Deduplicate multiple isHexDigit impls and optimize

src/workerd/api/util.c++

@@ -7,6 +7,7 @@
 #include "simdutf.h"
 
 #include <workerd/util/mimetype.h>
+#include <workerd/util/strings.h>
 
 #include <kj/encoding.h>
 
@@ -216,7 +217,6 @@ kj::String redactUrl(kj::StringPtr url) {
     bool isUpper = ('A' <= c && c <= 'Z');
     bool isLower = ('a' <= c && c <= 'z');
     bool isDigit = ('0' <= c && c <= '9');
-    bool isHexDigit = (isDigit || ('A' <= c && c <= 'F') || ('a' <= c && c <= 'f'));
     bool isSep = (c == '+' || c == '-' || c == '_');
     // These extra characters are used in the regular and url-safe versions of
     // base64, but might also be used for GUID-style separators in hex ids.
@@ -225,10 +225,10 @@ kj::String redactUrl(kj::StringPtr url) {
     // character.
 
     if (isUpper || isLower || isDigit || isSep) {
-      if (isHexDigit) {
+      if (isHexDigit(c)) {
         hexDigitCount++;
       }
-      if (!isHexDigit && !isSep) {
+      if (!isHexDigit(c) && !isSep) {
         sawNonHexChar = true;
       }
       if (isUpper) {

src/workerd/api/util.h

@@ -35,11 +35,6 @@ kj::String toLower(kj::String&& str);
 // Mutate `str` with all alphabetic ASCII characters uppercased. Returns `str`.
 kj::String toUpper(kj::String&& str);
 
-inline bool isHexDigit(uint32_t c) {
-  // Check if `c` is the ASCII code of a hexadecimal digit.
-  return ('0' <= c && c <= '9') || ('a' <= c && c <= 'f') || ('A' <= c && c <= 'F');
-}
-
 // Parse `rawText` as application/x-www-form-urlencoded name/value pairs and store in `query`. If
 // `skipLeadingQuestionMark` is true, any initial '?' will be ignored. Otherwise, it will be
 // interpreted as part of the first URL-encoded field.

src/workerd/jsg/BUILD.bazel

@@ -92,6 +92,7 @@ wd_cc_library(
     visibility = ["//visibility:public"],
     deps = [
         ":memory-tracker",
+        "//src/workerd/util:strings",
         "@capnp-cpp//src/kj",
     ],
 )

src/workerd/jsg/url.c++

@@ -1,5 +1,7 @@
 #include "url.h"
 
+#include <workerd/util/strings.h>
+
 #include <kj/hash.h>
 
 extern "C" {
@@ -488,11 +490,6 @@ inline bool isAsciiDigit(char c) {
   return c >= '0' && c <= '9';
 };
 
-inline bool isHexDigit(char c) {
-  // Check if `c` is the ASCII code of a hexadecimal digit.
-  return isAsciiDigit(c) || ('a' <= c && c <= 'f') || ('A' <= c && c <= 'F');
-}
-
 inline bool isAscii(char codepoint) {
   return codepoint >= 0x00 && codepoint <= 0x7f;
 };

src/workerd/util/strings.h

@@ -23,4 +23,21 @@ inline kj::String toLowerCopy(kj::ArrayPtr<const char> ptr) {
   return kj::mv(str);
 }
 
+constexpr kj::FixedArray<uint8_t, 256> kHexDigitTable = []() consteval {
+  kj::FixedArray<uint8_t, 256> result{};
+  for (uint8_t c: {'1', '2', '3', '4', '5', '6', '7', '8', '9', '0'}) {
+    result[c] = true;
+  }
+  for (uint8_t c: {'A', 'B', 'C', 'D', 'E', 'F'}) {
+    result[c] = true;         // Uppercase
+    result[c + 0x20] = true;  // Lowercase
+  }
+  return result;
+}();
+
+// Check if `c` is the ASCII code of a hexadecimal digit.
+constexpr bool isHexDigit(char c) {
+  return kHexDigitTable[static_cast<int>(c)] == 1;
+}
+
 }  // namespace workerd