Backported for VSD 3.3.x / Gpg4win 4.4.1

I'll backport this because this issue is a really bad usability problem for people using the Windows Explorer integration to decrypt files. And the regression risk is minimal.

For me the change fixes the problem on Windows. (I haven't checked if there was a problem on Linux/X11, but I have verified that the change also works on Linux/X11.)

This also happens if one activates the Kleopatra main window, then activates the Windows Explorer and then starts a decrypt&verify operation.

I have seen this when doing the following:

• Start Kleopatra and minimize its window
• Start decrypt&verify operation for an encrypted archive in Windows Explorer with GpgEX

Gpg4win 4.4.0 has just been released. Creation of encrypted archives has been completely reworked with Gpg4win 4.3.0 already. I could create an encrypted archive containing the two files with GpgEX without problems.

I tried to reproduce this with Gpg4win 4.4.0, but it worked for all PKCS#15 cards I tried (dark blue R&S, Genua, ipt.fraunhofer). As soon as the certificates were imported from the card (and a "Do you trust this root certificate" dialog was canceled) the buttons became enabled.

Gpg4win 4.4.0 has just been released. I cannot reproduce unexpected results when doing a lookup on server. Unfortunately, many keyservers do not publish the user IDs (i.e. name and/or email address) anymore. Kleopatra does not list results without user IDs because GnuPG won't import them anyway.

This has been fixed in the meantime. If you should still experience this problem with Gpg4win 4.4.0 then reopen this ticket.

Gpg4win 4.4.0 has just been released. I assume that this has been fixed in the meantime. In particular, "General error" should happen a lot less. If the problem still occurs with Gpg4win 4.4.0 then reopen the ticket.

I assume the problem has been resolved because we never got feedback that the problem persists.

Interesting. Just a few days ago I noticed that Kleopatra creates a new tab with the imported certificates except when the import is done via the notepad (or via the clipboard). I guess we could add an option to open a tab with the imported certificates to the import result window.

This problem has been fixed quite some time ago. See https://bugs.kde.org/show_bug.cgi?id=415168.

This doesn't look like a problem in Kleopatra.

We have to take the list of OIDs to ignore into account that can be configured for gpgsm with ignore-cert-with-oid.

We won't revert to the old UI which had its own share of problems. In the meantime Kleopatra supports certificate groups which should help if you often need to encrypt for the same keys.

Kleopatra now let's gpg deal with I/O directly. This makes decryption with Kleopatra as fast (or slow) as with gpg on the command line. Hence I'll remove the Kleopatra tag.

Fixed.

I also checked this on Windows 10 with light, dark and high contrast (theme No. 1) colors:

Fixed. After deleting the certificates belonging to the first two card slots the smartcard window shows

Kleopatra does now read the certificates from the card and import them itself instead of relying on gpgsm --learn-card.

In T7273#195063, @ebo wrote:

We could discuss how to make this a bit more easy to parse visually, but that would be a new task.

Should probably be backported for VSD 3.3 because very likely this happens always when a new PKCS#15 card (with unknown certificates) is inserted. Workaround: Close and re-open the smart card window.

Fixed (for Gpg4win 4.4 / VSD 3.3) by patching Qt's Windows Vista style.

Backported for VSD 3.3

Shall we use the narrower tooltips in general? I'm undecided. On one hand, we'd avoid an unholy mix of wider (but not too long) one-line tooltips (at least for English and maybe German) and narrower multi-line tooltips. On the other hand, I think the multi-line tooltips are too narrow and a narrow two-liner might look uglier than a slightly wider one-liner.

I have added the NEWS for Kleopatra for Gpg4win 4.4.0. I have excluded the following tickets from the NEWS for various reasons:

gnupg24 (gnupg-2.4.5) is a milestone and means "fixed in gnupg24 (gnupg-2.4.5)". I suggest not to question the workflow of the people running this bug tracker.

For T6567 I changed the way that Kleopatra runs "gpgconf --launch gpg-agent". This change is not yet in Eva's test build. It seems my change is not good because running "gpgconf --launch gpg-agent" timed out after 5 seconds in 3 of 3 tests starting Kleopatra after a reboot of the VM. To check if "gpgconf --launch gpg-agent" really takes that long I measured the time in PowerShell after another reboot of the VM. The result is shocking.

Result of first iteration sorting the tickets by Features, Bug Fixes, and different reasons why they are not relevant for the release notes of 4.4.0

Noteworthy changes in Version 4.4.0 (unreleased)
------------------------------------------------

This doesn't need to be verified with VSD 3.3 after it was verified with Gpg4win.

This was fixed in gpgme and is therefore automatically in VSD 3.3

Ready for testing. Note that you also need gpgme master.