This is a bug fix release.

- for the X11 driver: backport of Mariano's MGRX improvement: this driver now responds to Expose events (i.e. repaints parts of the GRX window brought to front)

- for the win32 driver, correction of several annoying bugs . no more enqueue WM_PAINT events to give GR_OS_DRAW_REQUEST this is of no use and it had a great drawback: it saturated the grx's event queue, gobbling mouse and keyboard events. Compare responsivity to mouse and keyboard hits of tests which produce fast frequent drawings like test/mousetst.c test/bgi/bccbgi.c and pascal/bgi/demo.pas . some bitmap characters drawn were missing in various test programs. This was due to a Window's bug: the GetUpdateRect() function gives wrong UpdateRect. The BeginPaint() function gives a correct UpdateRect. This is contrary to Microsoft specs. Do'nt ask me why ! . Some Sleep() are needed to yield control to system to draw. Compare behaviour of PlayRGBpalette() in test/bgi/bccbgi.c (for this driver which has not a true palette working)

- for the djgpp driver: Now the syntax /dev/env/DJDIR works as expected for loading fonts: GRX_DEFAULT_FONT_PATH or direct loading by GrLoadFont() with an absolute path. Beware that many tests work because they give relative pathes towards the grx248/fonts directory. This is so to enable to run tests before installing the fonts.

- some random improvements

See the changes file

Enjoy, Maurice