This is related to the KTcpSocket removal. There are however a few users of this independent of KTcpSocket, such as ktp-auth-handler, and the KIO certificate manager and SSL error UI.
This might require making KSslError a thin wrapper around QSslError as a transition step during KF5 time without breaking API, so we can port the users already.
Both QSslError and KSslError have their own human readable error messages:
The Qt ones look a lot more comprehensive than ours, the KDE ones might have better translation coverage. How to proceed here:
Code-wise the first option looks best, but I have no idea how big the impact on translations would be.
The impact is relatively big
tsdgeos@yoga:~:$ pacman -Ql kio | grep kio5.mo | wc -l
83
tsdgeos@yoga:~:$ pacman -Ql qt5-translations | grep qtbase | wc -l
22
Mostly because translating qt is kind of hard mode translation/community work.
But on the other hand Qt *needs* to be translated since we use it for OK/Cancel of [some] dialogs so i *think* it would be OK to remove them and then we need to remind translators that Qt is our foundation and it probably needs to be translated too.
@ltoscano what's your take?
Sorry, I missed it. I'd say that if we fail to remind translators that we need to translated Qt, we have much bigger issues than a bunch of SSL *error* messages.
Related Qt change to clean up manual error enum to string mapping in KIO: https://codereview.qt-project.org/c/qt/qtbase/+/277684