Thanks a lot for your testing. So, apparently, the PC/SC behavior is different between GNU/Linux and Windows.
Thus, I pushed another change: rG1e27c0e04cd3: scd: More fix with PC/SC for Windows.. Please test this. (Both of previous version and this version work well on GNU/Linux for operations not including suspend/resume with Yubikey and Gnuk Token, while my Yubikey with PC/SC doesn't work well for suspend/resume.)

Thanks, this version of scdaemon executes.

I think the problem is more that NSIS uses this arcane build system which makes it hard to cross compile.
NSIS 3.0 is also not in experimental.

About Debian: Stable releases are only updated for bug fixes and not for new features. This is an important for almost all production systems. Rolling release distros do not provide a platform which can be used to replicate use cases or problems.

there is only NSIS 2.51 in debian

Cool! Thanks for testing :-)

Thanks for the help,

I still don't have any clue what could be wrong here.

I can't reproduce this. I sent myself a Mail with capitalized "Andre.Heinecke@intevation.de" while my key only has an identity for "andre.heinecke@intevation.de" and it worked as expected:

Mh, that is strange and indeed a bug if that is so. GpgOL should do some simple normalisation which should prevent exactly such a problem. I'll look into it.

Question has been answered. Closing this.

Thanks for the hints.
The problem for us is that we want to rely on Debian Stable for building Gpg4win and there is only NSIS 2.51 in debian :-/ Maybe we make an exception and package NSIS 3 ourself for debian.

With 3.1 ( https://www.gpg4win.org/version3.1.html ) the problem should be gone. We still have to block outlook when the inline editor is used but that left no artifacts in the past. And if a Mail Window is opened we do not block outlook anymore. We only disable it to show a modal dialog.

At some point we really need to look at better error handling so that such an error would be more visible in the UI.

As this is easy to test and I tested it myself I think I can mark it as resolved.

Thanks for your help / report.

With Gpg4win 3.1.0 ( https://files.gpg4win.org/Beta/gpg4win-3.1.0-beta-current.exe ) GpgOL no longer uses Kleopatra for signing. So this problem can no longer exist.

I'm lowering the priority to Normal. I've done a lot of GpgOL work and Testing for the upcoming 3.1.0 release and have not seen this problem.

Leaving this open until we have a new version of GnuPG in the installer. While Kleopatra should no longer crash it won't properly work without the patch to GnuPG.

Got confirmation In Bugs.kde.org that this is fixed https://bugs.kde.org/show_bug.cgi?id=389792 as my tests also showed this -> resolved.

We have this now. There might be bugs but in general this works.

Sorry, my build was not good even if it's for x86_64 (I used development version of libassuan, etc.).

I realized that: once KDF-DO is written to smartcard/token, factory-reset command won't work because it assumes standard PIN format than hashed.

Sorry again. My script was still wrong (didn't work).

Probably you are right but I don't know Windows internals that much.

I installed 3.1.0-beta and tested all use cases, everything is working properly now.

A Beta of 3.1.0 is now available:
https://files.gpg4win.org/Beta/gpg4win-3.1.0-beta-current.exe

Should work with the current beta: https://files.gpg4win.org/Beta/gpg4win-3.1.0-beta-current.exe Although the window mangement is still a bit "iffy" but we at least switch back the focus.

Yes sorry, I decided against a release specially for that is it is not super critical, no data loss.

A beta of the next version where this is fixed is available now: https://files.gpg4win.org/Beta/gpg4win-3.1.0-beta-current.exe

I've uploaded a beta for the upcoming 3.1.0 Version: https://files.gpg4win.org/Beta/gpg4win-3.1.0-beta-current.exe

Version 3.1.0 now supports sign & encrypt and sign only for G Suite accounts as long as there are no attachments on the mail.

I wonder if this also works similar in a multi user system:

It doesn't work because I did mistake for the salt of reset code, it should be 8-byte instead of 4-byte.
Here is a fixed version, which I tested with Gnuk 1.2.8: