trunc(768.3156) == 770?!

[Jiri Svoboda]

Adding the code fragment

TPRINTF("trunc(768.3156)=%lf\n", trunc(768.3156));

to tester yields the output

trunc(768.3156)=770.000000

However the float2 test succeeds, because after multiplication by precision and conversion to integer the result is as expected.

[Martin Decky]

On what platform do you observe this? Could it be an issue in the printf?

If I port both the generic code (uspace/lib/math/generic/trunc.c) and the ia32-specific FPU code to Linux, I don't observe this behavior.

[Martin Decky]

OK, it's very probably a bug in printf and not in the arithmetic functions. The following code

printf("768.3156=%lf\n", 768.3156);
printf("768.3150=%lf\n", 768.3150);
printf("768.3100=%lf\n", 768.3100);
printf("768.3000=%lf\n", 768.3000);
printf("768.0000=%lf\n", 768.0000);

produces the following output

768.3156=768.315600
768.3150=768.315000
768.3100=768.310000
768.3000=768.300000
768.0000=770.000000