Merge tag 'v0.11.0-beta1' into release-0.11.0

The merge of the MinGW Port updates accidentally didn't make
it into the GitHub version of the release branch, even though
the tag was correct.

dmd2/expression.c

@@ -2567,7 +2567,7 @@ void floatToBuffer(OutBuffer *buf, Type *type, real_t value)
     real_t r = strtold(buffer, NULL);
     __locale_decpoint = save;
 #else
-    real_t r = strtold(buffer, NULL);
+    real_t r = Port::strtold(buffer, NULL);
 #endif
 #if IN_LLVM
     if (r == value)                     // if exact duplication

dmd2/lexer.c

@@ -2365,7 +2365,7 @@ done:
 #ifdef IN_GCC
     t->float80value = real_t::parse((char *)stringbuffer.data, real_t::LongDouble);
 #else
-    t->float80value = strtold((char *)stringbuffer.data, NULL);
+    t->float80value = Port::strtold((char *)stringbuffer.data, NULL);
 #endif
     errno = 0;
     float strtofres;

dmd2/root/longdouble.h

@@ -34,6 +34,12 @@ typedef volatile long double volatile_longdouble;
 // template<typename T> longdouble ldouble(T x) { return (longdouble) x; }
 #define ldouble(x) ((longdouble)(x))
 
+#if __MINGW32__
+// MinGW supports 80 bit reals, but the formatting functions map to versions
+// from the MSVC runtime by default which don't.
+#define sprintf __mingw_sprintf
+#endif
+
 inline size_t ld_sprint(char* str, int fmt, longdouble x)
 {
     char sfmt[4] = "%Lg";
@@ -41,6 +47,10 @@ inline size_t ld_sprint(char* str, int fmt, longdouble x)
     return sprintf(str, sfmt, x);
 }
 
+#if __MINGW32__
+#undef sprintf
+#endif
+
 #else
 
 #include <float.h>

dmd2/root/port.c

@@ -348,9 +348,201 @@ int Port::stricmp(const char *s1, const char *s2)
     return ::stricmp(s1, s2);
 }
 
+// See vcbuild/strtold.c.
+longdouble strtold(const char *p, char **endp);
+
+longdouble Port::strtold(const char *p, char **endp)
+{
+    return ::strtold(p, endp);
+}
+
 #endif
 
-#if linux || __APPLE__ || __FreeBSD__ || __OpenBSD__ || __HAIKU__ || __MINGW32__
+#if __MINGW32__
+
+#include <math.h>
+#include <time.h>
+#include <sys/time.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <ctype.h>
+#include <wchar.h>
+#include <float.h>
+#include <assert.h>
+
+static double zero = 0;
+double Port::nan = copysign(NAN, 1.0);
+double Port::infinity = 1 / zero;
+double Port::dbl_max = 1.7976931348623157e308;
+double Port::dbl_min = 5e-324;
+longdouble Port::ldbl_max = LDBL_MAX;
+
+struct PortInitializer
+{
+    PortInitializer();
+};
+
+static PortInitializer portinitializer;
+
+PortInitializer::PortInitializer()
+{
+    assert(!signbit(Port::nan));
+}
+
+int Port::isNan(double r)
+{
+    return isnan(r);
+}
+
+int Port::isNan(longdouble r)
+{
+    return isnan(r);
+}
+
+int Port::isSignallingNan(double r)
+{
+    /* A signalling NaN is a NaN with 0 as the most significant bit of
+     * its significand, which is bit 51 of 0..63 for 64 bit doubles.
+     */
+    return isNan(r) && !((((unsigned char*)&r)[6]) & 8);
+}
+
+int Port::isSignallingNan(longdouble r)
+{
+    /* A signalling NaN is a NaN with 0 as the most significant bit of
+     * its significand, which is bit 62 of 0..79 for 80 bit reals.
+     */
+    return isNan(r) && !((((unsigned char*)&r)[7]) & 0x40);
+}
+
+int Port::isFinite(double r)
+{
+    return ::finite(r);
+}
+
+int Port::isInfinity(double r)
+{
+    return isinf(r);
+}
+
+int Port::Signbit(double r)
+{
+    union { double d; long long ll; } u;
+    u.d =  r;
+    return u.ll < 0;
+}
+
+double Port::floor(double d)
+{
+    return ::floor(d);
+}
+
+double Port::pow(double x, double y)
+{
+    return ::pow(x, y);
+}
+
+longdouble Port::fmodl(longdouble x, longdouble y)
+{
+    return ::fmodl(x, y);
+}
+
+unsigned long long Port::strtoull(const char *p, char **pend, int base)
+{
+    return ::strtoull(p, pend, base);
+}
+
+char *Port::ull_to_string(char *buffer, ulonglong ull)
+{
+    sprintf(buffer, "%llu", ull);
+    return buffer;
+}
+
+wchar_t *Port::ull_to_string(wchar_t *buffer, ulonglong ull)
+{
+    swprintf(buffer, sizeof(ulonglong) * 3 + 1, L"%llu", ull);
+    return buffer;
+}
+
+double Port::ull_to_double(ulonglong ull)
+{
+    return (double) ull;
+}
+
+const char *Port::list_separator()
+{
+    return ",";
+}
+
+const wchar_t *Port::wlist_separator()
+{
+    return L",";
+}
+
+char *Port::strupr(char *s)
+{
+    char *t = s;
+
+    while (*s)
+    {
+        *s = toupper(*s);
+        s++;
+    }
+
+    return t;
+}
+
+int Port::memicmp(const char *s1, const char *s2, int n)
+{
+    int result = 0;
+
+    for (int i = 0; i < n; i++)
+    {   char c1 = s1[i];
+        char c2 = s2[i];
+
+        result = c1 - c2;
+        if (result)
+        {
+            result = toupper(c1) - toupper(c2);
+            if (result)
+                break;
+        }
+    }
+    return result;
+}
+
+int Port::stricmp(const char *s1, const char *s2)
+{
+    int result = 0;
+
+    for (;;)
+    {   char c1 = *s1;
+        char c2 = *s2;
+
+        result = c1 - c2;
+        if (result)
+        {
+            result = toupper(c1) - toupper(c2);
+            if (result)
+                break;
+        }
+        if (!c1)
+            break;
+        s1++;
+        s2++;
+    }
+    return result;
+}
+
+longdouble Port::strtold(const char *p, char **endp)
+{
+    return ::__mingw_strtold(p, endp);
+}
+
+#endif
+
+#if linux || __APPLE__ || __FreeBSD__ || __OpenBSD__ || __HAIKU__
 
