add support for unsigned long long

Want to be able to format the following without any casting

std::cout << fmt::str(fmt::Format("{:x}") << 0xAEull) << std::endl;
std::cout << fmt::str(fmt::Format("{:x}") << 0xAEul) << std::endl;
std::cout << fmt::str(fmt::Format("{:x}") << uint64_t(0xae)) <<
std::endl;

format.cc

@@ -400,7 +400,7 @@ void fmt::BasicFormatter<Char>::CheckSign(const Char *&s, const Arg &arg) {
     ReportError(s,
         Format("format specifier '{0}' requires numeric argument") << *s);
   }
-  if (arg.type == UINT || arg.type == ULONG) {
+  if (arg.type == UINT || arg.type == ULONG || arg.type == ULLONG) {
     ReportError(s,
         Format("format specifier '{0}' requires signed argument") << *s);
   }
@@ -519,7 +519,7 @@ void fmt::BasicFormatter<Char>::DoFormat() {
           ++s;
           ++num_open_braces_;
           const Arg &precision_arg = ParseArgIndex(s);
-          unsigned long value = 0;
+          unsigned long long value = 0;
           switch (precision_arg.type) {
           case INT:
             if (precision_arg.int_value < 0)
@@ -537,12 +537,15 @@ void fmt::BasicFormatter<Char>::DoFormat() {
           case ULONG:
             value = precision_arg.ulong_value;
             break;
+          case ULLONG:
+            value = precision_arg.ulong_long_value;
+            break;            
           default:
             ReportError(s, "precision is not integer");
           }
           if (value > INT_MAX)
             ReportError(s, "number is too big in format");
-          precision = value;
+          precision = static_cast<int>(value);
           if (*s++ != '}')
             throw FormatError("unmatched '{' in format");
           --num_open_braces_;
@@ -578,6 +581,9 @@ void fmt::BasicFormatter<Char>::DoFormat() {
     case ULONG:
       writer.FormatInt(arg.ulong_value, spec);
       break;
+    case ULLONG:
+      writer.FormatInt(arg.ulong_long_value, spec);
+      break;      
     case DOUBLE:
       writer.FormatDouble(arg.double_value, spec, precision);
       break;

format.h

@@ -189,6 +189,9 @@ struct SignedIntTraits {
 template <>
 struct IntTraits<int> : SignedIntTraits<int, unsigned> {};
 
+template <>
+struct IntTraits<uint32_t> : SignedIntTraits<uint32_t, unsigned> {};
+
 template <>
 struct IntTraits<long> : SignedIntTraits<long, unsigned long> {};
 
@@ -444,6 +447,7 @@ DEFINE_INT_FORMATTERS(int)
 DEFINE_INT_FORMATTERS(long)
 DEFINE_INT_FORMATTERS(unsigned)
 DEFINE_INT_FORMATTERS(unsigned long)
+DEFINE_INT_FORMATTERS(unsigned long long)
 
 template <typename Char>
 class BasicFormatter;
@@ -811,7 +815,7 @@ class BasicFormatter {
 
   enum Type {
     // Numeric types should go first.
-    INT, UINT, LONG, ULONG, DOUBLE, LONG_DOUBLE,
+    INT, UINT, LONG, ULONG, ULLONG, DOUBLE, LONG_DOUBLE,
     LAST_NUMERIC_TYPE = LONG_DOUBLE,
     CHAR, STRING, WSTRING, POINTER, CUSTOM
   };
@@ -846,6 +850,7 @@ class BasicFormatter {
       double double_value;
       long long_value;
       unsigned long ulong_value;
+      unsigned long long ulong_long_value;
       long double long_double_value;
       const void *pointer_value;
       struct {
@@ -865,6 +870,7 @@ class BasicFormatter {
     Arg(unsigned value) : type(UINT), uint_value(value), formatter(0) {}
     Arg(long value) : type(LONG), long_value(value), formatter(0) {}
     Arg(unsigned long value) : type(ULONG), ulong_value(value), formatter(0) {}
+    Arg(unsigned long long value) : type(ULLONG), ulong_long_value(value), formatter(0) {}
     Arg(float value) : type(DOUBLE), double_value(value), formatter(0) {}
     Arg(double value) : type(DOUBLE), double_value(value), formatter(0) {}
     Arg(long double value)
@@ -1148,7 +1154,7 @@ class FormatInt {
   enum {BUFFER_SIZE = std::numeric_limits<uint64_t>::digits10 + 3};
   char buffer_[BUFFER_SIZE];
   char *str_;
-
+ 
   // Formats value in reverse and returns the number of digits.
   char *FormatDecimal(uint64_t value) {
     char *buffer_end = buffer_ + BUFFER_SIZE;
@@ -1183,6 +1189,7 @@ class FormatInt {
       *--str_ = '-';
   }
   explicit FormatInt(unsigned value) : str_(FormatDecimal(value)) {}
+  explicit FormatInt(uint64_t value) : str_(FormatDecimal(value)) {}
 
   const char *c_str() const { return str_; }
   std::string str() const { return str_; }