Summary: | Account settings lost at upgrade time. | ||
---|---|---|---|
Product: | [Unmaintained] kmail | Reporter: | Ian Turner <vectro> |
Component: | config compatibility | Assignee: | kdepim bugs <kdepim-bugs> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | CC: | bugs.kde.org, jim.lieb, lemma, mail |
Priority: | NOR | Keywords: | triaged |
Version: | 1.10.3 | ||
Target Milestone: | --- | ||
Platform: | Ubuntu | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
Ian Turner
2008-11-10 01:47:14 UTC
After restarting kmail, it appears that actually only "Receiving" accounts were lost, and that "Sending" accounts remain. I am having the same problem. Just upgrade my gentoo system to 4.1.2 to 4.1.3 and lost all recieving accounts. Same thing happens when I went from 3.x.x to 4.1.2. Did you upgrade from a kde3 version to a kde4 version? or you were using the kde4 version before the upgrade too? Are you sure the ~/.kde/share/config/kmailrc file is preserved? Maybe you were using ~/.kde4 as a settings prefix before and now you are using ~/.kde ? My upgrade was from 3.x to 4.1.2; it is quite possible that the kmailrc file was not preserved, though I certainly didn't delete it manually. I have managed to cause this problem on a full filesystem as well. Although I have not inspected the code itself, it seems that kmail re-writes ~/.kde/share/config/kmailrc with app state such as current message in folder etc. If this write fails for any reason, the kmailrc file is corrupted. If this analysis is true, I strongly suggest that configuration information that is both persistent and only modified under Settings-> be in kmailrc and the rest of the semi-persistent data such as current folder state be somewhere else. This would both reduce the extra I/O overhead and harden kmail to incidents such as this. If a local cache/persistent state store is needed, and it is useful, why not use something like sqlite which is hardened better than writing a full flat file from walking some internal copy of everything. @Jim I've also had the problem with a full filesystem. see bug 175680 Except the full filesystem, this bug may be related to bug 165320 > Except the full filesystem, this bug may be related to bug 165320 Right, it is, thanks. *** This bug has been marked as a duplicate of bug 165320 *** |