add -input-codec option for qm reader

basing the input codec on the locale was broken by design and didn't
really work anyway. so adding this option really is a bugfix, kind of.

2 files changed, 10 insertions, 3 deletions

diff --git a/tools/linguist/lconvert/main.cpp b/tools/linguist/lconvert/main.cpp
index 336bf31..ddde578 100644
--- a/tools/linguist/lconvert/main.cpp
+++ b/tools/linguist/lconvert/main.cpp
@@ -85,6 +85,9 @@ static int usage(const QStringList &args)
         "    -of <outformat>\n"
         "    --output-format <outformat>\n"
         "           Specify output format. See -if.\n\n"
+        "    --input-codec <codec>\n"
+        "           Specify encoding for .qm input files. Default is 'Latin1'.\n"
+        "           UTF-8 is always tried as well, corresponding to the trUtf8() function.\n\n"
         "    --drop-tags <regexp>\n"
         "           Drop named extra tags when writing 'ts' or 'xlf' files.\n"
         "           May be specified repeatedly.\n\n"
@@ -136,6 +139,7 @@ int main(int argc, char *argv[])
     bool verbose = false;
 
     ConversionData cd;
+    cd.m_codecForSource = "Latin1";
     Translator tr;
 
     for (int i = 1; i < args.size(); ++i) {
@@ -164,6 +168,10 @@ int main(int argc, char *argv[])
             if (++i >= args.size())
                 return usage(args);
             inFormat = args[i];
+        } else if (args[i] == QLatin1String("-input-codec")) {
+            if (++i >= args.size())
+                return usage(args);
+            cd.m_codecForSource = args[i].toLatin1();
         } else if (args[i] == QLatin1String("-drop-tag")) {
             if (++i >= args.size())
                 return usage(args);
diff --git a/tools/linguist/shared/qm.cpp b/tools/linguist/shared/qm.cpp
index 14f4c2c..c6f3602 100644
--- a/tools/linguist/shared/qm.cpp
+++ b/tools/linguist/shared/qm.cpp
@@ -545,8 +545,7 @@ bool loadQM(Translator &translator, QIODevice &dev, ConversionData &cd)
     size_t numItems = offsetLength / (2 * sizeof(quint32));
     //qDebug() << "NUMITEMS: " << numItems;
 
-    // FIXME: that's just a guess, the original locale data is lost...
-    QTextCodec *codec = QTextCodec::codecForLocale();
+    QTextCodec *codec = QTextCodec::codecForName(cd.m_codecForSource);
     QTextCodec *utf8Codec = 0;
     if (codec->name() != "UTF-8")
         utf8Codec = QTextCodec::codecForName("UTF-8");
@@ -649,7 +648,7 @@ bool loadQM(Translator &translator, QIODevice &dev, ConversionData &cd)
             }
             if (!(contextIsSystem && sourcetextIsSystem && commentIsSystem)) {
                 cd.appendError(QLatin1String(
-                        "Cannot read file with current system character codec"));
+                        "Cannot read file with specified input codec"));
                 return false;
             }
             // The message is 8-bit in the file's encoding (utf-8 or not).