[muscle] Re: muscle5.71,

  • From: Jeremy Friesner <jfriesne@xxxxxxxxx>
  • To: muscle@xxxxxxxxxxxxx
  • Date: Tue, 12 Jun 2012 10:47:00 -0700

Hi Gilles,

I think I have a patch to fix this problem.  Here is the patch -- let me know 
if it doesn't work for you.

Index: support/MuscleSupport.h
===================================================================
--- support/MuscleSupport.h     (revision 10293)
+++ support/MuscleSupport.h     (working copy)
@@ -603,7 +603,7 @@
       mov val, ax;
    };
 #else
-   __asm__ ("xchgb %b0,%h0" : "=q" (val) : "0" (val));
+   __asm__ ("xchgb %b0,%h0" : "=Q" (val) : "0" (val));
 #endif
    return val;
 }

Thanks,
Jeremy

On Jun 12, 2012, at 8:50 AM, Gilles J. Seguin wrote:

> fedora 17, gcc-4.7.0-5.fc17.x86_64
> 
> g++ -I.. -DMUSCLE_SINGLE_THREAD_ONLY -DMUSCLE_ENABLE_ZLIB_ENCODING \
>  -fno-exceptions -DMUSCLE_NO_EXCEPTIONS -W -Wall -Wno-multichar  -O3 \
>  -DMUSCLE_ENABLE_MEMORY_TRACKING    -c -o ByteBuffer.o \
>  ../util/ByteBuffer.cpp
> 
> ../util/ByteBuffer.cpp: In member function ‘uint32 \
>   muscle::ByteBuffer::ReadInt16s(int16*, uint32, uint32&) const’: \
>   ../util/ByteBuffer.cpp:235:1: error: extended registers have no \
>   high halves
> make: *** [ByteBuffer.o] Error 1
> 
> gilles
> 
> 
> 


Other related posts: