i686-pc-cygwin results

[Andreas Tobler <andreast-list at fgznet dot ch>]

Hi all,

I think we have hit already a similar topic:
Testresults of libffi-3.0.9rc10 for x64 mingw target

Here the results from:
Native configuration is i686-pc-cygwin (rc11)

                === libffi Summary ===

# of expected passes            1658
# of unexpected failures        1
# of expected failures          10


The failure I see is this one:

/home/andreast/libffi-3.0.9rc10/testsuite/libffi.call/cls_align_pointer.c:29: 
warning: format ?~@~X%lu?~@~Y expects type ?~@~Xlong unsigned int?~@~Y, 
but argument 3 has type ?~@~Xunsigned int?~@~Y^M
/home/andreast/libffi-3.0.9rc10/testsuite/libffi.call/cls_align_pointer.c:29: 
warning: format ?~@~X%lu?~@~Y expects type ?~@~Xlong unsigned int?~@~Y, 
but argument 6 has type ?~@~Xunsigned int?~@~Y^M
/home/andreast/libffi-3.0.9rc10/testsuite/libffi.call/cls_align_pointer.c:29: 
warning: format ?~@~X%lu?~@~Y expects type ?~@~Xlong unsigned int?~@~Y, 
but argument 9 has type ?~@~Xunsigned int?~

Sorry for the noise, but this is windows c&p.

In the inttypes.h the PRIuPTR is defined as "lu", when I override this 
to "u" the testcase passes.

The offending part is the void* in

typedef struct cls_struct_align {
  unsigned char a;
  void *b;
  unsigned char c;
} cls_struct_align;

It is 'unsigned int' iso. of 'long unsigned int' as expected by the 
printf modifier. I do not know who is correct nor how to 'fix' this.

Just for the record.
Andreas