Re: bugs with count_leading_zeros and gmp-4.0

[Torbjorn Granlund]

Nathalie Revol <address@hidden> writes:

  Hello
  
  we (me by chance and Paul by analyzing "my" bug) encountered a problem
  with count_leading_zeros when applied to 42 (= 101010 in binary),
  cf. C program below.
  The result should be 26 = 32-6 and it is 27:
  
  pastaga% gcc -I. -L.libs toto.c -lgmp  (in the gmp-4.0 directory)
  pastaga% ./a.out
  k=27
  
I am unsable to reproduce.  Using a 32-bit build, I get 26, and with a
64-bit build I get 58, as expected.

Perhaps your build was in 64-bit mode and then you're not using the
right compiler?

-- 
Torbjörn