[avr-libc-dev] Re: testsuite failures

[Joerg Wunsch]

As Joerg Wunsch wrote:

> Simulate: math/isinf-01.c atmega128 ... *** simulate failed: 22
> Simulate: math/isinf-01.c at90s8515 ... *** simulate failed: 22
> Simulate: math/modf-np.c atmega128 ... *** simulate failed: 47
> Simulate: math/modf-np.c at90s8515 ... *** simulate failed: 47
> Simulate: math/signbit-01.c atmega128 ... *** simulate failed: 2
> Simulate: math/signbit-01.c at90s8515 ... *** simulate failed: 2
> Simulate: avr/eeprom-1.c at90s2313 ... storage.c:97: ERROR: address out of 
> bounds: 0xff16

I could fix the last one by removing the old AT90S2313 simulator bug
workaround, as the original code now works as expected, for any of the
tested MCU types.

This leaves the floating-point testcases open as "fail".  I've really
got no idea how serious they are.  math/isinf-01.c also fails when run
on the host, while modf-np.c and signbit-01.c only fail on the AVR.

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)