Resolves: fdo#82904 non-Japanese ww95 documents claiming ms932 encoding

Change-Id: I62f8d5c3cac71f83f5cdde114f66e8554a780538

1 files changed, 44 insertions, 1 deletions

diff --git a/sw/source/filter/ww8/ww8par.cxx b/sw/source/filter/ww8/ww8par.cxx
index 4fc41eb40f4b..54cfbcc3a7fd 100644
--- a/sw/source/filter/ww8/ww8par.cxx
+++ b/sw/source/filter/ww8/ww8par.cxx
@@ -3037,8 +3037,51 @@ bool SwWW8ImplReader::ReadPlainChars(WW8_CP& rPos, sal_Int32 nEnd, sal_Int32 nCp
     // the correct FilePos has already been reached.
     const sal_Int32 nStrLen = std::min(nValidStrLen, SAL_MAX_INT32-1);
 
-    const rtl_TextEncoding eSrcCharSet = bVer67 ? GetCurrentCharSet() :
+    rtl_TextEncoding eSrcCharSet = bVer67 ? GetCurrentCharSet() :
         RTL_TEXTENCODING_MS_1252;
+    if (bVer67 && eSrcCharSet == RTL_TEXTENCODING_MS_932)
+    {
+        /*
+         fdo#82904
+
+         Older documents exported as word 95 that use unicode aware fonts will
+         have the charset of those fonts set to RTL_TEXTENCODING_MS_932 on
+         export as the conversion from RTL_TEXTENCODING_UNICODE. This is a serious
+         pain.
+
+         We will try and use a fallback encoding if the conversion from
+         RTL_TEXTENCODING_MS_932 fails, but you can get unlucky and get a document
+         which isn't really in RTL_TEXTENCODING_MS_932 but parts of it form
+         valid RTL_TEXTENCODING_MS_932 by chance :-(
+
+         We're not the only ones that struggle with this: Here's the help from
+         MSOffice 2003 on the topic:
+
+         <<
+          Earlier versions of Microsoft Word were sometimes used in conjunction with
+          third-party language-processing add-in programs designed to support Chinese or
+          Korean on English versions of Microsoft Windows. Use of these add-ins sometimes
+          results in incorrect text display in more recent versions of Word.
+
+          However, you can set options to convert these documents so that text is
+          displayed correctly. On the Tools menu, click Options, and then click the
+          General tab. In the English Word 6.0/95 documents list, select Contain Asian
+          text (to have Word interpret the text as Asian code page data, regardless of
+          its font) or Automatically detect Asian text (to have Word attempt to determine
+          which parts of the text are meant to be Asian).
+        >>
+
+        What we can try here is to ignore a RTL_TEXTENCODING_MS_932 codepage if
+        the language is not Japanese
+        */
+
+        const SfxPoolItem * pItem = GetFmtAttr(RES_CHRATR_CJK_LANGUAGE);
+        if (pItem != NULL && LANGUAGE_JAPANESE != static_cast<const SvxLanguageItem *>(pItem)->GetLanguage())
+        {
+            SAL_WARN("sw.ww8", "discarding word95 RTL_TEXTENCODING_MS_932 encoding");
+            eSrcCharSet = GetCharSetFromLanguage();
+        }
+    }
     const rtl_TextEncoding eSrcCJKCharSet = bVer67 ? GetCurrentCJKCharSet() :
         RTL_TEXTENCODING_MS_1252;