page with uuid 13: page not found

That I hadn't been able to reproduce reliably. I'll do some more testing 
and get back to you.

> When saving, the following error messages appear:*
> (com.github.scantpaper:266239): Gtk-WARNING **: 13:25:07.349: Failed to 
> measure available space: Fehler beim Einlesen der Dateisystem- 
> Information für /home/martin/pCloudDrive: Der Socket ist nicht verbunden
> 
> [ I specified a completely different save location than “/home/martin/ 
> pCloudDrive” ]
> However, the GUI shows: “2257 of 1000 being processed (save_pdf)”
> 
> 
> *The saving process ends without completing, displaying the error message:*
> ERROR:basethread:Error running process 'save_pdf':
> ERROR:session_mixins:Error running 'error' callback for 'save_pdf' process:

Can you give me a log file as above, please.

Regards

Jeff

*Hello Jeff, hello All,*

*I’ve been a big fan of gscan2pdf for many years! Thank you so much, Jeff!

For about three years now, I’ve been using gscan2pdf for historical 
research, particularly when working with file formats created in archives.

gscan2pdf is a great tool for historians!

For example, the German Federal Archives makes digitized documents 
available for download on the web in JPG format. For some archival 
materials, several hundred pages need to be processed; for others (more 
rarely), over a thousand pages.
I import these files into Gscan2pdf, use Tesseract to add an OCR layer, 
and save them in PDF format.
In my experience, Gscan2pdf has trouble processing more than 300 pages. 
That’s why I “split” the jobs into batches of 250 to 300 pages—depending 
on the quality of the original—and then “merge” the individual PDF files 
using a PDF editor.

A second example is data sets that you create yourself on-site at a 
state archive using a microfilm reader. A third example is scan 
collections created on-site in the archives using smartphone apps.

GScan2pdf works well for all these purposes.

As of today, I’ve been testing Scantpaper v3.0.0 on a desktop running 
the Debian 13 operating system. Scantpaper was installed using a 
pre-built deb package.

The first thing I noticed was an error message when launching Scantpaper 
from the console:
"Error retrieving scanner options: expected string or bytes-like object, 
got ‘NoneType’"

Subjectively, Scantpaper seems to me to work more slowly than GScan2pdf 
when loading files and performing text recognition.

When loading JPG files, I receive the following error messages for each 
file or page (excerpt for illustrative purposes):*
"WARNING:py.warnings:/usr/lib/python3/dist-packages/gi/_propertyhelper.py:220: 
Warning: value "137" of type 'gint' is invalid or out of range for 
property 'page-number-start' of type 'gint'
   instance.set_property(self.name, value)

WARNING:py.warnings:/usr/lib/python3/dist-packages/gi/_propertyhelper.py:220: 
Warning: value "138" of type 'gint' is invalid or out of range for 
property 'page-number-start' of type 'gint'
   instance.set_property(self.name, value)

WARNING:py.warnings:/usr/lib/python3/dist-packages/gi/_propertyhelper.py:220: 
Warning: value "139" of type 'gint' is invalid or out of range for 
property 'page-number-start' of type 'gint'
   instance.set_property(self.name, value)"


*During the text recognition step with Tesseract, error messages are 
also “normal” for GScan2pdf, since scanned documents from archives are 
often of very poor quality.
When using Scantpaper, I receive the following error messages (excerpt 
for illustrative purposes):*
Estimating resolution as 765
ERROR:tools_menu_mixins:Can't display page with uuid 13: page not found
Estimating resolution as 110
Estimating resolution as 773
ERROR:tools_menu_mixins:Can't display page with uuid 14: page not found
Estimating resolution as 132
Estimating resolution as 774
Estimating resolution as 136
ERROR:tools_menu_mixins:Can't display page with uuid 15: page not found
Estimating resolution as 780
Estimating resolution as 142
Estimating resolution as 160
ERROR:tools_menu_mixins:Can't display page with uuid 16: page not found
Estimating resolution as 159
Estimating resolution as 341
Estimating resolution as 1125


*However, text recognition seems to have worked on the GUI.


