Version: 4.7 (using KDE 4.7.0) OS: Linux Have kmail configured to fetch mail from a pop3 account. Sometimes it happens that kmail/akonadi fatches mail from the pop3 server but the mail never appears in kmail. I noticed this because on some mailing-lists I am subscribed to, mails in a thread were missing in kmail. As I am affected by bug 271231 (akoandi stores mails twice locally) I did a grep for some unique string in such a missing mail (after I configured kmail to leave the messages on the server and no longer delete them) for the relevant folders. The string was not found in ~/.local/share/.local-mail.directory and ~/.local/share/akonadi/file_db_data but in the binary-files ~/.local/share/akonadi/db_data/akonadi/parttable.ibd and ~/.local/share/akonadi/db_data/ib_logfile1 I am using the mysql backend for akonadi. I never noticed this with kmail2 from KDE 4.6! If I would have to guess I would say something between 1 out of 20 to 50 mails is affected by this. I see this behaviour with two accounts. Please tell me which additional information I can provide to solve this bug. Reproducible: Sometimes Steps to Reproduce: see above Actual Results: kmail does not show some mails after fetching them from a pop3 server and they are also not saved! This means data loss, usually without a chance to notice it. Expected Results: kmail saves and shows all mails. OS: Linux (x86_64) release 2.6.37.6-0.7-desktop Compiler: gcc
One important additional information: It seems to be that the data loss only happens for mails for which a filter *for incoming mail* is applied. Today I disabled all filter for incoming mail and only applied the filters manually via "Ctrl+J". So far there has not been any new mails lost. Additionally the fact that mails are shown twice in kmail, see https://bugs.kde.org/show_bug.cgi?id=271231#c9, does also not happen any longer. If my observations turn out to be wrong, I will report back.
I have to correct myself. The bug also happens when no filter (which moves the mail directly into another folder) applies. But for me it seems to happen less often in this case. But that is a bit hard to determine.
This seems to be the same as bug 279297. I added some steps that would help with debugging there. Maybe you can try those steps too? *** This bug has been marked as a duplicate of bug 279297 ***