Merge pull request #220 from mayawarrier/main

Add support for parsing numbers according to JSON format

include/fast_float/ascii_number.h

@@ -280,7 +280,7 @@ parsed_number_string_t<UC> parse_number_string(UC const *p, UC const * pend, par
   answer.too_many_digits = false;
   answer.negative = (*p == UC('-'));
 #ifdef FASTFLOAT_ALLOWS_LEADING_PLUS // disabled by default
-  if ((*p == UC('-')) || (*p == UC('+'))) {
+  if ((*p == UC('-')) || (!(fmt & FASTFLOAT_JSONFMT) && *p == UC('+'))) {
 #else
   if (*p == UC('-')) { // C++17 20.19.3.(7.1) explicitly forbids '+' sign here
 #endif
@@ -288,8 +288,14 @@ parsed_number_string_t<UC> parse_number_string(UC const *p, UC const * pend, par
     if (p == pend) {
       return answer;
     }
-    if (!is_integer(*p) && (*p != decimal_point)) { // a sign must be followed by an integer or the dot
-      return answer;
+    if (fmt & FASTFLOAT_JSONFMT) {
+      if (!is_integer(*p)) { // a sign must be followed by an integer
+        return answer;
+      }    
+    } else {
+      if (!is_integer(*p) && (*p != decimal_point)) { // a sign must be followed by an integer or the dot
+        return answer;
+      }
     }
   }
   UC const * const start_digits = p;
@@ -306,8 +312,16 @@ parsed_number_string_t<UC> parse_number_string(UC const *p, UC const * pend, par
   UC const * const end_of_integer_part = p;
   int64_t digit_count = int64_t(end_of_integer_part - start_digits);
   answer.integer = span<const UC>(start_digits, size_t(digit_count));
+  if (fmt & FASTFLOAT_JSONFMT) {
+    // at least 1 digit in integer part, without leading zeros
+    if (digit_count == 0 || (start_digits[0] == UC('0') && digit_count > 1)) {
+      return answer;
+    }
+  }
+
   int64_t exponent = 0;
-  if ((p != pend) && (*p == decimal_point)) {
+  const bool has_decimal_point = (p != pend) && (*p == decimal_point);
+  if (has_decimal_point) {
     ++p;
     UC const * before = p;
     // can occur at most twice without overflowing, but let it occur more, since
@@ -323,8 +337,13 @@ parsed_number_string_t<UC> parse_number_string(UC const *p, UC const * pend, par
     answer.fraction = span<const UC>(before, size_t(p - before));
     digit_count -= exponent;
   }
-  // we must have encountered at least one integer!
-  if (digit_count == 0) {
+  if (fmt & FASTFLOAT_JSONFMT) {
+    // at least 1 digit in fractional part
+    if (has_decimal_point && exponent == 0) {
+      return answer;
+    }
+  } 
+  else if (digit_count == 0) { // we must have encountered at least one integer!
     return answer;
   }
   int64_t exp_number = 0;            // explicit exponential part

include/fast_float/float_common.h

@@ -12,10 +12,15 @@
 
 namespace fast_float {
 
+#define FASTFLOAT_JSONFMT (1 << 5)
+
 enum chars_format {
   scientific = 1 << 0,
   fixed = 1 << 2,
   hex = 1 << 3,
+  no_infnan = 1 << 4,
+  json = FASTFLOAT_JSONFMT | fixed | scientific | no_infnan,
+  json_or_infnan = FASTFLOAT_JSONFMT | fixed | scientific,
   general = fixed | scientific
 };
 

include/fast_float/parse_number.h

@@ -164,7 +164,13 @@ from_chars_result_t<UC> from_chars_advanced(UC const * first, UC const * last,
   }
   parsed_number_string_t<UC> pns = parse_number_string<UC>(first, last, options);
   if (!pns.valid) {
-    return detail::parse_infnan(first, last, value);
+    if (options.format & chars_format::no_infnan) {
+      answer.ec = std::errc::invalid_argument;
+      answer.ptr = first;
+      return answer;
+    } else {
+      return detail::parse_infnan(first, last, value);
+    }
   }
 
   answer.ec = std::errc(); // be optimistic

tests/CMakeLists.txt

@@ -72,6 +72,8 @@ fast_float_add_cpp_test(long_test)
 fast_float_add_cpp_test(powersoffive_hardround)
 fast_float_add_cpp_test(string_test)
 
+fast_float_add_cpp_test(json_fmt)
+
 
 
 option(FASTFLOAT_EXHAUSTIVE "Exhaustive tests" OFF)

tests/json_fmt.cpp

@@ -0,0 +1,40 @@
+
+#include <cstdlib>
+#include <iostream>
+#include <vector>
+
+// test that this option is ignored
+#define FASTFLOAT_ALLOWS_LEADING_PLUS
+
+#include "fast_float/fast_float.h"
+
+int main()
+{
+  const std::vector<double> expected{ -0.2, 0.02, 0.002, 1., 0., std::numeric_limits<double>::infinity() };
+  const std::vector<std::string> accept{ "-0.2", "0.02", "0.002", "1e+0000", "0e-2", "inf" };
+  const std::vector<std::string> reject{ "-.2", "00.02", "0.e+1", "00.e+1", ".25", "+0.25", "inf", "nan(snan)" };
+
+  for (std::size_t i = 0; i < accept.size(); ++i)
+  {
+    const auto& f = accept[i];
+    double result;
+    auto answer = fast_float::from_chars(f.data(), f.data() + f.size(), result, fast_float::chars_format::json_or_infnan);
+    if (answer.ec != std::errc() || result != expected[i]) {
+      std::cerr << "json fmt rejected valid json " << f << std::endl;
+      return EXIT_FAILURE;
+    }
+  }
+
+  for (std::size_t i = 0; i < reject.size(); ++i)
+  {
+    const auto& f = reject[i];
+    double result;
+    auto answer = fast_float::from_chars(f.data(), f.data() + f.size(), result, fast_float::chars_format::json);
+    if (answer.ec == std::errc()) {
+      std::cerr << "json fmt accepted invalid json " << f << std::endl;
+      return EXIT_FAILURE;
+    }
+  }
+
+  return EXIT_SUCCESS;
+}