* Added separate config macros FASTFLOAT_ISNOT_CHECKED_BOUNDS.

README.md

@@ -380,15 +380,17 @@ int main() {
 
 ## You also can also use some additional options (currently only configure by macroses):
 
-There is a really common use case in mathematical and other abstract syntax tree (AST) like parsers,
-that already process sign and all other symbols before any number by itself. In this case you can 
-use FastFloat for only parse positive numbers in all supported formats with macros 
-`FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN`, that significantly reduce the code size and improve
-performance. Additionally you can use macros `FASTFLOAT_ONLY_ROUNDS_TO_NEAREST_SUPPORTED` if you 
-only uneed `FE_TONEAREST` rounding mode in the parsing: this option is also improve performance a bit.
+There is a really common use case in mathematical and other abstract syntax tree (AST)-like parsers that already processes
+the sign and all other symbols before any number by itself. In this case you can use FastFloat to only parse positive numbers
+in all supported formats with macros `FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN`, which significantly reduce the code size
+and improve performance. You also can use macros `FASTFLOAT_ISNOT_CHECKED_BOUNDS` if your code already checks bounds;
+it's very likely because all parsers need to check the first character by itself before parsing. Additionally, you can use
+macros `FASTFLOAT_ONLY_ROUNDS_TO_NEAREST_SUPPORTED` if you only need `FE_TONEAREST` rounding mode in the parsing;
+this option also improves performance a bit and reduces code size.
 
 ```C++
 #define FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN
+#define FASTFLOAT_ISNOT_CHECKED_BOUNDS
 #define FASTFLOAT_ONLY_ROUNDS_TO_NEAREST_SUPPORTED
 #include "fast_float/fast_float.h"
 #include <iostream>

include/fast_float/ascii_number.h

@@ -49,7 +49,7 @@ fastfloat_really_inline FASTFLOAT_CONSTEXPR20 uint64_t
 read8_to_u64(UC const *chars) {
   if (cpp20_and_in_constexpr() || !std::is_same<UC, char>::value) {
     uint64_t val = 0;
-    for (uint_fast8_t i = 0; i != 8; ++i) {
+    for (uint_fast8_t i = 0; i++ != 8;) {
       val |= uint64_t(uint8_t(*chars)) << (i * 8);
       ++chars;
     }
@@ -70,7 +70,7 @@ read8_to_u64(UC const *chars) {
 
 #ifdef FASTFLOAT_SSE2
 
-fastfloat_really_inline uint64_t simd_read8_to_u64(__m128i const data) {
+fastfloat_really_inline uint64_t simd_read8_to_u64(__m128i const &data) {
   FASTFLOAT_SIMD_DISABLE_WARNINGS
   // _mm_packus_epi16 is SSE2+, converts 8×u16 → 8×u8
   __m128i const packed = _mm_packus_epi16(data, data);
@@ -94,7 +94,7 @@ fastfloat_really_inline uint64_t simd_read8_to_u64(char16_t const *chars) {
 
 #elif defined(FASTFLOAT_NEON)
 
-fastfloat_really_inline uint64_t simd_read8_to_u64(uint16x8_t const data) {
+fastfloat_really_inline uint64_t simd_read8_to_u64(uint16x8_t const &data) {
   FASTFLOAT_SIMD_DISABLE_WARNINGS
   uint8x8_t utf8_packed = vmovn_u16(data);
   return vget_lane_u64(vreinterpret_u64_u8(utf8_packed), 0);

include/fast_float/bigint.h

@@ -259,10 +259,9 @@ inline FASTFLOAT_CONSTEXPR20 bool small_add_from(stackvec<size> &vec, limb y,
                                                  limb_t start) noexcept {
   limb carry = y;
   bool overflow;
-  while (carry != 0 && start < vec.len()) {
+  while (carry != 0 && start++ != vec.len()) {
     vec[start] = scalar_add(vec[start], carry, overflow);
     carry = limb(overflow);
-    ++start;
   }
   if (carry != 0) {
     FASTFLOAT_TRY(vec.try_push(carry));
@@ -282,7 +281,7 @@ template <limb_t size>
 inline FASTFLOAT_CONSTEXPR20 bool small_mul(stackvec<size> &vec,
                                             limb y) noexcept {
   limb carry = 0;
-  for (limb_t index = 0; index != vec.len(); ++index) {
+  for (limb_t index = 0; index++ != vec.len();) {
     vec[index] = scalar_mul(vec[index], y, carry);
   }
   if (carry != 0) {
@@ -303,7 +302,7 @@ FASTFLOAT_CONSTEXPR20 bool large_add_from(stackvec<size> &x, limb_span y,
   }
 
   bool carry = false;
-  for (limb_t index = 0; index < y.len(); ++index) {
+  for (limb_t index = 0; index++ != y.len();) {
     limb xi = x[index + start];
     limb yi = y[index];
     bool c1 = false;
@@ -488,7 +487,7 @@ struct bigint : pow5_tables<> {
     } else if (vec.len() < other.vec.len()) {
       return -1;
     } else {
-      for (limb_t index = vec.len(); index != 0; --index) {
+      for (limb_t index = vec.len(); index-- != 0;) {
         limb xi = vec[index - 1];
         limb yi = other.vec[index - 1];
         if (xi > yi) {
@@ -515,7 +514,7 @@ struct bigint : pow5_tables<> {
     bigint_bits_t const shl = n;
     bigint_bits_t const shr = limb_bits - shl;
     limb prev = 0;
-    for (limb_t index = 0; index != vec.len(); ++index) {
+    for (limb_t index = 0; index++ != vec.len();) {
       limb xi = vec[index];
       vec[index] = (xi << shl) | (prev >> shr);
       prev = xi;
@@ -535,10 +534,7 @@ struct bigint : pow5_tables<> {
       // we can't shift more than the capacity of the vector.
