Bring back arg_list_utf_16to8(), but have it just do UTF-16-to-UTF-8 mapping.

Call it from wmain() in the command-line tools, passing it the input
argument count and vector, and call it from main() in Wireshark, after
getting a UTF-16 argument vector from passing the result of
GetCommandLineW() to CommandLineToArgvW().

Change-Id: I0e51703c0a6c92f7892d196e700ab437bd702514
Reviewed-on: https://code.wireshark.org/review/30063
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Tested-by: Petri Dish Buildbot
Reviewed-by: Guy Harris <guy@alum.mit.edu>

capinfos.c

@@ -1726,15 +1726,12 @@ exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-  char **argv_utf8;
+  char **argv;
 
-  /* Convert our arg list from UTF-16LE to UTF-8. */
-  argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-  for (int i = 0; i < argc; i++)
-    argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-  return real_main(argc, argv_utf8);
+  argv = arg_list_utf_16to8(argc, wc_argv);
+  return real_main(argc, argv);
 }
 #else
 int

captype.c

@@ -207,15 +207,12 @@ real_main(int argc, char *argv[])
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-  char **argv_utf8;
+  char **argv;
 
-  /* Convert our arg list from UTF-16LE to UTF-8. */
-  argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-  for (int i = 0; i < argc; i++)
-    argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-  return real_main(argc, argv_utf8);
+  argv = arg_list_utf_16to8(argc, wc_argv);
+  return real_main(argc, argv);
 }
 #else
 int

dumpcap.c

@@ -5273,15 +5273,12 @@ real_main(int argc, char *argv[])
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-  char **argv_utf8;
+  char **argv;
 
-  /* Convert our arg list from UTF-16LE to UTF-8. */
-  argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-  for (int i = 0; i < argc; i++)
-    argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-  return real_main(argc, argv_utf8);
+  argv = arg_list_utf_16to8(argc, wc_argv);
+  return real_main(argc, argv);
 }
 #else
 int

editcap.c

@@ -1954,15 +1954,12 @@ clean_exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-    char **argv_utf8;
+    char **argv;
 
-    /* Convert our arg list from UTF-16LE to UTF-8. */
-    argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-    for (int i = 0; i < argc; i++)
-        argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-    return real_main(argc, argv_utf8);
+    argv = arg_list_utf_16to8(argc, wc_argv);
+    return real_main(argc, argv);
 }
 #else
 int

mergecap.c

@@ -473,15 +473,12 @@ clean_exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-  char **argv_utf8;
+  char **argv;
 
-  /* Convert our arg list from UTF-16LE to UTF-8. */
-  argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-  for (int i = 0; i < argc; i++)
-    argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-  return real_main(argc, argv_utf8);
+  argv = arg_list_utf_16to8(argc, wc_argv);
+  return real_main(argc, argv);
 }
 #else
 int

randpkt.c

@@ -248,15 +248,12 @@ clean_exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t **argv)
+wmain(int argc, wchar_t **wc_argv)
 {
-	char **argv_utf8;
+	char **argv;
 
-	/* Convert our arg list from UTF-16LE to UTF-8. */
-	argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-	for (int i = 0; i < argc; i++)
-		argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-	return real_main(argc, argv_utf8);
+	argv = arg_list_utf_16to8(argc, wc_argv);
+	return real_main(argc, argv);
 }
 #else
 int

rawshark.c

@@ -832,15 +832,12 @@ clean_exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-    char **argv_utf8;
+    char **argv;
 
-    /* Convert our arg list from UTF-16LE to UTF-8. */
-    argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-    for (int i = 0; i < argc; i++)
-        argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-    return real_main(argc, argv_utf8);
+    argv = arg_list_utf_16to8(argc, wc_argv);
+    return real_main(argc, argv);
 }
 #else
 int

text2pcap.c

@@ -1933,15 +1933,12 @@ clean_exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-    char **argv_utf8;
+    char **argv;
 
