Merge #13671: Remove the boost/algorithm/string/case_conv.hpp dependency

b193d5a443bfd994936ad21b807b2bb37756ef2c Removes the Boost case_conv.hpp dependency. (251)
7a208d9fade56e2347891daff2f6b903923c9d50 Implements custom tolower and toupper functions. (251)
e2ba043b8d852d3f465bc293d6e494c8c6f75dfd Implements ParseNetwork unit test. (251)

Pull request description:

  This pull request removes the `boost/algorithm/string/case_conv.hpp` dependency from the project.

  `boost/algorithm/string/case_conv.hpp` is included for the `boost::to_lower` and `boost::to_upper` template functions.

  We can replace the calls to these functions with straightforward alternative implementations that use the C++ Standard Library, because the functions are called with `std::string` objects that use standard 7-bit ASCII characters as argument.

  The refactored implementation should work without the explicit `static_cast<unsigned char>` cast and `unsigned char` lambda return type. Both have been added defensively and to be explicit. Especially in case of the former, behaviour is undefined (potentially result in a crash) if the `std::toupper` argument is not an `unsigned char`.

  A potential alternative, maybe even preferred, implementation to address the `boost::to_lower` function call in `ParseNetwork(std::string)` could have been:

  ```c++
  if (net == "ipv4" || net == "IPv4") return NET_IPV4;
  if (net == "ipv6" || net == "IPv6") return NET_IPV6;
  ```
  This alternative implementation would however change the external behaviour of `ParseNetwork(std::string)`.

  This pull requests includes a unit test to validate the implementation of `ParseNetwork(std::string)`  prior and after the removal of the `case_conv.hpp` dependency.

  `boost/algorithm/string/case_conv.hpp` has been removed from the `EXPECTED_BOOST_INCLUDES` in `test/lint/lint-includes.sh` because it is no longer required.

Tree-SHA512: d803ae709f2368a3efb223097384a722436955bce0c44a1a5cffd0abb3164be0cce85ba0e9ebd9408166df3f1a95ea0c0d29e3a2534af2fae206c0419d67fde9

8 files changed, 116 insertions, 10 deletions

diff --git a/src/netbase.cpp b/src/netbase.cpp
index 9750173987..4b63757f3d 100644
--- a/src/netbase.cpp
+++ b/src/netbase.cpp
@@ -19,8 +19,6 @@
 #include <fcntl.h>
 #endif
 
-#include <boost/algorithm/string/case_conv.hpp> // for to_lower()
-
 #if !defined(MSG_NOSIGNAL)
 #define MSG_NOSIGNAL 0
 #endif
@@ -37,7 +35,7 @@ static const int SOCKS5_RECV_TIMEOUT = 20 * 1000;
 static std::atomic<bool> interruptSocks5Recv(false);
 
 enum Network ParseNetwork(std::string net) {
-    boost::to_lower(net);
+    Downcase(net);
     if (net == "ipv4") return NET_IPV4;
     if (net == "ipv6") return NET_IPV6;
     if (net == "onion") return NET_ONION;
diff --git a/src/rpc/server.cpp b/src/rpc/server.cpp
index e46bf2f765..7d7e83fea8 100644
--- a/src/rpc/server.cpp
+++ b/src/rpc/server.cpp
@@ -16,7 +16,6 @@
 
 #include <boost/bind.hpp>
 #include <boost/signals2/signal.hpp>
-#include <boost/algorithm/string/case_conv.hpp> // for to_upper()
 #include <boost/algorithm/string/classification.hpp>
 #include <boost/algorithm/string/split.hpp>
 
