[gscan2pdf-help] Looking for beta testers for the gscan2pdf rewrite, scantpaper v3.0.0-rc1

[Jeff <jf...@po...>]

After something like four years of work, I've finally got the gscan2pdf 
rewrite to a stage where people can test it. I've been using it for over 
a year, but I'm sure there are plenty of bugs to find.

I've given the rewrite a new name - scantpaper. The UI is basically 
identical to gscan2pdf, but I have completely rewritten it in Python. 
"scant" (https://en.wiktionary.org/wiki/scant) in this sense means 
"short (of)", as I am trying to digitalise my paperwork, and I liked the 
play on "scan".

Improvements compared to gscan2pdf:

* scantpaper uses OCRmyPDF as the backend to create PDFs. These means we 
get PDF/A out of the box.
* scantpaper uses an SQLite database to store all session data. This 
eliminates the need for gscan2pdf's big hairy temporary directory, and 
does not suffer from gscan2pdf's requirement to hold an open file handle 
for each page, which meant that some users ran out of file handles when 
they created several hundred pages.
* Storing sessions in a database also means that the number of undo/redo 
steps is only limited by available storage. gscan2pdf can undo/redo one 
step.

Known issues:

* OCRmyPDF does not currently support non-ASCII characters. This is fine 
for English and most Western European languages, but the hidden text 
layer is probably going to break for non-Latin scripts, e.g. Chinese, 
Japanese, Cyrillic. I intend to help the author of OCRmyPDF fix this as 
soon as possible.
* The Python bindings for SANE don't seem to support pressing (scanner) 
buttons in software. But I don't think that feature was widely used. I 
think I know how to fix it, I just didn't have the tuits up to now, as 
my scanners don't have software buttons.
* Some of the dependencies (at least python3-iso639) are currently only 
provided by Debian unstable, and are not yet in Ubuntu-latest. Until 
then, they can be installed manually from 
https://packages.debian.org/sid/amd64/python3-iso639

Roadmap:

* Package the dependencies for the PPA
* Fix enough of the bugs people find to get a release out
* Add support for non-Latin scripts
* Deprecate gscan2pdf
* Migrate to Gtk4.

Thanks to:

* Chris Mayo, who has already found lots of bugs, and provided fixes for 
many of them.

scantpaper is hosted on https://github.com/carygravel/scantpaper

The CI/CD pipeline (Github Action) builds a Debian/Ubuntu/Mint package 
every time code is pushed, and of course for every release.

https://github.com/carygravel/scantpaper/releases

Please report bugs as Github issues: 
https://github.com/carygravel/scantpaper/issues

If you for some reason really can't use Github issues, I'll will try to 
keep up with stuff reported on this mailing list, but bugs are more 
likely to get lost here.

I'm happy to accept pull requests, especially for workflows to create 
packages for non-Debian systems, e.g. RPM, NixOS, Arch, etc. Just note 
that I can't test the resulting packages; so I would ask the submitters 
to maintain them.

Regards

Jeff