-    /* Convert our arg list from UTF-16LE to UTF-8. */
-    argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-    for (int i = 0; i < argc; i++)
-        argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-    return real_main(argc, argv_utf8);
+    argv = arg_list_utf_16to8(argc, wc_argv);
+    return real_main(argc, argv);
 }
 #else
 int

tfshark.c

@@ -1007,15 +1007,12 @@ clean_exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-  char **argv_utf8;
+  char **argv;
 
-  /* Convert our arg list from UTF-16LE to UTF-8. */
-  argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-  for (int i = 0; i < argc; i++)
-    argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-  return real_main(argc, argv_utf8);
+  argv = arg_list_utf_16to8(argc, wc_argv);
+  return real_main(argc, argv);
 }
 #else
 int

tshark.c

@@ -2260,15 +2260,12 @@ clean_exit:
 
 #ifdef _WIN32
 int
-wmain(int argc, wchar_t *argv[])
+wmain(int argc, wchar_t *wc_argv[])
 {
-  char **argv_utf8;
+  char **argv;
 
-  /* Convert our arg list from UTF-16LE to UTF-8. */
-  argv_utf8 = g_malloc(argc * sizeof *argv_utf8);
-  for (int i = 0; i < argc; i++)
-    argv_utf8[i] = g_utf16_to_utf8(argv[i], -1, NULL, NULL, NULL);
-  return real_main(argc, argv_utf8);
+  argv = arg_list_utf_16to8(argc, wc_argv);
+  return real_main(argc, argv);
 }
 #else
 int

ui/qt/main.cpp

@@ -357,7 +357,6 @@ int main(int argc, char *qt_argv[])
     MainWindow *main_w;
 
 #ifdef _WIN32
-    int                  opt;
     LPWSTR              *wc_argv;
     int                  wc_argc;
 #endif
@@ -420,13 +419,11 @@ int main(int argc, char *qt_argv[])
     // strings into UTF-8.
     //
     wc_argv = CommandLineToArgvW(GetCommandLineW(), &wc_argc);
-    if (wc_argv && wc_argc == argc) {
-        argv = (char **) g_malloc(sizeof(char *) * argc);
-        for (opt = 0; opt < argc; opt++) {
-            argv[opt] = g_utf16_to_utf8((const gunichar2 *)wc_argv[opt], -1, NULL, NULL, NULL);
-        }
+    if (wc_argv) {
+        argc = wc_argc;
+        argv = arg_list_utf_16to8(wc_argc, wc_argv);
+        LocalFree(wc_argv);
     } /* XXX else bail because something is horribly, horribly wrong? */
-    LocalFree(wc_argv);
 
     create_app_running_mutex();
 #endif /* _WIN32 */

wsutil/unicode-utils.c

@@ -141,6 +141,20 @@ utf_16to8(const wchar_t *utf16str)
 
   return utf8buf[idx];
 }
+
+/* Convert our argument list from UTF-16 to UTF-8. */
+char **
+arg_list_utf_16to8(int argc, wchar_t *wc_argv[]) {
+  char               **argv;
+  int                  i;
+
+  argv = (char **) g_malloc(sizeof(char *) * argc);
+  for (i = 0; i < argc; i++) {
+    argv[i] = g_utf16_to_utf8(wc_argv[i], -1, NULL, NULL, NULL);
+  }
+  return argv;
+}
+
 #endif
 
 /*

wsutil/unicode-utils.h

@@ -62,6 +62,17 @@ void utf_8to16_snprintf(TCHAR *utf16buf, gint utf16buf_len, const gchar* fmt,
  */
 WS_DLL_PUBLIC
 gchar * utf_16to8(const wchar_t *utf16str);
+
+/** Convert the supplied program argument list from UTF-16 to UTF-8
+ * return a pointer to the array of UTF-8 arguments. This is intended
+ * to be used to normalize command line arguments at program startup.
+ *
+ * @param argc The number of arguments.
+ * @param argv The argument values (vector).
+ */
+WS_DLL_PUBLIC
+char ** arg_list_utf_16to8(int argc, wchar_t *wc_argv[]);
+
 #endif /* _WIN32 */
 
 /*