Make careful_string_filter a little better and use it for KDM filenames.

[dcpomatic.git] / src / lib / util.cc

diff --git a/src/lib/util.cc b/src/lib/util.cc
index b5265519413976328254749d63cb5236caa200cd..c2089146431614bda8957b4b26e74434653f6414 100644 (file)
--- a/src/lib/util.cc
+++ b/src/lib/util.cc
@@ -62,6 +62,7 @@ extern "C" {
 #include <boost/range/algorithm/replace_if.hpp>
 #include <boost/thread.hpp>
 #include <boost/filesystem.hpp>
+#include <boost/locale.hpp>
 #ifdef DCPOMATIC_WINDOWS
 #include <boost/locale.hpp>
 #include <dbghelp.h>
@@ -80,6 +81,7 @@ extern "C" {
 #include "i18n.h"
 
 using std::string;
+using std::wstring;
 using std::setfill;
 using std::ostream;
 using std::endl;
@@ -743,15 +745,34 @@ careful_string_filter (string s)
           Safety first and all that.
        */
 
+       wstring ws = boost::locale::conv::utf_to_utf<wchar_t>(s);
+
        string out;
        string const allowed = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz-_%.+";
-       for (size_t i = 0; i < s.size(); ++i) {
-               if (allowed.find (s[i]) != string::npos) {
-                       out += s[i];
+       for (size_t i = 0; i < ws.size(); ++i) {
+
+               wchar_t c = ws[i];
+
+               /* Remove some accents */
+               if (wstring(L"áàâ").find(c) != string::npos) {
+                       c = 'a';
+               }
+               if (wstring(L"éèêë").find(c) != string::npos) {
+                       c = 'e';
+               }
+               if (wstring(L"ö").find(c) != string::npos) {
+                       c = 'o';
+               }
+               if (wstring(L"ü").find(c) != string::npos) {
+                       c = 'u';
+               }
+
+               if (allowed.find(c) != string::npos) {
+                       out += c;
                }
        }
 
-       return out;
+       return boost::locale::conv::utf_to_utf<char>(out);
 }
 
 /** @param mapped List of mapped audio channels from a Film.