BTW, do we really need a C++ API for this? Might make sense due to the need for a context.

If RestorePositionForNextInstance=false is still there after gpgpass has been quit then either KMainWindow::closeEvent() didn't run or the dirty state config wasn't sync'd to disk. When reading the code in KMainWindow I was wondering if a sync() was missing, but in Kleopatra it worked without this.

Window is open:

@ebo: What does the MainWindow section in gpgpasstaterc contain?

The versioninfo.txt line of 5.0Beta-145 for gpgpass is:

It seems to work here on Windows and Linux

Note: Restoring the last window location is a built-in feature of KMainWindow/KXmlGuiWindow. It should work out-of-the-box if one makes sure that KMainWindow::closeEvent() is called even if it's overwritten. (Instead of overwriting KMainWindow::closeEvent() the better approach is to overwrite KMainWindow::queryClose() which is what I changed in Kleopatra.)

New API gpgme_op_random_bytes is now in master (gpgme 2.0). Use tests/run-genrandom --help for testing. Extra features will come soon.

Why do we need a word list if we are a password storage tool?

Fixed by https://dev.gnupg.org/rGPGPASS7587429ed4da7b5002cdbb35c4d1e7e649c2bac5

(ignore the last commit, I assigned the wrong task to it)

Fixed with https://git.gnupg.org/cgi-bin/gitweb.cgi?p=gpgpass.git;a=commit;h=51a227de27d53626ff87151b3cd619efd19b1804

Done with https://git.gnupg.org/cgi-bin/gitweb.cgi?p=gpgpass.git;a=commit;h=6e5d0ef58fff0277cb072b46b56080eb3f130ad9 and https://git.gnupg.org/cgi-bin/gitweb.cgi?p=gpgpass.git;a=commit;h=5ecb1e2e8eb0d4608be1e7fe04f6327b84753fcf