GNOME Bugzilla – Bug 421091
Evolution does not properly cache offline mail
Last modified: 2008-04-10 12:15:23 UTC
Please describe the problem: I am currently using Evo to access an account with a 2 Gbyte inbox. Since the server connection is slow (60kbyte/sec, through vpn tunnel) I would really like to have everything synced for offline usage. At the same time, to sync the entire mail base, I would have to leave the computer on for probably 48 hours in a row. This is not feasible. My only option is to synchronise piecemal. Today I let synchronisation run until Evolution indicated that 27% were downloaded. At that point, the local cache of the inbox had some 700 MByte. At that point the VPN connection went away and I quit Evolution (no crash, regular exit). After restarting, I was shocked to see that the inbox had now shrunk to 136 Megs. Evolution had obviously thrown 550 Megs away, which had taken hours to download. I have made some more tests to see if this would repeat itself. At first, things went better and I had hope, that this would have been a temporary hickup. But just now, as I'm writing, the local inbox cache has again shrunk by 250 megs after regularly quitting and restarting the problem Steps to reproduce: 1. Download mail from a relatively large account 2. Measure disk usage of ~/.evolution/mail/echange and its subfolders 3. Quit the program and restart Actual results: As said above, sometime it works, but other times Evolution throws away hundreds of megabytes of mail, thus effectively preventing me from ever successfully completing synchronisation. Expected results: Evolution should preserve mail which has been downloaded. Does this happen every time? No. But frequently enough to seriously annoy me. Other information:
For now, before quitting, as a step 2.5) Click on any small folder (this switching-between folders forces a local sync and thus cache and summary are updated). This is just a workaround and a fix will be provided soon.
this seems to have been fixed. I have used Evolution 2.20 and 2.22 for some time and I found that the exchange caching behaviour is now quite robust and reliable. So this bug can be safely closed.