@@ -192,9 +191,7 @@ std::string CRPCTable::help(const std::string& strCommand, const JSONRPCRequest&
                     if (!category.empty())
                         strRet += "\n";
                     category = pcmd->category;
-                    std::string firstLetter = category.substr(0,1);
-                    boost::to_upper(firstLetter);
-                    strRet += "== " + firstLetter + category.substr(1) + " ==\n";
+                    strRet += "== " + Capitalize(category) + " ==\n";
                 }
             }
             strRet += strHelp + "\n";
diff --git a/src/test/netbase_tests.cpp b/src/test/netbase_tests.cpp
index 9c1af6bdf6..8072eb922d 100644
--- a/src/test/netbase_tests.cpp
+++ b/src/test/netbase_tests.cpp
@@ -302,4 +302,22 @@ BOOST_AUTO_TEST_CASE(netbase_getgroup)
     BOOST_CHECK(CreateInternal("baz.net").GetGroup() == internal_group);
 }
 
+BOOST_AUTO_TEST_CASE(netbase_parsenetwork)
+{
+    BOOST_CHECK_EQUAL(ParseNetwork("ipv4"), NET_IPV4);
+    BOOST_CHECK_EQUAL(ParseNetwork("ipv6"), NET_IPV6);
+    BOOST_CHECK_EQUAL(ParseNetwork("onion"), NET_ONION);
+    BOOST_CHECK_EQUAL(ParseNetwork("tor"), NET_ONION);
+
+    BOOST_CHECK_EQUAL(ParseNetwork("IPv4"), NET_IPV4);
+    BOOST_CHECK_EQUAL(ParseNetwork("IPv6"), NET_IPV6);
+    BOOST_CHECK_EQUAL(ParseNetwork("ONION"), NET_ONION);
+    BOOST_CHECK_EQUAL(ParseNetwork("TOR"), NET_ONION);
+
+    BOOST_CHECK_EQUAL(ParseNetwork(":)"), NET_UNROUTABLE);
+    BOOST_CHECK_EQUAL(ParseNetwork("tÖr"), NET_UNROUTABLE);
+    BOOST_CHECK_EQUAL(ParseNetwork("\xfe\xff"), NET_UNROUTABLE);
+    BOOST_CHECK_EQUAL(ParseNetwork(""), NET_UNROUTABLE);
+}
+
 BOOST_AUTO_TEST_SUITE_END()
diff --git a/src/test/util_tests.cpp b/src/test/util_tests.cpp
index b2960446eb..8e2f5abe66 100644
--- a/src/test/util_tests.cpp
+++ b/src/test/util_tests.cpp
@@ -1217,4 +1217,43 @@ BOOST_AUTO_TEST_CASE(test_DirIsWritable)
     fs::remove(tmpdirname);
 }
 
+BOOST_AUTO_TEST_CASE(test_ToLower)
+{
+    BOOST_CHECK_EQUAL(ToLower('@'), '@');
+    BOOST_CHECK_EQUAL(ToLower('A'), 'a');
+    BOOST_CHECK_EQUAL(ToLower('Z'), 'z');
+    BOOST_CHECK_EQUAL(ToLower('['), '[');
+    BOOST_CHECK_EQUAL(ToLower(0), 0);
+    BOOST_CHECK_EQUAL(ToLower(255), 255);
+
+    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_AUTO_TEST_CASE(test_ToUpper)
+{
+    BOOST_CHECK_EQUAL(ToUpper('`'), '`');
+    BOOST_CHECK_EQUAL(ToUpper('a'), 'A');
+    BOOST_CHECK_EQUAL(ToUpper('z'), 'Z');
+    BOOST_CHECK_EQUAL(ToUpper('{'), '{');
+    BOOST_CHECK_EQUAL(ToUpper(0), 0);
+    BOOST_CHECK_EQUAL(ToUpper(255), 255);
+}
+
+BOOST_AUTO_TEST_CASE(test_Capitalize)
+{
+    BOOST_CHECK_EQUAL(Capitalize(""), "");
+    BOOST_CHECK_EQUAL(Capitalize("bitcoin"), "Bitcoin");
+    BOOST_CHECK_EQUAL(Capitalize("\x00\xfe\xff"), "\x00\xfe\xff");
+}
+
 BOOST_AUTO_TEST_SUITE_END()
diff --git a/src/utilstrencodings.cpp b/src/utilstrencodings.cpp
index e66e2c238c..2326383586 100644
--- a/src/utilstrencodings.cpp
+++ b/src/utilstrencodings.cpp
@@ -7,6 +7,7 @@
 