When saving, the following error messages appear:*
(com.github.scantpaper:266239): Gtk-WARNING **: 13:25:07.349: Failed to 
measure available space: Fehler beim Einlesen der 
Dateisystem-Information für /home/martin/pCloudDrive: Der Socket ist 
nicht verbunden

[ I specified a completely different save location than 
“/home/martin/pCloudDrive” ]
However, the GUI shows: “2257 of 1000 being processed (save_pdf)”


*The saving process ends without completing, displaying the error message:*
ERROR:basethread:Error running process 'save_pdf':
ERROR:session_mixins:Error running 'error' callback for 'save_pdf' process:


*Comparison with gscan2pdf, v.2.13.4: When launched from the console, 
the same error as above appears:*
(net.sourceforge.gscan2pdf:274582): Gtk-WARNING **: 13:45:14.772: Failed 
to measure available space: Fehler beim Einlesen der 
Dateisystem-Information für /home/martin/pCloudDrive: Der Socket ist 
nicht verbunden

*However, the files load without any errors and much faster than with 
Scantpaper.*
ERROR - 3d3cd0a0-2aea-4dc5-9e6c-488f13c0b049, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - dada053f-10b6-486b-bd7f-bc3dbc945fcf, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - 98e70d0a-ae6a-4a00-abcf-dfae35307bf2, tesseract, Detected 71 
diacritics
ERROR - 94f1260e-afd6-4b14-97bf-65e3e6df4157, tesseract, Detected 70 
diacritics
ERROR - df423c95-d498-4365-82a9-e597dbe81548, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - 9d691361-c61b-437a-8c97-d2e008c3f40c, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - cfdff534-256c-46fd-a49c-536eaea6878f, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - 88d76958-b801-4cfb-8f42-baf9c207d7bd, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - 1ba5fc3f-92dd-444e-84db-428b58b5d1d0, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - 5ff7b239-bfdd-450c-a29e-55cf2b362a54, tesseract, Detected 386 
diacritics
ERROR - fec6ec58-0af9-434f-8e73-1070540f0bd7, tesseract, Detected 298 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - 08b05dca-9059-44f3-8930-125f2ce3d2bc, tesseract, Detected 185 
diacritics
ERROR - 7b392f03-0e1d-4323-9537-aeb7211c6d61, tesseract, Detected 345 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - 3744de25-4be6-45e8-ac00-333fa09a8a7e, tesseract, Detected 321 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - ea03244b-47cb-4180-bc76-86f0620822c5, tesseract, Detected 589 
diacritics
ERROR - e510a717-0c27-4a19-bc65-c5b5c86d3774, tesseract, Detected 279 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - ce97adc8-2d4a-4971-94fb-d7d6d6f1285e, tesseract, Detected 451 
diacritics
ERROR - 12cb3785-6e7e-45a7-bc75-e7cf7575abb0, tesseract, Detected 364 
diacritics
ERROR - f91aa7e9-5a68-46c2-8f63-0bd9541daeed, tesseract, Detected 212 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - 936951c5-99e1-42da-ac89-11d5b4815a05, tesseract, Detected 184 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - 17d7d17b-cf44-4107-bf6b-9aa29729ec88, tesseract, Detected 160 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - ef542d62-b784-4836-aca4-cb2aea909537, tesseract, Detected 196 
diacritics
ERROR - 051e2b35-9a03-4710-9143-8ba0fab843d6, tesseract, Detected 332 
diacritics
ERROR - 3e2c4220-7f7f-46ea-8ec9-dfab30ac6f2d, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - 263b918c-0a30-4bd2-8d41-a3fa157428e2, tesseract, Detected 12 
diacritics
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - 9bc817f9-45d2-4028-9b7e-1b4a3bcdc19c, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - b478c1ba-932a-4223-b4d2-d0f10ed58292, tesseract, Empty page!!
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []
ERROR - deleting corrupt text layer: []


*When saving, GScan2pdf displays the “usual error messages,” but the 
process completes successfully.*

*
*

*Regards *

*Martin*




Am 11.03.26 um 22:26 schrieb Jeff:
> Changes compared to v3.0.0-rc5
>
> * + minimal en_US translation to prevent warnings. Closes #41
> * Be graceful if the previous current working directory no longer
>   exists. Closes #42
>
> Available on Github:
>
> https://github.com/carygravel/scantpaper/releases/tag/v3.0.0
>
> Or via the PPA:
>
> https://launchpad.net/~jeffreyratcliffe/+archive/ubuntu/ppa