From 8cd856d9705fbcd61ad902859769fc98bf6d7a69 Mon Sep 17 00:00:00 2001
From: Michael Stahl <mstahl@redhat.com>
Date: Mon, 2 Jun 2014 23:57:13 +0200
Subject: fdo#79384: RTF import: fix literal Shift-JIS text

This is a variable-length encoding, and the second byte may be a RTF
syntax character like \, {, }.

Change-Id: I813ccafda18388af3bf05eb7ce9a0253c627b1c4
(cherry picked from commit 061190a62fcdbfb3a0b266d5afffbd257a3e692e)
---
 sw/qa/extras/rtfimport/data/fdo79384.rtf       |  9 +++++++
 sw/qa/extras/rtfimport/rtfimport.cxx           |  8 ++++++
 writerfilter/source/rtftok/rtfdocumentimpl.cxx | 34 +++++++++++++++++++++++---
 writerfilter/source/rtftok/rtfdocumentimpl.hxx |  2 +-
 4 files changed, 48 insertions(+), 5 deletions(-)
 create mode 100644 sw/qa/extras/rtfimport/data/fdo79384.rtf

diff --git a/sw/qa/extras/rtfimport/data/fdo79384.rtf b/sw/qa/extras/rtfimport/data/fdo79384.rtf
new file mode 100644
index 000000000000..2a900852861d
--- /dev/null
+++ b/sw/qa/extras/rtfimport/data/fdo79384.rtf
@@ -0,0 +1,9 @@
+{\rtf1\ansi
+{\fonttbl{\f5\fnil\fprq0\fcharset128 OpenSymbol{\*\falt Arial Unicode MS};}}
+
+\pard\plain
+
+\dbch\f5 „M„p„‚„{„u„‚„ „ƒ„„y„ƒ„{„p
+„}„\
+
+\par }
diff --git a/sw/qa/extras/rtfimport/rtfimport.cxx b/sw/qa/extras/rtfimport/rtfimport.cxx
index e96f03723f72..49e9687385bb 100644
--- a/sw/qa/extras/rtfimport/rtfimport.cxx
+++ b/sw/qa/extras/rtfimport/rtfimport.cxx
@@ -290,6 +290,14 @@ DECLARE_RTFIMPORT_TEST(testN751020, "n751020.rtf")
     CPPUNIT_ASSERT_EQUAL(sal_Int32(convertTwipToMm100(200)), getProperty<sal_Int32>(xParaEnum->nextElement(), "ParaBottomMargin"));
 }
 
+DECLARE_RTFIMPORT_TEST(testFdo79384, "fdo79384.rtf")
+{
+    uno::Reference<text::XTextRange> xTextRange = getRun(getParagraph(1), 1);
+
+    CPPUNIT_ASSERT_EQUAL(OUString("ΠœΠ°Ρ€ΠΊΠ΅Ρ€Ρ‹ спискамЫ", 31, RTL_TEXTENCODING_UTF8),
+            xTextRange->getString());
+}
+
 DECLARE_RTFIMPORT_TEST(testFdo47326, "fdo47326.rtf")
 {
     // This was 15 only, as \super buffered text, then the contents of it got lost.
diff --git a/writerfilter/source/rtftok/rtfdocumentimpl.cxx b/writerfilter/source/rtftok/rtfdocumentimpl.cxx
index 686901483b6a..cf989c79f2d2 100644
--- a/writerfilter/source/rtftok/rtfdocumentimpl.cxx
+++ b/writerfilter/source/rtftok/rtfdocumentimpl.cxx
@@ -980,9 +980,33 @@ int RTFDocumentImpl::resolveChars(char ch)
                 m_aStates.top().nCharsToSkip--;
             }
         }
+
         // read a single char if we're in hex mode
         if (m_aStates.top().nInternalState == INTERNAL_HEX)
             break;
+
+        if (RTFParserState::DBCH == m_aStates.top().eRunType &&
+            RTL_TEXTENCODING_MS_932 == m_aStates.top().nCurrentEncoding)
+        {
+            unsigned char uch = ch;
+            if ((uch >= 0x80 && uch <= 0x9F) || uch >= 0xE0)
+            {
+                // read second byte of 2-byte Shift-JIS - may be \ { }
+                Strm().ReadChar(ch);
+                if (m_aStates.top().nCharsToSkip == 0)
+                {
+                    assert(bUnicodeChecked);
+                    aBuf.append(ch);
+                }
+                else
+                {
+                    assert(bSkipped);
+                    // anybody who uses \ucN with Shift-JIS is insane
+                    m_aStates.top().nCharsToSkip--;
+                }
+            }
+        }
+
         Strm().ReadChar(ch);
     }
     if (m_aStates.top().nInternalState != INTERNAL_HEX && !Strm().IsEof())
@@ -2980,12 +3004,13 @@ int RTFDocumentImpl::dispatchFlag(RTFKeyword nKeyword)
         break;
     case RTF_LOCH:
         // Noop, dmapper detects this automatically.
+        m_aStates.top().eRunType = RTFParserState::LOCH;
         break;
     case RTF_HICH:
-        m_aStates.top().bIsCjk = true;
+        m_aStates.top().eRunType = RTFParserState::HICH;
         break;
     case RTF_DBCH:
-        m_aStates.top().bIsCjk = false;
+        m_aStates.top().eRunType = RTFParserState::DBCH;
         break;
     case RTF_TITLEPG:
     {
@@ -3471,7 +3496,8 @@ int RTFDocumentImpl::dispatchValue(RTFKeyword nKeyword, int nParam)
         if (nKeyword == RTF_F)
             nSprm = NS_ooxml::LN_CT_Fonts_ascii;
         else
-            nSprm = (m_aStates.top().bIsCjk ? NS_ooxml::LN_CT_Fonts_eastAsia : NS_ooxml::LN_CT_Fonts_cs);
+            nSprm = (m_aStates.top().eRunType == RTFParserState::HICH
+                ? NS_ooxml::LN_CT_Fonts_eastAsia : NS_ooxml::LN_CT_Fonts_cs);
         if (m_aStates.top().nDestinationState == DESTINATION_FONTTABLE || m_aStates.top().nDestinationState == DESTINATION_FONTENTRY)
         {
             m_aFontIndexes.push_back(nParam);
@@ -5714,7 +5740,7 @@ RTFParserState::RTFParserState(RTFDocumentImpl* pDocumentImpl)
       aShape(),
       aDrawingObject(),
       aFrame(this),
-      bIsCjk(false),
+      eRunType(LOCH),
       nYear(0),
       nMonth(0),
       nDay(0),
diff --git a/writerfilter/source/rtftok/rtfdocumentimpl.hxx b/writerfilter/source/rtftok/rtfdocumentimpl.hxx
index 3c7d2f318008..d069baa2ca1d 100644
--- a/writerfilter/source/rtftok/rtfdocumentimpl.hxx
+++ b/writerfilter/source/rtftok/rtfdocumentimpl.hxx
@@ -256,7 +256,7 @@ public:
     RTFFrame aFrame;
 
     /// CJK or CTL?
-    bool bIsCjk;
+    enum { LOCH, HICH, DBCH } eRunType;
 
     // Info group.
     int nYear;
-- 
cgit