 #include <math.h>
 #if linux
@@ -408,7 +600,7 @@ int Port::isNan(double r)
 #else
     return __inline_isnan(r);
 #endif
-#elif __HAIKU__ || __OpenBSD__ || __MINGW32__
+#elif __HAIKU__ || __OpenBSD__
     return isnan(r);
 #else
     #undef isnan
@@ -424,7 +616,7 @@ int Port::isNan(longdouble r)
 #else
     return __inline_isnan(r);
 #endif
-#elif __HAIKU__ || __OpenBSD__ || __MINGW32__
+#elif __HAIKU__ || __OpenBSD__
     return isnan(r);
 #else
     #undef isnan
@@ -460,7 +652,7 @@ int Port::isInfinity(double r)
 {
 #if __APPLE__
     return fpclassify(r) == FP_INFINITE;
-#elif defined __HAIKU__ || __OpenBSD__ || __MINGW32__
+#elif defined __HAIKU__ || __OpenBSD__
     return isinf(r);
 #else
     #undef isinf
@@ -586,6 +778,11 @@ int Port::stricmp(const char *s1, const char *s2)
     return result;
 }
 
+longdouble Port::strtold(const char *p, char **endp)
+{
+    return ::strtold(p, endp);
+}
+
 #endif
 
 #if __sun
@@ -731,4 +928,9 @@ char *Port::strupr(char *s)
     return t;
 }
 
+longdouble Port::strtold(const char *p, char **endp)
+{
+    return ::strtold(p, endp);
+}
+
 #endif

dmd2/root/port.h

@@ -19,8 +19,6 @@
 #if _MSC_VER
 #include <float.h>  // for _isnan
 #include <malloc.h> // for alloca
-// According to VC 8.0 docs, long double is the same as double
-longdouble strtold(const char *p,char **endp);
 #define strtof  strtod
 #define isnan   _isnan
 
@@ -41,14 +39,6 @@ struct Port
     static double dbl_min;
     static longdouble ldbl_max;
 
-#if !defined __HAIKU__ || __OpenBSD__
-#elif __GNUC__
-    // These conflict with macros in math.h, should rename them
-    #undef isnan
-    #undef isfinite
-    #undef isinfinity
-    #undef signbit
-#endif
     static int isNan(double);
     static int isNan(longdouble);
 
@@ -80,6 +70,8 @@ struct Port
 
     static int memicmp(const char *s1, const char *s2, int n);
     static int stricmp(const char *s1, const char *s2);
+
+    static longdouble strtold(const char *p, char **endp);
 };
 
 #endif