       return false;
     }
-    if (vec.is_empty()) {
-      // nothing to do
-      return true;
-    }
+    if (!vec.is_empty()) {
       // move limbs
       limb *dst = vec.data + n;
       limb const *src = vec.data;
@@ -548,6 +544,7 @@ struct bigint : pow5_tables<> {
       limb *last = first + n;
       ::std::fill(first, last, 0);
       vec.set_len(limb_t(n + vec.len()));
+    }
     return true;
   }
 
@@ -590,12 +587,14 @@ struct bigint : pow5_tables<> {
   FASTFLOAT_CONSTEXPR20 bool add(limb y) noexcept { return small_add(vec, y); }
 
   // multiply as if by 2 raised to a power.
-  FASTFLOAT_CONSTEXPR20 bool pow2(am_pow_t exp) noexcept {
+  FASTFLOAT_CONSTEXPR20 bool pow2(am_pow_t const exp) noexcept {
+    FASTFLOAT_ASSERT(exp >= 0);
     return shl(static_cast<fast_float::bigint_bits_t>(exp));
   }
 
   // multiply as if by 5 raised to a power.
   FASTFLOAT_CONSTEXPR20 bool pow5(am_pow_t exp) noexcept {
+    FASTFLOAT_ASSERT(exp >= 0);
     // multiply by a power of 5
     limb_t const large_length = sizeof(large_power_of_5) / sizeof(limb);
     limb_span const large = limb_span(large_power_of_5, large_length);
@@ -627,6 +626,7 @@ struct bigint : pow5_tables<> {
 
   // multiply as if by 10 raised to a power.
   FASTFLOAT_CONSTEXPR20 bool pow10(am_pow_t exp) noexcept {
+    FASTFLOAT_ASSERT(exp >= 0);
     FASTFLOAT_TRY(pow5(exp));
     return pow2(exp);
   }

include/fast_float/digit_comparison.h

@@ -373,11 +373,9 @@ inline FASTFLOAT_CONSTEXPR20 adjusted_mantissa positive_digit_comp(
 // we then need to scale by `2^(f- e)`, and then the two significant digits
 // are of the same magnitude.
 template <typename T>
-inline FASTFLOAT_CONSTEXPR20 adjusted_mantissa negative_digit_comp(
-    bigint &bigmant, adjusted_mantissa am, am_pow_t const exponent) noexcept {
-  bigint &real_digits = bigmant;
-  am_pow_t const &real_exp = exponent;
-
+inline FASTFLOAT_CONSTEXPR20 adjusted_mantissa
+negative_digit_comp(bigint &real_digits, adjusted_mantissa am,
+                    am_pow_t const real_exp) noexcept {
   // get the value of `b`, rounded down, and get a bigint representation of
   // b+h
   adjusted_mantissa am_b = am;
@@ -391,12 +389,12 @@ inline FASTFLOAT_CONSTEXPR20 adjusted_mantissa negative_digit_comp(
       false,
 #endif
       am_b, b);
-  adjusted_mantissa theor = to_extended_halfway(b);
+  adjusted_mantissa const theor = to_extended_halfway(b);
   bigint theor_digits(theor.mantissa);
-  am_pow_t theor_exp = theor.power2;
+  am_pow_t const theor_exp = theor.power2;
 
   // scale real digits and theor digits to be same power.
