From: Erik P. <epe...@iv...> - 2012-05-18 20:49:36
|
On Fri, 18 May 2012, Philipp Klaus Krause wrote: > Dear sdcc developers, > > I have added two new ports to sdcc: The r3ka and s08 ports. > Currently they are still dummy ports: No Rabbit 3000A or S08-specific > instructions are used, so r3ka generates code as r2k does and s08 > generates code as hc08 does. > For s08 I also added the regression testing infrastructure. > Summary for 's08': 0 failures, 6884 tests, 1540 test cases, 2408132 > bytes, 29426282 ticks > Summary for 'hc08': 0 failures, 6884 tests, 1540 test cases, 2409859 > bytes, 29466533 ticks > The difference in code size and ticks is due to > 1) bug3223041.c behaves differently for hc08 than for other ports, but > I don't know why, so I did not make it behave for s08 like it does for > hc08. It looks like this is just a remnant of when hc08 did not yet support bool. > 2) I did not break the string functions for s08. We probably should > discuss if string functions should be broken on s08 (and if they > should stay broken on hc08). There wasn't anything intentionally broken (other than the previously discussed reentrant vs non-reentrant calling conventions) with the library functions. What needs to be fixed? Erik |