Moved kpilot from kdepim to applications, as the core Trinity libraries should not contain hardware-dependent software

git-svn-id: svn://anonsvn.kde.org/home/kde/branches/trinity/applications/kpilot@1221127 283d02a7-25f6-0310-bc7c-ecb5cbfe19da

1 files changed, 201 insertions, 0 deletions

diff --git a/TODO b/TODO
new file mode 100644
index 0000000..1729faf
--- /dev/null
+++ b/TODO
@@ -0,0 +1,201 @@
+This is a TODO file. Open issues are unmarked, partly resolved issues
+are marked with a *, ignored issues get an -, and resolved issues a +
+(or they are completely removed from the list). Check the ChangeLog
+for details on how an issue is resolved, since resolved issues are
+ periodically purged from this list.
+
+
+* Import pilot-link 0.12 CVS again
+* Implement correct full syncing in memofile conduit.  honestly, we should
+  really have a generic sync engine that knows how to do all of the
+  different sync types and only calls conduits for comparisons, list
+  fetching, pilotid->whateverId persistence and retrieval, etc.  *sigh*  is
+  that what opensync gives us?
+
+Essential
+=========
+3b) Implement reset to last sync and ignore conflict resolution options in
+    vcal conduits (bug #59221)
+4a) Config troubles: Start kpilotDaemon and then kpilot. Configure a conduit
+    in kpilot, then do a sync (which writes something to the  config, e.g.
+    run the docconduit). Now configure another conduit in kpilot. The config
+    file will be changed again, but the changes from the sync are lost (i.e.
+    after a sync, kpilot needs to reload the config). (seems connected to bug #56115)
+*4b) Conduits should not use the backup databases, otherwise a backup run will
+    screw up the conduit's mechanism to detect changes. (bug #59219)
+4c) Need to find a better way to detect runs without a conduit, and switch to
+    full sync in that case to prevent data loss.
+4d) Need to find a good way to avoid a mess when implementing new features
+    (e.g. after implementing the correct category sync, how do I ensure that
+    no category settings are lost?)
+5)  After the special sync is done, ask if the next sync should again be the
+    same special sync, a fast or a hot sync. (wish #59218)
+
+
+Important (hight priority)
+============================
+6a) The abook conduit currently only syncs one email address and one
+    phone number of each type. I need to find a way to sync two email
+    addresses to and from the handheld. (Bug #56364)
+7)  Check the first-time syncing behavior of all the conduits.
+15) Add force-first-time to the KNotes conduit.
+17) Include a calendar viewer, or get rid of the internal viewers
+    altogether.
+19) GJJ: Replace newlines in TODO summaries (not allowed in libkcal),
+    also do the same for phone numbers etc. (bug #56379)
+19b)Make the doc conduit use the new global settings instead of its own 
+
+
+Medium (nice to have, but not essential)
+========================================
+18) VCal conduit: events going over midnight need to be split up
+    on the handheld (bug #59223)
+20) GJJ: Check all debug output, what should be put in the log (i.e.
+    be visible to the user) and what is just debug output to track
+    down problems.
+21) GJJ: Hunt down and kill printfs. Hunt down bad i18n.
+26) ABook conduit: sync with preferred address (not home or work address)
+27) Update the conduit programming howto (some changes in the database
+    and record classes)
+31) GJJ: Extend all conduits to allow arbitrary URLs instead of local
+    files.
+32) JPilot conduit: fix the configure.in.in checks for gtk/glibc
+    (although the conduit isn't even remotely finished)
+32a)Find a replacement for the NULL conduit (skip databases when
+    doing a backup)
+
+
+Unimportant (lowest priority)
+=============================
+33) JPilot conduit: Implement the conduit as an external GTK application,
+    that is embedded into kpilot using qxembed
+34) JJ: Add tooltips to all controls in all conduit configure dialogs
+35) Do not store the RecordID with the vcalendar entry, but in the kpilot
+    config
+37) JJ: Show the sync log as a KPassivePopup
+39) Make an application that just has a LogWidget and KPilot's dcop
+    interface (and registration!) so that you can follow the daemon's
+    log output without starting KPilot at all.
+40) Allow the todo conduit to use the manana database (see jpilot)
+41) Allow the memo conduit to use the pedit32 database (see jpilot)
+42) Use pdf2text application to also install pdb files with the palmdoc
+    conduit.
+
+
+
+
+
+
+FINISHED TODO jobs
+==================
+
++(1)  VCal conduit: Currently using the ResourceCalendar (standard
+    calendar) doesn't work for some (unknown) reason. Every sync
+    happens with the filename given in the setup dialog. Probably
+    very easy to fix.
++(2) 4d) Need to find a good way to avoid a mess when implementing new features
+    (e.g. after implementing the correct category sync, how do I ensure that
+    no category settings are lost?)
++(3a) Implement the CopyToPC and CopyToHH directions in the addressbook conduit (bug #59220)