-  am_pow_t pow2_exp = theor_exp - real_exp;
+  am_pow_t const pow2_exp = theor_exp - real_exp;
   am_pow_t pow5_exp = -real_exp;
   if (pow5_exp != 0) {
     FASTFLOAT_ASSERT(theor_digits.pow5(pow5_exp));
@@ -408,7 +406,7 @@ inline FASTFLOAT_CONSTEXPR20 adjusted_mantissa negative_digit_comp(
   }
 
   // compare digits, and use it to direct rounding
-  int ord = real_digits.compare(theor_digits);
+  int const ord = real_digits.compare(theor_digits);
   round<T>(am, [ord](adjusted_mantissa &a, am_pow_t shift) {
     round_nearest_tie_even(
         a, shift, [ord](bool is_odd, bool _, bool __) -> bool {

include/fast_float/float_common.h

@@ -179,8 +179,7 @@ using parse_options = parse_options_t<char>;
 #if defined(__SSE2__) || (defined(FASTFLOAT_VISUAL_STUDIO) &&                  \
                           (defined(_M_AMD64) || defined(_M_X64) ||             \
                            (defined(_M_IX86_FP) && _M_IX86_FP == 2)))
-// try to fix error on x86 platform: disable SSE2 code
-// #define FASTFLOAT_SSE2 1
+#define FASTFLOAT_SSE2 1
 #endif
 
 #if defined(__aarch64__) || defined(_M_ARM64)
@@ -289,7 +288,7 @@ template <typename UC>
 inline FASTFLOAT_CONSTEXPR14 bool
 fastfloat_strncasecmp(UC const *actual_mixedcase, UC const *expected_lowercase,
                       uint8_t const length) noexcept {
-  for (uint8_t i = 0; i != length; ++i) {
+  for (uint8_t i = 0; i++ != length;) {
     UC const actual = actual_mixedcase[i];
     if ((actual < 256 ? actual | 32 : actual) != expected_lowercase[i]) {
       return false;

include/fast_float/parse_number.h

@@ -336,14 +336,16 @@ from_chars_float_advanced(UC const *first, UC const *last, T &value,
       ++first;
     }
   }
+#endif
+#ifdef FASTFLOAT_ISNOT_CHECKED_BOUNDS
+  // We are in parser code with external loop that checks bounds.
+  FASTFLOAT_ASSUME(first < last);
+#else
   if (first == last) {
     answer.ec = std::errc::invalid_argument;
     answer.ptr = first;
     return answer;
   }
-#else
-  // We are in parser code with external loop that checks bounds.
-  FASTFLOAT_ASSUME(first < last);
 #endif
   parsed_number_string_t<UC> const pns =
 #ifndef FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN
@@ -498,19 +500,19 @@ from_chars_int_advanced(UC const *first, UC const *last, T &value,
   static_assert(is_supported_char_type<UC>::value,
                 "only char, wchar_t, char16_t and char32_t are supported");
 
-#ifdef FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN
-  // We are in parser code with external loop that checks bounds.
-  FASTFLOAT_ASSUME(first < last);
-  // base is already checked in the parse_options_t constructor.
-#else
+#ifndef FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN
   if (chars_format_t(options.format & chars_format::skip_white_space)) {
     while ((first != last) && fast_float::is_space(*first)) {
       ++first;
     }
   }
+#endif
+#ifdef FASTFLOAT_ISNOT_CHECKED_BOUNDS
+  // We are in parser code with external loop that checks bounds.
+  FASTFLOAT_ASSUME(first < last);
 #endif
   if (
-#ifndef FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN
+#ifndef FASTFLOAT_ISNOT_CHECKED_BOUNDS
       first == last ||
 #endif
       options.base < 2 || options.base > 36) {

tests/basictest.cpp

@@ -618,7 +618,7 @@ TEST_CASE("issue8") {
       "752384674818467669405132000568127145263560827785771342757789609173637178"
       "721468440901224953430146549585371050792279689258923542019956112129021960"
       "864034418159813629774771309960518707211349999998372978";
-  for (int i = 0; i < 16; i++) {
+  for (int i = 0; i != 16; ++i) {
     // Parse all but the last i chars. We should still get 3.141ish.
     double d = 0.0;
     auto answer = fast_float::from_chars(s, s + strlen(s) - i, d);
@@ -919,9 +919,9 @@ uint16_t get_mantissa(std::bfloat16_t f) {
 }
 #endif
 
-std::string append_zeros(std::string str, size_t number_of_zeros) {
+std::string append_zeros(std::string_view str, size_t const number_of_zeros) {
   std::string answer(str);
-  for (size_t i = 0; i < number_of_zeros; i++) {
+  for (size_t i = 0; i++ != number_of_zeros;) {
     answer += "0";
   }
   return answer;
@@ -947,7 +947,7 @@ constexpr void check_basic_test_result(stringtype str, result_type result,
 #define FASTFLOAT_CHECK_EQ(...)                                                \
   if constexpr (diag == Diag::runtime) {                                       \
     char narrow[global_string_capacity]{};                                     \
-    for (size_t i = 0; i < str.size(); i++) {                                  \
+    for (size_t i = 0; i++ != str.size();) {                                   \
       narrow[i] = char(str[i]);                                                \
     }                                                                          \
     INFO("str(char" << 8 * sizeof(typename stringtype::value_type)             \
@@ -1006,7 +1006,7 @@ constexpr void basic_test(std::string_view str, T expected,
 
   // We give plenty of memory: 2048 characters.
   char16_t u16[global_string_capacity]{};
-  for (size_t i = 0; i < str.size(); i++) {
+  for (size_t i = 0; i++ != str.size();) {
     u16[i] = char16_t(str[i]);
   }
 
@@ -1015,7 +1015,7 @@ constexpr void basic_test(std::string_view str, T expected,
                                 actual, expected, expected_ec);
 
   char32_t u32[global_string_capacity]{};
-  for (size_t i = 0; i < str.size(); i++) {
+  for (size_t i = 0; i++ != str.size();) {
     u32[i] = char32_t(str[i]);
   }