[Mono-list] PowerPC issues and floating points + casts

PRINGLE Chris chris.pringle at miranda.com
Wed Feb 16 05:54:57 EST 2011


Hi All,

I'm having a few problems with mono under PowerPC. I cannot get anything floating point. When I run the regression tests for mono, it passes most, but fails all the float/cast related ones. I tried running the regression tests after I found that Math.Round didn't work properly. I put GDB onto mono and found that the value passed into ves_icall_System_Math_Round (in sysmath.c) is always the same (2.0132...e-230), regardless of the constant expressed in the C#.

I've tried using both a cross compiled and native compiled version of mono, and both produce the same results. I had to add the following compiler flags during build "-D__powerpc__ -D__ppc__ -Xassembler -mregnames".

Regression tests fail as follows:
Test run: image=/mnt/cpringle/rcheck/basic-float.exe, opts=
Test 'test_0_conv_i' failed result (got 2, expected 0).
Test 'test_2_rem' failed result (got 8, expected 2).
Test 'test_4_long_cast' failed result (got 0, expected 4).
Test 'test_4_ulong_cast' failed result (got 0, expected 4).
Test 'test_4_single_long_cast' failed result (got 0, expected 4).
Test 'test_0_lconv_to_r8' failed result (got 1, expected 0).
Test 'test_0_lconv_to_r4' failed result (got 1, expected 0).
Test 'test_0_ftol_clobber' failed result (got 2, expected 0).
Test 'test_0_rounding' failed result (got 1, expected 0).
Results: total tests: 30, failed: 9, cfailed: 0 (pass: 70.00%)
Elapsed time: 0.017374 secs (0.003467, 0.013907), Code size: 9544

Any ideas?

Cheers,
Chris

____________________________

Miranda Technologies Limited
Registered in England and Wales CN 02017053
Registered Office: James House, Mere Park, Dedmere Road, Marlow, Bucks, SL7 1FJ
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ximian.com/pipermail/mono-list/attachments/20110216/774e0a82/attachment.html 


More information about the Mono-list mailing list