+5)  After the special sync is done, ask if the next sync should again be the
+    same special sync, a fast or a hot sync. (wish #59218)
+
+ Redesign the config dialog of kpilot (get rid of deprecated
+    options, include conflict resolution, etc.)
++(3)  Fix all conduits to use the new global options: conflict resolution,
+    emergency sync, sync mode etc.
++(4)  VCal conduit: Fields not supported on the handheld are erased
+    in the calendar (e.g. categories). I need to get the Event
+    before I copy all fields over, and only overwrite the fields
+    from the handheld, so the others remain unchanged.
+	 RESOLUTION: Found out that this is not true any longer, and everything
+	 works just fine!
++(5)  AvantGo conduit: fix the configure.in.in checks for libmal
++(6)  ABook conduit: fix the first sync setting
++(8)  Find out what's the reason of the "Invalid unique ID" crashes in
+    the calendar conduits (and the addresssbook conduit) (bug #51216)
+-(9)  GJJ: PalmDOC conduit: kpalmdoc is run as a modal dialog, not as
+    a standalone dialog (e.g. create a link  on the desktop to
+    kpalmdoc and start kpalmdoc from there. You can't access the
+    desktop until you quit kpalmdoc)
+    RESOLUTION: This is a general KDE Bug!!!!
+-(10) GJJ: Add a flag to the log{Error,Message} functions to prevent
+    the addition of the trailing <br> and the date to each output
+    (e.g. to put all output from the AvantGo conduit in one line,
+    instead of one separate line for each .)
+    RESOLUTION: This is not possible in a QTextEdit in LogText mode.
+    However, I got rid of the pseudoo-progressbar, so this i no longer
+    needed.
++(11) Merge between HEAD and BRANCH (again).
++(12) ABook conduit: Check all fields for differences before asking
+    for conflict resolution. (i.e. don't ask for conflict resolution
+    of single fields, just for the whole record) (bug #59222)
++(13) ABook conduit: custom field sync doesn't work with birthdate
+    (how do I convert a date to a string and back in an arbitrary
+    format???) (Bug #50871)
++(14) ABook conduit: Fix categories sync
+    RESOLUTION: Fixed for the abook and the todo conduit
++(16) Remove the edit capabilities from the built-in viewers. (fixes bug #54765)
+    Or write a SyncAction childclass that merges  the changes from the internal
+    editors to the databases on the handheld before the sync takes place
+		RESOLUTION: Implemented the InternalEditorAction that syncs these modifications
++(19a)Implement the internal editor sync
++(19a)When syncing, also backup the databases (if option is set)
++(19b)Add all codecs that are supported by Qt, automatically fill the
+    combobox in the config dialog. Change the combobox to an editable
+	 Combobox so that the user can also give other encodings (if he
+	 installed the corresponding plugins).
++(22) Update README, TODO, NEWS, AUTHORS in HEAD.
++(23) GJJ: Move the conduit configuration dialog (for the umpteenth time)
+    to a list of QCheckBoxes, like noatun's plain playlist or
+    kdebugdialog
++(24) PalmDOC conduit: choosing the bookmark type of the pdb->txt
+    conversion is not yet implemented. Only create .bm file, if there
+    are bookmarks in the text.
+-(25) ABook conduit: sync with multiple email addresses (space-separated
+    list received from handheld)
+		RESOLUTION: Duplicate of 6a)
++(28) GJJ: The question asked when kpilot and the pilot's usernames don't
+    match isn't very intuitive. It should be changed to a yes/no/cancel
+    with "Use KPilot name" "Use Pilot name" "Cancel". This is in
+    interactiveSync.cc.
+-(29) Fix the NULL conduit's tooltip - OR - make it actually do what the
+    current tooltip says.
+    RESOLUTION: The NULL conduit is no longer needed at all, so I disabled
+    it completely
+-(30) GJJ: Create good icons for the kpalmdoc converter application.
++(38) Should interactiveSync and syncStack be moved to lib/ ?
+
+
+
+
+
+Specific JOBS
+=============
+
+***
+*** PilotLocalDatabase fixups
+***
+
+The PilotLocalDatabase class is really ugly and clunky -- and far worse,
+it has a hard-coded limit of 10000 records per database. This may have hurt
+us once or twice already (consider that a Visor w/ 16Mb could *easily*
+hold 10000 addresses). It needs careful reworking.
+
+I think using a Q(Ptr)List is the way to go here -- well, maybe a QVector
+would work as well; that depends on whether QVectors resize well. The tricky
+part is implementing the "current record pointer" with the same behavior
+as the current fCurrentRecord member, since conduits may depend on it.
+
+
+***
+*** Popmail Conduit & IOSlaves
+***
+
+The popmail conduit duplicates a lot of code found elsewhere -- it does
+SMTP sending and POP3 retrieval of mail messages. That should really be
+left to IOSlaves, which are better debugged. Using tickle() will be
+necessary while waiting for responses from the slaves, since you must keep
+the pilot awake. (This opens the door to multi-threading conduits, with
+possibly an asynchronous "pre-sync" action).
+