8.2.9

reject non-digit wide code units in uint8/uint16 integer fast path

CMakeLists.txt

@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.14)
 
 
-project(fast_float VERSION 8.2.8 LANGUAGES CXX)
+project(fast_float VERSION 8.2.9 LANGUAGES CXX)
 set(FASTFLOAT_CXX_STANDARD 11 CACHE STRING "the C++ standard to use for fastfloat")
 set(CMAKE_CXX_STANDARD ${FASTFLOAT_CXX_STANDARD})
 option(FASTFLOAT_TEST "Enable tests" OFF)

README.md

@@ -531,7 +531,7 @@ sufficiently recent version of CMake (3.11 or better at least):
 FetchContent_Declare(
   fast_float
   GIT_REPOSITORY https://github.com/fastfloat/fast_float.git
-  GIT_TAG tags/v8.2.8
+  GIT_TAG tags/v8.2.9
   GIT_SHALLOW TRUE)
 
 FetchContent_MakeAvailable(fast_float)
@@ -547,7 +547,7 @@ You may also use [CPM](https://github.com/cpm-cmake/CPM.cmake), like so:
 CPMAddPackage(
   NAME fast_float
   GITHUB_REPOSITORY "fastfloat/fast_float"
-  GIT_TAG v8.2.8)
+  GIT_TAG v8.2.9)
 ```
 
 ## Using as single header
@@ -559,7 +559,7 @@ if desired as described in the command line help.
 
 You may directly download automatically generated single-header files:
 
-<https://github.com/fastfloat/fast_float/releases/download/v8.2.8/fast_float.h>
+<https://github.com/fastfloat/fast_float/releases/download/v8.2.9/fast_float.h>
 
 ## Benchmarking
 

include/fast_float/ascii_number.h

@@ -600,7 +600,8 @@ parse_int_string(UC const *p, UC const *pend, T &value,
 
   UC const *const start_digits = p;
 
-  FASTFLOAT_IF_CONSTEXPR17((std::is_same<T, std::uint8_t>::value)) {
+  FASTFLOAT_IF_CONSTEXPR17(
+      (std::is_same<T, std::uint8_t>::value && sizeof(UC) == 1)) {
     if (base == 10) {
       const size_t len = (size_t)(pend - p);
       if (len == 0) {
@@ -692,7 +693,8 @@ parse_int_string(UC const *p, UC const *pend, T &value,
     }
   }
 
-  FASTFLOAT_IF_CONSTEXPR17((std::is_same<T, std::uint16_t>::value)) {
+  FASTFLOAT_IF_CONSTEXPR17(
+      (std::is_same<T, std::uint16_t>::value && sizeof(UC) == 1)) {
     if (base == 10) {
       const size_t len = size_t(pend - p);
       if (len == 0) {

include/fast_float/float_common.h

@@ -18,7 +18,7 @@
 
 #define FASTFLOAT_VERSION_MAJOR 8
 #define FASTFLOAT_VERSION_MINOR 2
-#define FASTFLOAT_VERSION_PATCH 8
+#define FASTFLOAT_VERSION_PATCH 9
 
 #define FASTFLOAT_STRINGIZE_IMPL(x) #x
 #define FASTFLOAT_STRINGIZE(x) FASTFLOAT_STRINGIZE_IMPL(x)

tests/fast_int.cpp

@@ -1295,6 +1295,59 @@ int main() {
       return EXIT_FAILURE;
     }
   }
+  // The uint8_t and uint16_t base-10 paths use a byte-oriented fast path. A
+  // wider code unit whose low byte is an ASCII digit (e.g. U+2131..U+2139) must
+  // not be mistaken for that digit. The generic path already rejects these for
+  // int; the fixed-width fast paths must agree. Lengths of 1..5 exercise both
+  // the uint8_t path and the 4-digit uint16_t SWAR path.
+  {
+    const std::u16string bad16[] = {
+        u"ℹ", u"ℱℲ", u"ℱℲℳ", u"ℱℲℳℴ", u"ℱℲℳℴℵ",
+    };
+    const std::u32string bad32[] = {
+        U"ℹ",
+        U"ℱℲℳ",
+        U"ℱℲℳℴ",
+        U"ℱℲℳℴℵ",
+    };
+    bool failed = false;
+    for (auto const &s : bad16) {
+      uint8_t r8 = 123;
+      auto a8 = fast_float::from_chars(s.data(), s.data() + s.size(), r8);
+      if (a8.ec == std::errc()) {
+        failed = true;
+        std::cerr << "Incorrectly parsed wide units as uint8_t " << unsigned(r8)
+                  << "." << std::endl;
+      }
+      uint16_t r16 = 123;
+      auto a16 = fast_float::from_chars(s.data(), s.data() + s.size(), r16);
+      if (a16.ec == std::errc()) {
+        failed = true;
+        std::cerr << "Incorrectly parsed wide units as uint16_t " << r16 << "."
+                  << std::endl;
+      }
+    }
+    for (auto const &s : bad32) {
+      uint8_t r8 = 123;
+      auto a8 = fast_float::from_chars(s.data(), s.data() + s.size(), r8);
+      if (a8.ec == std::errc()) {
+        failed = true;
+        std::cerr << "Incorrectly parsed wide units as uint8_t " << unsigned(r8)
+                  << "." << std::endl;
+      }
+      uint16_t r16 = 123;
+      auto a16 = fast_float::from_chars(s.data(), s.data() + s.size(), r16);
+      if (a16.ec == std::errc()) {
+        failed = true;
+        std::cerr << "Incorrectly parsed wide units as uint16_t " << r16 << "."
+                  << std::endl;
+      }
+    }
+
+    if (failed) {
+      return EXIT_FAILURE;
+    }
+  }
 
   return EXIT_SUCCESS;
 }