Merge #13883: utils: Convert Windows args to utf-8 string

380c843217 utils: Convert Windows args to utf-8 string (Chun Kuan Lee)

Pull request description:

  Create a new class `WinCmdLineArgs` when building for Windows. It converts all command line arguments to utf8 string.

Tree-SHA512: f098520fd123a8a452bc84a55dc8c0b88f0c475410efe57f2ccc393f86c396eed59ea1575ddc1b920323792e390fdb092061d80cdcd9b682f0ac79a22a22ff82

6 files changed, 62 insertions, 1 deletions

diff --git a/src/bitcoin-cli.cpp b/src/bitcoin-cli.cpp
index 09507fd249..f466505114 100644
--- a/src/bitcoin-cli.cpp
+++ b/src/bitcoin-cli.cpp
@@ -17,6 +17,7 @@
 
 #include <memory>
 #include <stdio.h>
+#include <tuple>
 
 #include <event2/buffer.h>
 #include <event2/keyvalq_struct.h>
@@ -511,6 +512,10 @@ static int CommandLineRPC(int argc, char *argv[])
 
 int main(int argc, char* argv[])
 {
+#ifdef WIN32
+    util::WinCmdLineArgs winArgs;
+    std::tie(argc, argv) = winArgs.get();
+#endif
     SetupEnvironment();
     if (!SetupNetworking()) {
         fprintf(stderr, "Error: Initializing networking failed\n");
diff --git a/src/bitcoind.cpp b/src/bitcoind.cpp
index bf04d95b50..18fcd9bc2a 100644
--- a/src/bitcoind.cpp
+++ b/src/bitcoind.cpp
@@ -185,6 +185,10 @@ static bool AppInit(int argc, char* argv[])
 
 int main(int argc, char* argv[])
 {
+#ifdef WIN32
+    util::WinCmdLineArgs winArgs;
+    std::tie(argc, argv) = winArgs.get();
+#endif
     SetupEnvironment();
 
     // Connect bitcoind signal handlers
diff --git a/src/qt/bitcoin.cpp b/src/qt/bitcoin.cpp
index 1e950e2686..61a9f390e9 100644
--- a/src/qt/bitcoin.cpp
+++ b/src/qt/bitcoin.cpp
@@ -552,6 +552,10 @@ static void SetupUIArgs()
 #ifndef BITCOIN_QT_TEST
 int main(int argc, char *argv[])
 {
+#ifdef WIN32
+    util::WinCmdLineArgs winArgs;
+    std::tie(argc, argv) = winArgs.get();
+#endif
     SetupEnvironment();
 
     std::unique_ptr<interfaces::Node> node = interfaces::MakeNode();
diff --git a/src/util.cpp b/src/util.cpp
index fa624aee90..1002302904 100644
--- a/src/util.cpp
+++ b/src/util.cpp
@@ -61,6 +61,7 @@
 #include <codecvt>
 
 #include <io.h> /* for _commit */
+#include <shellapi.h>
 #include <shlobj.h>
 #endif
 
@@ -1200,6 +1201,10 @@ void SetupEnvironment()
     } catch (const std::runtime_error&) {
         setenv("LC_ALL", "C", 1);
     }
+#elif defined(WIN32)
+    // Set the default input/output charset is utf-8
+    SetConsoleCP(CP_UTF8);
+    SetConsoleOutputCP(CP_UTF8);
 #endif
     // The path locale is lazy initialized and to avoid deinitialization errors
     // in multithreading environments, it is set explicitly by the main thread.
@@ -1265,3 +1270,30 @@ int ScheduleBatchPriority()
     return 1;
 #endif
 }
+
+namespace util {
+#ifdef WIN32
+WinCmdLineArgs::WinCmdLineArgs()
+{
+    wchar_t** wargv = CommandLineToArgvW(GetCommandLineW(), &argc);
+    std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>, wchar_t> utf8_cvt;
+    argv = new char*[argc];
+    args.resize(argc);
+    for (int i = 0; i < argc; i++) {
+        args[i] = utf8_cvt.to_bytes(wargv[i]);
+        argv[i] = &*args[i].begin();
+    }
+    LocalFree(wargv);
+}
+
+WinCmdLineArgs::~WinCmdLineArgs()
+{
+    delete[] argv;
+}
+
+std::pair<int, char**> WinCmdLineArgs::get()
+{
+    return std::make_pair(argc, argv);
+}
+#endif
+} // namespace util
diff --git a/src/util.h b/src/util.h
index f119385e48..fa6d2cd489 100644
--- a/src/util.h
+++ b/src/util.h
@@ -29,6 +29,7 @@
 #include <stdint.h>
 #include <string>
 #include <unordered_set>
+#include <utility>
 #include <vector>
 
 #include <boost/thread/condition_variable.hpp> // for boost::thread_interrupted
@@ -361,6 +362,21 @@ inline void insert(std::set<TsetT>& dst, const Tsrc& src) {
     dst.insert(src.begin(), src.end());
 }
 
+#ifdef WIN32
+class WinCmdLineArgs
+{
+public:
+    WinCmdLineArgs();
+    ~WinCmdLineArgs();
+    std::pair<int, char**> get();
+
+private:
+    int argc;
+    char** argv;
+    std::vector<std::string> args;
+};
+#endif
+
 } // namespace util
 
 #endif // BITCOIN_UTIL_H
diff --git a/test/functional/feature_uacomment.py b/test/functional/feature_uacomment.py
index 691a39b825..fb4ad21359 100755
--- a/test/functional/feature_uacomment.py
+++ b/test/functional/feature_uacomment.py
@@ -31,7 +31,7 @@ class UacommentTest(BitcoinTestFramework):
         self.nodes[0].assert_start_raises_init_error(["-uacomment=" + 'a' * 256], expected, match=ErrorMatch.FULL_REGEX)
 
         self.log.info("test -uacomment unsafe characters")
-        for unsafe_char in ['/', ':', '(', ')']:
+        for unsafe_char in ['/', ':', '(', ')', '₿', '🏃']:
             expected = "Error: User Agent comment \(" + re.escape(unsafe_char) + "\) contains unsafe characters."
             self.nodes[0].assert_start_raises_init_error(["-uacomment=" + unsafe_char], expected, match=ErrorMatch.FULL_REGEX)