Mercurial > mxe-octave
comparison src/mingw-w64-2-float-h.patch @ 3554:b1ca68ddd02c
build native-gcc using untared dependancies within gcc source tree
* src/native-gcc.mk: untar gmp, mpc, mpfr, cloog and isl into the gcc source tree before configure and build, set sysroot.
* src/native-binutils.mk: set sysroot path.
* src/mingw-w64-1-math-h.patch, mingw-w64-2-float-h.patch: New files.
| author | John Donoghue <john.donoghue@ieee.org> |
|---|---|
| date | Fri, 21 Feb 2014 06:05:39 -0500 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| 3553:c27b154dab6c | 3554:b1ca68ddd02c |
|---|---|
| 1 This file is part of MXE. | |
| 2 See index.html for further information. | |
| 3 | |
| 4 workaround for -isystem flag messing up include order | |
| 5 https://bugzilla.redhat.com/show_bug.cgi?id=843436 | |
| 6 | |
| 7 diff -ur a/mingw-w64-headers/crt/float.h b/mingw-w64-headers/crt/float.h | |
| 8 --- a/mingw-w64-headers/crt/float.h 2014-01-11 21:38:50.640418500 +1100 | |
| 9 +++ b/mingw-w64-headers/crt/float.h 2014-01-11 21:44:08.528393903 +1100 | |
| 10 @@ -105,6 +105,15 @@ | |
| 11 #define DBL_MAX_10_EXP __DBL_MAX_10_EXP__ | |
| 12 #define LDBL_MAX_10_EXP __LDBL_MAX_10_EXP__ | |
| 13 | |
| 14 + /* The difference between 1 and the least value greater than 1 that is | |
| 15 + representable in the given floating point type, b**1-p. */ | |
| 16 + #undef FLT_EPSILON | |
| 17 + #undef DBL_EPSILON | |
| 18 + #undef LDBL_EPSILON | |
| 19 + #define FLT_EPSILON __FLT_EPSILON__ | |
| 20 + #define DBL_EPSILON __DBL_EPSILON__ | |
| 21 + #define LDBL_EPSILON __LDBL_EPSILON__ | |
| 22 + | |
| 23 /* Addition rounds to 0: zero, 1: nearest, 2: +inf, 3: -inf, -1: unknown. */ | |
| 24 /* ??? This is supposed to change with calls to fesetround in <fenv.h>. */ | |
| 25 #undef FLT_ROUNDS |