SUMMARY The signing dialog that is accessible from "Tools>Digitally Sign..." repeatedly asks for a certificate password even though the certificate doesn't have any, and the user enters nothing. STEPS TO REPRODUCE 1. Open a PDF 2. Select "Tools>Digitally Sign..." 3. Click "OK" OBSERVED RESULT The application goes into an endless loop asking for a certificate password EXPECTED RESULT Since no certificate is selected, either disable the "OK" button or select the first certificate at dialog init. SOFTWARE/OS VERSIONS Okular: 25.04.1 KDE Frameworks: 6.14.0 Qt: Using 6.9.0 and built against 6.9.0 Manjaro Linux (Xcb) Build ABI: x86_64-little_endian-lp64 Kernel: linux 6.12.28-1-MANJARO ADDITIONAL INFORMATION The bug is not present if one selects a certificate from the list, in my case the certificate I have in the NSS database. This is the location of the loop: https://github.com/KDE/okular/blob/b1d97424246e43a230a2875458eab28231028abc/part/signaturepartutils.cpp#L223
Similar to #486680 but in a newer version of Okular that should have presumably fixed it
Are you saying it is a exact replica of the other bug you mentioned? That if you select first, then it works as expected?
(In reply to Sune Vuorela from comment #2) > Are you saying it is a exact replica of the other bug you mentioned? That if > you select first, then it works as expected? I would not say so. The steps to reproduce and observations do match. The setup varies: The other bug is about okular deployed with Flatpak, in my setup it's through Archlinux pacman.
(In reply to Sune Vuorela from comment #2) > That if you select first, then it works as expected? This is correct
If I can give you a couple of pointers, would you be willing to give it a go in trying to fix it ? I'm 95% certain that the selection that is supposed to happen here: https://invent.kde.org/graphics/okular/-/blob/master/part/signaturepartutils.cpp?ref_type=heads#L81 somehow leads to reading it back here: https://invent.kde.org/graphics/okular/-/blob/master/part/signaturepartutils.cpp?ref_type=heads#L207 fails
(In reply to Sune Vuorela from comment #5) > If I can give you a couple of pointers, would you be willing to give it a go > in trying to fix it ? sure, let's try
Quick drive-by comment: The last working version is 24.12.3. 25.04.0 shows the same erroneous behaviour.
I suffered this too, and by chance came to a post somewhere mentioning to check $ cat ~/.pki/nssdb/pkcs11.txt there, I got rid of the *onepin* entry and magically this started acting normal again. (naturally, this presumes you're using nss)