Merge #16566: util: refactor upper/lowercase functions

0481fa2584 util: refactor upper/lowercase functions (Karl-Johan Alm)

Pull request description:

  This includes renaming Downcase() to ToLower() and make it return a string rather than modify referenced arg.
  Also adds ToUpper() string version.

  Additionally, it clarifies that the locale independency of the case functions is a *feature* and not a limitation. I interpreted it as the latter and rewrote code to be locale-aware before realizing this.

  This is done in preparation for #11413 and as a general refactor. I don't think the optimization that the pre-refactor state gave warrants the unwieldy usage.

ACKs for top commit:
  laanwj:
    ACK 0481fa2584 from me
  practicalswift:
    ACK 0481fa2584 -- diff looks correct
  l2a5b1:
    ACK 0481fa2 - Although, I think @luke-jr's [feedback](https://github.com/bitcoin/bitcoin/pull/16566#issuecomment-519580760) is spot on; `Downcase` is just an artifact of `ParseNetwork`, which has been happily downcasing `net` via a string argument for over 7 years; and I do recommend to add `ToLower` *when* somebody actually needs it in new work, there is no point in keeping a trivial utility function if it is not appreciated.
  promag:
    ACK 0481fa2584.

Tree-SHA512: 9b834ecc1b97db043e261bcbc59e42372e11e2fb9a6943688f18a835bf5c9205f68e4614f58e90ba260d1b8f0e060c6f67b390b62436c21b56891db23bc41628

src/netbase.cpp

@@ -37,8 +37,8 @@ bool fNameLookup = DEFAULT_NAME_LOOKUP;
 static const int SOCKS5_RECV_TIMEOUT = 20 * 1000;
 static std::atomic<bool> interruptSocks5Recv(false);
 
-enum Network ParseNetwork(std::string net) {
-    Downcase(net);
+enum Network ParseNetwork(const std::string& net_in) {
+    std::string net = ToLower(net_in);
     if (net == "ipv4") return NET_IPV4;
     if (net == "ipv6") return NET_IPV6;
     if (net == "onion") return NET_ONION;

src/netbase.h

@@ -37,7 +37,7 @@ public:
     bool randomize_credentials;
 };
 
-enum Network ParseNetwork(std::string net);
+enum Network ParseNetwork(const std::string& net);
 std::string GetNetworkName(enum Network net);
 bool SetProxy(enum Network net, const proxyType &addrProxy);
 bool GetProxy(enum Network net, proxyType &proxyInfoOut);

src/test/util_tests.cpp

@@ -1532,17 +1532,9 @@ BOOST_AUTO_TEST_CASE(test_ToLower)
     BOOST_CHECK_EQUAL(ToLower(0), 0);
     BOOST_CHECK_EQUAL(ToLower('\xff'), '\xff');
 
-    std::string testVector;
-    Downcase(testVector);
-    BOOST_CHECK_EQUAL(testVector, "");
-
-    testVector = "#HODL";
-    Downcase(testVector);
-    BOOST_CHECK_EQUAL(testVector, "#hodl");
-
-    testVector = "\x00\xfe\xff";
-    Downcase(testVector);
-    BOOST_CHECK_EQUAL(testVector, "\x00\xfe\xff");
+    BOOST_CHECK_EQUAL(ToLower(""), "");
+    BOOST_CHECK_EQUAL(ToLower("#HODL"), "#hodl");
+    BOOST_CHECK_EQUAL(ToLower("\x00\xfe\xff"), "\x00\xfe\xff");
 }
 
 BOOST_AUTO_TEST_CASE(test_ToUpper)
@@ -1553,6 +1545,10 @@ BOOST_AUTO_TEST_CASE(test_ToUpper)
     BOOST_CHECK_EQUAL(ToUpper('{'), '{');
     BOOST_CHECK_EQUAL(ToUpper(0), 0);
     BOOST_CHECK_EQUAL(ToUpper('\xff'), '\xff');
+
+    BOOST_CHECK_EQUAL(ToUpper(""), "");
+    BOOST_CHECK_EQUAL(ToUpper("#hodl"), "#HODL");
+    BOOST_CHECK_EQUAL(ToUpper("\x00\xfe\xff"), "\x00\xfe\xff");
 }
 
