Hmm, calling QProcess::setStandardInputFile(QProcess::nullDevice()) before start() may be even better than calling closeWriteChannel() after start().

On the Cli, gpg does not give an error but an info message informing that the subkey already exists so there is no error message for Kleopatra to act on.

Fix for the crash https://invent.kde.org/pim/kleopatra/-/merge_requests/318

It works in Gpg4win-Beta-75 to add the ADSK to more than one Key.

The QProcesses we start should be configured with closed stdin so that the started process don't hang waiting for input, e.g. by calling closeWriteChannel() immediately after start(). Or by calling start() with QIODeviceBase::ReadOnly, but, sadly, the documentation of QProcess doesn't document what effect that has.

Gpg4win-Beta-75, works:

Gpg4win-Beta-75: This is in principle ok. Even adding further commands to
C:\Program Files (x86)\Gpg4win\etc\xdg\kleopatradebugcommandsrc works:

@ebo @ikloecker Let me explain my thoughts. If you have time, please help me doing some tests in your environment.

This shell script running gpg-connect-agent should run successfully:

gpg4win-Beta-75:
German tooltip translations are all complete

For Beta-75 it looks similar judging from my first tries.

Gpg4win-Beta-75:
Looks good, I see no "EXCEPTION_ACCESS_VIOLATION" after quitting any more.
Instead only: "process started at 14:11:03.007 has terminated with exit code 0"

Gpg4win-Beta-75: This does not work.

Gpg4win-Beta-75:
Checked the property details of some files in "Progams (X86)" and found several without product name, E.g. libgpg-error-0.dll and other dlls located at C:\Program Files (x86)\GnuPG\bin.
Or gpgolconfig.exe at C:\Program Files (x86)\Gpg4win\bin and the examble from above, libKF5JobWidgets.dll, shows no product name either.

Gpg4win-Beta-75: Symbol is there.

The umlauts are ok again!