use the same format string in the test program as in LEGACY_PLUGIN for record keeping purposes

TODO.txt

@@ -48,37 +48,30 @@ E6)  Start initial work for changing component library file format to use Dick's
 PCBNew
 ------
 
-* Make the zone hit testing be done in screen coordinates, not internal units.
-  See the @todos in class_zone.cpp
+*)  Make the zone hit testing be done in screen coordinates, not internal units.
+    See the @todos in class_zone.cpp.  A fixed distance in internal units becomes
+    a variable distance based on zoom factor, leading to inconsistent results at
+    various zoom factors.  I believe that a fixed distance in pixels might make
+    for a friendlier UI.
 
-
-Dick's Peronal TODO Items  (Last Update: 24-April-2012)
------------------------------------------------------
-
-1)  Work through some scroll, pan, zoom overflows in PCBNEW's nanometer build mode.
-    It is thought that if we can constrain the virtual IU space to within
-    INT_MIN to INT_MAX then a whole host of other problems will go away.  Most
-    of the evil is in EDA_DRAW_FRAME::AdjustScrollBars() which assumes the
-    virtual IU space is infinite.  This function triggers a movement of the
-    viewport within the virtual IU space and also a change in size of the virtual
-    IU space.  Once this happens, you can end up thinking there are problems in
-    functions like EDA_DRAW_PANE::DrawCrossHair(), but this may be an artifact
-    of having traveled outside a limited virtual IU space.
-
-
-2)  Check that the new load visibility BOARD settings is properly setting the toolbar
+*)  Check that the new load visibility BOARD settings is properly setting the toolbar
     buttons like show grid or ratsnest.  Add PCB_EDIT_FRAME::SetVisibleElements() so
     toolbar crap is not known to a BOARD.
 
-3)  Finish removing global access requirements from KICAD_PLUGIN, so that:
+*)  Finish removing global access requirements from PLUGINs, so that:
     *)  a BOARD is a fully self contained document description.
     *)  plugin developers do not have to access globals, since a plugin could
-        very well be a dynamically loaded DLL/DSO.
-        One final problem remains with BASE_SCREEN's grid origin, easy solution is to
-        move just that one field into the BOARD.
+        very well be a dynamically loaded DLL/DSO in the future.
+        One final problem remains is the BASE_SCREEN's grid origin.  An easy
+        solution is to move just that one field into the BOARD.
+
+*)  Add ::Footprint*() functions to EAGLE_PLUGIN, so that Eagle footprint libraries
+    can be used in situ.
+
+*)  Add a library table for Pcbnew like that in the sweet library and get rid of the
+    damn search path strategy.  This will enable concurrent usage of various types
+    of PLUGIN::Footprint*() functions.  At least LEGACY and KICAD are both needed
+    concurrently.
 
-4) Do an EAGLE XML import PCBNEW PLUGIN, and possibly add export support to it.
-   This is PLUGIN::Load() and maybe PLUGIN::Save().
 
-5) Get back to the SWEET work.
 

pcbnew/legacy_plugin.cpp

@@ -2659,7 +2659,7 @@ BIU LEGACY_PLUGIN::biuParse( const char* aValue, const char** nptrptr )
 
     // fval is up into the whole number realm here, and should be bounded
     // within INT_MIN to INT_MAX since BIU's are nanometers.
-    return BIU( KiROUND( double( fval ) ) );
+    return KiROUND( fval );
 }
 
 

tools/test-nm-biu-to-ascii-mm-round-tripping.cpp

@@ -43,7 +43,7 @@ std::string biuFmt( BIU aValue )
 
     if( engUnits != 0.0 && fabsl( engUnits ) <= 0.0001 )
     {
-        len = snprintf( temp, sizeof( temp ), "%.16f", engUnits );
+        len = snprintf( temp, sizeof( temp ), "%.10f", engUnits );
 
         while( --len > 0 && temp[len] == '0' )
             temp[len] = '\0';
@@ -52,7 +52,7 @@ std::string biuFmt( BIU aValue )
     }
     else
     {
-        len = snprintf( temp, sizeof( temp ), "%.16g", engUnits );
+        len = snprintf( temp, sizeof( temp ), "%.10g", engUnits );
     }
 
     return std::string( temp, len );;