 BOOST_AUTO_TEST_CASE(test_Capitalize)

src/util/strencodings.cpp

@@ -546,9 +546,18 @@ bool ParseFixedPoint(const std::string &val, int decimals, int64_t *amount_out)
     return true;
 }
 
-void Downcase(std::string& str)
+std::string ToLower(const std::string& str)
 {
-    std::transform(str.begin(), str.end(), str.begin(), [](char c){return ToLower(c);});
+    std::string r;
+    for (auto ch : str) r += ToLower((unsigned char)ch);
+    return r;
+}
+
+std::string ToUpper(const std::string& str)
+{
+    std::string r;
+    for (auto ch : str) r += ToUpper((unsigned char)ch);
+    return r;
 }
 
 std::string Capitalize(std::string str)

src/util/strencodings.h

@@ -199,6 +199,8 @@ bool ConvertBits(const O& outfn, I it, I end) {
  * Converts the given character to its lowercase equivalent.
  * This function is locale independent. It only converts uppercase
  * characters in the standard 7-bit ASCII range.
+ * This is a feature, not a limitation.
+ *
  * @param[in] c     the character to convert to lowercase.
  * @return          the lowercase equivalent of c; or the argument
  *                  if no conversion is possible.
@@ -209,17 +211,22 @@ constexpr char ToLower(char c)
 }
 
 /**
- * Converts the given string to its lowercase equivalent.
+ * Returns the lowercase equivalent of the given string.
  * This function is locale independent. It only converts uppercase
  * characters in the standard 7-bit ASCII range.
- * @param[in,out] str   the string to convert to lowercase.
+ * This is a feature, not a limitation.
+ *
+ * @param[in] str   the string to convert to lowercase.
+ * @returns         lowercased equivalent of str
  */
-void Downcase(std::string& str);
+std::string ToLower(const std::string& str);
 
 /**
  * Converts the given character to its uppercase equivalent.
  * This function is locale independent. It only converts lowercase
  * characters in the standard 7-bit ASCII range.
+ * This is a feature, not a limitation.
+ *
  * @param[in] c     the character to convert to uppercase.
  * @return          the uppercase equivalent of c; or the argument
  *                  if no conversion is possible.
@@ -229,13 +236,25 @@ constexpr char ToUpper(char c)
     return (c >= 'a' && c <= 'z' ? (c - 'a') + 'A' : c);
 }
 
+/**
+ * Returns the uppercase equivalent of the given string.
+ * This function is locale independent. It only converts lowercase
+ * characters in the standard 7-bit ASCII range.
+ * This is a feature, not a limitation.
+ *
+ * @param[in] str   the string to convert to uppercase.
+ * @returns         UPPERCASED EQUIVALENT OF str
+ */
+std::string ToUpper(const std::string& str);
+
 /**
  * Capitalizes the first character of the given string.
- * This function is locale independent. It only capitalizes the
- * first character of the argument if it has an uppercase equivalent
- * in the standard 7-bit ASCII range.
+ * This function is locale independent. It only converts lowercase
+ * characters in the standard 7-bit ASCII range.
+ * This is a feature, not a limitation.
+ *
  * @param[in] str   the string to capitalize.
- * @return          string with the first letter capitalized.
+ * @returns         string with the first letter capitalized.
  */
 std::string Capitalize(std::string str);
 

src/util/system.cpp

@@ -388,7 +388,7 @@ bool ArgsManager::ParseParameters(int argc, const char* const argv[], std::strin
             key.erase(is_index);
         }
 #ifdef WIN32
-        std::transform(key.begin(), key.end(), key.begin(), ToLower);
+        key = ToLower(key);
         if (key[0] == '/')
             key[0] = '-';
 #endif