1 files changed, 51 insertions, 0 deletions
diff --git a/libkonq/konq_string_compare.h b/libkonq/konq_string_compare.h
new file mode 100644
index 000000000..8d17d409f
--- /dev/null
+++ b/libkonq/konq_string_compare.h
@@ -0,0 +1,51 @@
+#ifndef KONQ_STRING_COMPARE_H
+#define KONQ_STRING_COMPARE_H
+
+#include "konq_sort_constants.h"
+
+static inline int stringCompare(
+  const TextSortOrder sortorder,
+  const TQString& a,
+  const TQString& b
+)
+{
+  // Our caller probably determined sortorder from KonqPropsView::getSortOrder()
+  // but we have a reasonable fallback position for bogus values.
+
+  switch(sortorder) {
+
+    case UNICODE_UNMODIFIED:
+      /*
+       * Strictly character code(point) numeric comparison as defined
+       * by the Unicode Standard that is backward compatible with the
+       * the ASCII standard.
+   . */
+      return a.compare( b );
+      break;
+
+    case UNICODE_CASEINSENSITIVE:
+      /*
+       * This is the traditional "case-insensitive" variation on character
+       * code order that ensures that ASCII lowercase and uppercase alphabetic
+       * characters are are grouped together instead of being separated by
+       * non-alphabetic ASCII characters [ \ ] ^ _ `
+      */
+      return a.lower().compare( b.lower() );
+      break;
+
+    case LOCALE_UNMODIFIED:
+      /*
+       * This is the pure locale-aware comparison as defined by ICU.
+       * Note: if LC_COLLATE == 'C' or 'Posix', this will produce same
+       * result as UNICODE_UNMODIFIED.
+      */
+      return a.localeAwareCompare( b );
+      break;
+
+    default: // Treat as UNICODE_UNMODIFIED
+      return a.compare( b );
+      break;
+  }
+}
+
+#endif // KONQ_STRING_COMPARE_H