                               ---------
                    Approximate bug list for RoadMap
                               ---------
                               ---------
                             October 2007


KNOWN BUGS -- patches welcome

Desktop and rendering bugs:

 * Gtk2 interface reports error:

   The following error message appears:

   (gtkroadmap:3487): WARNING **: Invalid UTF8 string passed to
        pango_layout_set_text()

 * Long lines are sometimes incorrectly displayed
 
   The symptom is that all of a sudden a long straight line that you know
   isn't really present will appear from edge to edge of your screen.

 * QT3,QT4 - no display periodic running

   Periodics seem to be attached to windows (?), so the new periodic started
   in roadmap_display_initialize() doesn't commence since there's no window
   yet.

 * QT3,QT4 - no signal handling

   No signal handling has been added to the QT versions.  This means user
   changes to routes and waypoints, and the track log, can easily be lost.

   (The GTK versions handle signals properly, and so RoadMap will shut down
   cleanly in the event of the user hitting ^C in the window from which
   RoadMap was started, or if RoadMap is killed with SIGTERM, as when the
   system is shutting down.)

 * WinCE,Win32 - Executables built with the arm-wince-mingw32ce
   cross-compiler don't run.
   
   It would be great if someone could finish that work.  And win32 hasn't even
   been attempted in cross-compilation.  Using a cross-compiler has big
   advantages for the principal developers, so please don't just toss that
   notion aside without some discussion.

 * GTK/GTK2 - Deprecated interfaces in use.

   Both flavors of GTK build make fairly heavy use of deprecated API pieces. 
   Someone with good GTK skills should clean these up.

 * Edge cases.
 
   If you get too close to either the north or south pole, RoadMap will
   get confused, and may crash.  If you go past the 180th meridian, in either
   direction, RoadMap may or may not get confused, but you probably will.
   
Map building bugs:

 * Shape point are ignored when drawing polygons:

   The Tiger files define polygons as delineated by a list of lines.
   Any of these lines might be defined by a serie of shape points.

   RoadMap uses the shape points when drawing the lines, but not when
   filling the polygons.

Map data bugs:

 * City search is broken:

   The roadmap files include a table to convert a place into a county
   subdivision.  This table causes a lot of problems because places can span
   across county subdivisions, with name clashes.

 * The "address by ZIP" method does not work anymore:

   Since the automatic selection of the map, because the file usdir.rdm does
   not have yet any ZIP code location information (TBD).

   This feature has been hidden from the user interface  for now.

 * Street address conflicts:

   A few street address conflicts appears when building maps.  Some seems to
   be a TIGER map database problem.  Others might be caused by the address
   range format issue (see below).  A few still need to be investigated.

   NOTE:  3 of these addresses conflicts have been confirmed to be TIGER
   database errors by the census bureau.

 * Street numbers are not always numerical:

   Some street numbers include letters, which is not well supported by
   RoadMap:  in such a case, roadmap generates a street number which is hoped
   to be unique enough by converting the letter into a "number".

   This is a design problem that is not going to be solved any time soon. 
   Current kludge is deemed good enough for the time being, until someone give
   a practical case where this does not work.

 * Some urban areas are missing:

   For example:  Woodland Hills (in Los Angeles county).  This is a TIGER file
   problem:  the string "WOODLAND HILLS" never appears in the LA county files.

 * The maps for the state of Hawaii do not use the NAD83 datum:

   The offset can be significant (a few hundred feet), making these maps
   basically unusable for navigation.  The USGS bureau has announced that
   their maps of Hawaii are being redone using the NAD83 datum, lets hope the
   US Census Bureau will do the same.  [As of 2007, I don't know whether this
   is still true.  -pgf]