 #include <tinyformat.h>
 
+#include <algorithm>
 #include <cstdlib>
 #include <cstring>
 #include <errno.h>
@@ -584,3 +585,15 @@ bool ParseHDKeypath(const std::string& keypath_str, std::vector<uint32_t>& keypa
     }
     return true;
 }
+
+void Downcase(std::string& str)
+{
+    std::transform(str.begin(), str.end(), str.begin(), [](unsigned char c){return ToLower(c);});
+}
+
+std::string Capitalize(std::string str)
+{
+    if (str.empty()) return str;
+    str[0] = ToUpper(str.front());
+    return str;
+}
diff --git a/src/utilstrencodings.h b/src/utilstrencodings.h
index 8f29f8f322..846a3917a2 100644
--- a/src/utilstrencodings.h
+++ b/src/utilstrencodings.h
@@ -186,4 +186,48 @@ bool ConvertBits(const O& outfn, I it, I end) {
 /** Parse an HD keypaths like "m/7/0'/2000". */
 bool ParseHDKeypath(const std::string& keypath_str, std::vector<uint32_t>& keypath);
 
+/**
+ * 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.
+ * @param[in] c     the character to convert to lowercase.
+ * @return          the lowercase equivalent of c; or the argument
+ *                  if no conversion is possible.
+ */
+constexpr unsigned char ToLower(unsigned char c)
+{
+    return (c >= 'A' && c <= 'Z' ? (c - 'A') + 'a' : c);
+}
+
+/**
+ * Converts the given string to its lowercase equivalent.
+ * 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.
+ */
+void Downcase(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.
+ * @param[in] c     the character to convert to uppercase.
+ * @return          the uppercase equivalent of c; or the argument
+ *                  if no conversion is possible.
+ */
+constexpr unsigned char ToUpper(unsigned char c)
+{
+    return (c >= 'a' && c <= 'z' ? (c - 'a') + 'A' : c);
+}
+
+/**
+ * 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.
+ * @param[in] str   the string to capitalize.
+ * @return          string with the first letter capitalized.
+ */
+std::string Capitalize(std::string str);
+
 #endif // BITCOIN_UTILSTRENCODINGS_H
diff --git a/test/lint/lint-includes.sh b/test/lint/lint-includes.sh
index a38ce6f4cf..346829e24b 100755
--- a/test/lint/lint-includes.sh
+++ b/test/lint/lint-includes.sh
@@ -47,7 +47,6 @@ fi
 
 EXPECTED_BOOST_INCLUDES=(
     boost/algorithm/string.hpp
-    boost/algorithm/string/case_conv.hpp
     boost/algorithm/string/classification.hpp
     boost/algorithm/string/replace.hpp
     boost/algorithm/string/split.hpp
diff --git a/test/lint/lint-locale-dependence.sh b/test/lint/lint-locale-dependence.sh
index cbe1143bd0..a5b97ca1e9 100755
--- a/test/lint/lint-locale-dependence.sh
+++ b/test/lint/lint-locale-dependence.sh
@@ -11,11 +11,9 @@ KNOWN_VIOLATIONS=(
     "src/dbwrapper.cpp:.*vsnprintf"
     "src/httprpc.cpp.*trim"
     "src/init.cpp:.*atoi"
-    "src/netbase.cpp.*to_lower"
     "src/qt/rpcconsole.cpp:.*atoi"
     "src/qt/rpcconsole.cpp:.*isdigit"
     "src/rest.cpp:.*strtol"
-    "src/rpc/server.cpp.*to_upper"
     "src/test/dbwrapper_tests.cpp:.*snprintf"
     "src/test/getarg_tests.cpp.*split"
     "src/torcontrol.cpp:.*atoi"