Mercurial > octave-libgccjit

comparison oct-conf-post.in.h @ 17993:ac9fd5010620

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
avoid including gnulib header in installed Octave header file (bug #40607) * oct-conf-post.in.h (OCTAVE_ENSURE_LONG_DOUBLE_OPERATIONS_ARE_NOT_TRUNCATED): New macro. * oct-inttypes.h, oct-inttypes.cc: Don't include fpucw.h from gnulib or the macros it defines. In functions that need to ensure that long double operations are not truncated to double precision, use external functions.
author John W. Eaton <jwe@octave.org>
date Sat, 23 Nov 2013 20:28:46 -0500
parents d63878346099
children 79653c5b6147
comparison
equal deleted inserted replaced
17992:b8cd17571a1d 17993:ac9fd5010620
105 105
106 /* To be able to use long doubles for 64-bit mixed arithmetics, we need 106 /* To be able to use long doubles for 64-bit mixed arithmetics, we need
107 them at least 80 bits wide and we need roundl declared in math.h. 107 them at least 80 bits wide and we need roundl declared in math.h.
108 FIXME: Maybe substitute this by a more precise check in the future? */ 108 FIXME: Maybe substitute this by a more precise check in the future? */
109 #if (SIZEOF_LONG_DOUBLE >= 10) && defined (HAVE_ROUNDL) 109 #if (SIZEOF_LONG_DOUBLE >= 10) && defined (HAVE_ROUNDL)
110 #define OCTAVE_INT_USE_LONG_DOUBLE 110 # define OCTAVE_INT_USE_LONG_DOUBLE
111 # if (SIZEOF_LONG_DOUBLE == 10 \
112 && (defined __i386__ || defined __x86_64__) && defined __GNUC__)
113 #define OCTAVE_ENSURE_LONG_DOUBLE_OPERATIONS_ARE_NOT_TRUNCATED 1
114 # endif
111 #endif 115 #endif
112 116
113 #define OCTAVE_EMPTY_CPP_ARG 117 #define OCTAVE_EMPTY_CPP_ARG
114 118
115 /* oct-dlldefs.h */ 119 /* oct-dlldefs.h */