Web lists-archives.com

Re: Request update GCC




On 27/06/2017 03:08, JonY wrote:
On 06/26/2017 10:16 AM, Marco Atzeri wrote:

Jon,

are you building the compilers on cygwin or cross compiling ?

I was trying to test the patch for
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=47030
as it is blocking PETSC porting,
but I was unable to build 6.3.0-1 from your
source package.


In file included from
/cygdrive/e/cyg_pub/devel/gcc/gcc-6.3.0-1.src/gcc-6.3.0-1.
x86_64/src/gcc-6.3.0/gcc/rtl.h:31:0,
                 from
/cygdrive/e/cyg_pub/devel/gcc/gcc-6.3.0-1.src/gcc-6.3.0-1.
x86_64/src/gcc-6.3.0/gcc/genopinit.c:25:
/cygdrive/e/cyg_pub/devel/gcc/gcc-6.3.0-1.src/gcc-6.3.0-1.x86_64/src/gcc-6.3.0/g

cc/real.h:36:35: warning: division by zero [-Wdiv-by-zero]
 #define SIGSZ   (SIGNIFICAND_BITS / HOST_BITS_PER_LONG)

/cygdrive/e/cyg_pub/devel/gcc/gcc-6.3.0-1.src/gcc-6.3.0-1.x86_64/src/gcc-6.3.0/g

cc/real.h:49:21: note: in expansion of macro ‘SIGSZ’
   unsigned long sig[SIGSZ];

On Cygwin itself, shouldn't that be 32bit and 64bit respectively for Cygwin?

In theory

gcc/real.h:#define SIGNIFICAND_BITS   (128 + HOST_BITS_PER_LONG)

gcc/hwint.h:#define HOST_BITS_PER_LONG  (CHAR_BIT * SIZEOF_LONG)


but I do not understand why SIGSZ is wrong defined.






--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple