[libc] Change ctype to be encoding independent (#110574)

The previous implementation of the ctype functions assumed ASCII.
This patch changes to a switch/case implementation that looks odd, but
actually is easier for the compiler to understand and optimize.

Author: michaelrj-google

libc/src/__support/ctype_utils.h

@@ -178,9 +178,11 @@ class HighPrecisionDecimal {
       if (digit_index >= this->num_digits) {
         return new_digits - 1;
       }
-      if (this->digits[digit_index] != power_of_five[digit_index] - '0') {
+      if (this->digits[digit_index] !=
+          internal::b36_char_to_int(power_of_five[digit_index])) {
         return new_digits -
-               ((this->digits[digit_index] < power_of_five[digit_index] - '0')
+               ((this->digits[digit_index] <
+                 internal::b36_char_to_int(power_of_five[digit_index]))
                     ? 1
                     : 0);
       }
@@ -337,8 +339,8 @@ class HighPrecisionDecimal {
         }
         ++total_digits;
         if (this->num_digits < MAX_NUM_DIGITS) {
-          this->digits[this->num_digits] =
-              static_cast<uint8_t>(num_string[num_cur] - '0');
+          this->digits[this->num_digits] = static_cast<uint8_t>(
+              internal::b36_char_to_int(num_string[num_cur]));
           ++this->num_digits;
         } else if (num_string[num_cur] != '0') {
           this->truncated = true;

libc/src/__support/high_precision_decimal.h

@@ -13,12 +13,13 @@
 #ifndef LLVM_LIBC_SRC___SUPPORT_INTEGER_LITERALS_H
 #define LLVM_LIBC_SRC___SUPPORT_INTEGER_LITERALS_H
 
-#include "src/__support/CPP/limits.h"        // CHAR_BIT
+#include "src/__support/CPP/limits.h" // CHAR_BIT
+#include "src/__support/ctype_utils.h"
 #include "src/__support/macros/attributes.h" // LIBC_INLINE
 #include "src/__support/macros/config.h"
-#include "src/__support/uint128.h"           // UInt128
-#include <stddef.h>                          // size_t
-#include <stdint.h>                          // uintxx_t
+#include "src/__support/uint128.h" // UInt128
+#include <stddef.h>                // size_t
+#include <stdint.h>                // uintxx_t
 
 namespace LIBC_NAMESPACE_DECL {
 
@@ -75,26 +76,13 @@ template <typename T, int base> struct DigitBuffer {
       push(*str);
   }
 
-  // Returns the digit for a particular character.
-  // Returns INVALID_DIGIT if the character is invalid.
-  LIBC_INLINE static constexpr uint8_t get_digit_value(const char c) {
-    const auto to_lower = [](char c) { return c | 32; };
-    const auto is_digit = [](char c) { return c >= '0' && c <= '9'; };
-    const auto is_alpha = [](char c) {
-      return ('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z');
-    };
-    if (is_digit(c))
-      return static_cast<uint8_t>(c - '0');
-    if (base > 10 && is_alpha(c))
-      return static_cast<uint8_t>(to_lower(c) - 'a' + 10);
-    return INVALID_DIGIT;
-  }
-
   // Adds a single character to this buffer.
   LIBC_INLINE constexpr void push(char c) {
     if (c == '\'')
       return; // ' is valid but not taken into account.
-    const uint8_t value = get_digit_value(c);
+    const int b36_val = internal::b36_char_to_int(c);
+    const uint8_t value = static_cast<uint8_t>(
+        b36_val < base && (b36_val != 0 || c == '0') ? b36_val : INVALID_DIGIT);
     if (value == INVALID_DIGIT || size >= MAX_DIGITS) {
       // During constant evaluation `__builtin_unreachable` will halt the
       // compiler as it is not executable. This is preferable over `assert` that

libc/src/__support/integer_literals.h

@@ -69,6 +69,7 @@
 #include "src/__support/CPP/type_traits.h"
 #include "src/__support/big_int.h" // make_integral_or_big_int_unsigned_t
 #include "src/__support/common.h"
+#include "src/__support/ctype_utils.h"
 #include "src/__support/macros/config.h"
 
 namespace LIBC_NAMESPACE_DECL {
@@ -214,9 +215,9 @@ template <typename T, typename Fmt = radix::Dec> class IntegerToString {
     using UNSIGNED_T = make_integral_or_big_int_unsigned_t<T>;
 
     LIBC_INLINE static char digit_char(uint8_t digit) {
-      if (digit < 10)
-        return '0' + static_cast<char>(digit);
-      return (Fmt::IS_UPPERCASE ? 'A' : 'a') + static_cast<char>(digit - 10);
+      const int result = internal::int_to_b36_char(digit);
+      return static_cast<char>(Fmt::IS_UPPERCASE ? internal::toupper(result)
+                                                 : result);
     }
 
     LIBC_INLINE static void

libc/src/__support/integer_to_string.h

@@ -909,7 +909,7 @@ decimal_string_to_float(const char *__restrict src, const char DECIMAL_POINT,
       cpp::numeric_limits<StorageType>::max() / BASE;
   while (true) {
     if (isdigit(src[index])) {
-      uint32_t digit = src[index] - '0';
+      uint32_t digit = b36_char_to_int(src[index]);
       seen_digit = true;
 
       if (mantissa < bitstype_max_div_by_base) {

libc/src/__support/str_to_float.h

@@ -42,22 +42,14 @@ first_non_whitespace(const char *__restrict src,
   return src + src_cur;
 }
 
-LIBC_INLINE int b36_char_to_int(char input) {
-  if (isdigit(input))
-    return input - '0';
-  if (isalpha(input))
-    return (input | 32) + 10 - 'a';
-  return 0;
-}
-
 // checks if the next 3 characters of the string pointer are the start of a
 // hexadecimal number. Does not advance the string pointer.
 LIBC_INLINE bool
 is_hex_start(const char *__restrict src,
              size_t src_len = cpp::numeric_limits<size_t>::max()) {
   if (src_len < 3)
     return false;
-  return *src == '0' && (*(src + 1) | 32) == 'x' && isalnum(*(src + 2)) &&
+  return *src == '0' && tolower(*(src + 1)) == 'x' && isalnum(*(src + 2)) &&
          b36_char_to_int(*(src + 2)) < 16;
 }
 

libc/src/__support/str_to_integer.h

@@ -16,7 +16,8 @@ namespace LIBC_NAMESPACE_DECL {
 
 LLVM_LIBC_FUNCTION(int, isxdigit, (int c)) {
   const unsigned ch = static_cast<unsigned>(c);
-  return static_cast<int>(internal::isdigit(ch) || (ch | 32) - 'a' < 6);
+  return static_cast<int>(internal::isalnum(ch) &&
+                          internal::b36_char_to_int(ch) < 16);
 }
 
 } // namespace LIBC_NAMESPACE_DECL

libc/src/ctype/isxdigit.cpp

@@ -16,7 +16,8 @@ namespace LIBC_NAMESPACE_DECL {
 
 LLVM_LIBC_FUNCTION(int, isxdigit_l, (int c, locale_t)) {
   const unsigned ch = static_cast<unsigned>(c);
-  return static_cast<int>(internal::isdigit(ch) || (ch | 32) - 'a' < 6);
+  return static_cast<int>(internal::isalnum(ch) &&
+                          internal::b36_char_to_int(ch) < 16);
 }
 
 } // namespace LIBC_NAMESPACE_DECL

libc/src/ctype/isxdigit_l.cpp

@@ -14,10 +14,6 @@
 
 namespace LIBC_NAMESPACE_DECL {
 
-LLVM_LIBC_FUNCTION(int, toupper, (int c)) {
-  if (internal::islower(c))
-    return c - ('a' - 'A');
-  return c;
-}
+LLVM_LIBC_FUNCTION(int, toupper, (int c)) { return internal::toupper(c); }
 
 } // namespace LIBC_NAMESPACE_DECL

libc/src/ctype/toupper.cpp

@@ -15,9 +15,7 @@
 namespace LIBC_NAMESPACE_DECL {
 
 LLVM_LIBC_FUNCTION(int, toupper_l, (int c, locale_t)) {
-  if (internal::islower(c))
-    return c - ('a' - 'A');
-  return c;
+  return internal::toupper(c);
 }
 
 } // namespace LIBC_NAMESPACE_DECL