- maintain a UID-serial number map during runtime to find messages fast
- use this map in the searchjob to speedup imap searches by a magnitude. You only get the speedup when the map is already build of course.
- use a progress item when messages need to be downloaded. The progress is currently not canceled when the search is stopped, I still have to implement that.
General search changes:
- include a matches bool in the searchDone signal and use that to check changed messages. This fixes the problem that status changes did not update the search.
BUGS:76181
svn path=/trunk/kdepim/; revision=409056
is quieted.
* Delay/compress UpdateCounts in FolderTree.
With this improvements,kmail don't eat the 100% CPU when refresh the cache from
a dimap server when this resides on the same lan.
svn path=/trunk/kdepim/; revision=405858
remain stable during the lifetime of a message. In other words, don't
crash when changing the subject header of a message which is part of
the current subject threading trees. Pfew, this one was a bugger to fix.
BUGS: 94043
svn path=/trunk/kdepim/; revision=390027
dialog and include it in the RMB menu of the foldertree. To make the check if the menu
should be shown a bit easier the folderstorage got a new "isMoveable()" function.
I noticed that moving of imap folders crashes again when the messages are moved (sigh)
but this is not related so I'll fix it afterwards.
svn path=/trunk/kdepim/; revision=385541
folder was get'ed for every single mail. Now the folderstorage has an
addMsg( QPtrList<KMMessage> ) method and the imapjob can handle put operations
with a msgList. And the thingy has a nice progressitem now.
svn path=/trunk/kdepim/; revision=385105
can only be moved when they have no children.
Along the way fix Bug 94125 and add the FolderRequester to the expire settings
in the folder dialog.
BUG: 94125
svn path=/trunk/kdepim/; revision=371419
happens when a mail that was supposed to be downloaded could not be,
because it was no longer on the server (deleted by another client).
svn path=/trunk/kdepim/; revision=366604
when simply opening mail folders) : don't look for the serial numbers
of messages that haven't been inserted into a folder yet, it can't find one,
and it's slow every time. Note that assign() is only called on construction
(or soon after), never on totally-unrelated instances.
svn path=/trunk/kdepim/; revision=361937
It simply combines the code from getMsg and readMsg, without the mMsgList.set() call.
(*) Callgrind shows that unGetMsg takes time too, mostly due to strippedSubjectMD5(),
so let's avoid it altogether, as discussed with Till.
svn path=/branches/KDE_3_3_BRANCH/kdepim/; revision=360560
- the search folder should of course also check added messages correctly
- finish the progressitem when the move to an imap folder fails
svn path=/trunk/kdepim/; revision=358300
This leads to writing uninitialized data out (!) since readConfig hasn't read everything yet.
Tested changing the type of a folder and creating a new folder with a contents type, both work.
svn path=/branches/KDE_3_3_BRANCH/kdepim/; revision=348874
the config to disk after changing the contents type. Please someone
implement parameterized groups in KConfigXT.
svn path=/trunk/kdepim/; revision=339001
Otherwise if the user recreates the same folder again, the settings will interfer.
In the case of cached imap this led to imappath being set, so kmail thought
the folder was deleted locally (insted of "just created"), and would delete it again.
svn path=/trunk/kdepim/; revision=324871
I created a new FolderStorage::listDirectory method to make it easier.
Add a Refresh folder list to the context menu of imap folders as the user
will probably not know how to refresh the listing (collapsing does the same).
This is needed together with the list-only-open-folders thing.
svn path=/trunk/kdepim/; revision=314858
Change the KMFolderMgr::find methods a bit to actually find all folders.
Oh well, and assign ids to all folders.
svn path=/trunk/kdepim/; revision=313955
Patch semi-approved by Carsten and semi-approved by Till, which makes it fully approved.
If it saved mGuessedUnreadMsgs it would get into mUnreadMsgs on re-reading,
which would be wrong I guess. But when we have a guess, don't save mUnreadMsgs,
to be closer to the old behavior (we'll guess again on reopening, I suppose?).
svn path=/trunk/kdepim/; revision=303284
Thomas Zander
Carsten Burghardt
Till Adam
Mario Teijeiro Otero
David Faure
Bo Thorsen
Ingo Klcker