sw/qa/extras/rtfexport/data/tdf95706.rtf | 36 ++++++++++++++
sw/qa/extras/rtfexport/rtfexport4.cxx | 61 +++++++++++++++++++++++++
writerfilter/source/rtftok/rtfcharsets.cxx | 9 +++
writerfilter/source/rtftok/rtfcharsets.hxx | 13 +++++
writerfilter/source/rtftok/rtfdocumentimpl.cxx | 39 +++++++++++++++
5 files changed, 156 insertions(+), 2 deletions(-)
New commits:
commit 965313b9efc761c70aacf6e3ebee60ffa2b1d5dd
Author: Vasily Melenchuk <vasily.melenc...@cib.de>
AuthorDate: Tue Apr 5 19:13:05 2022 +0300
Commit: Miklos Vajna <vmik...@collabora.com>
CommitDate: Thu Apr 7 14:29:04 2022 +0200
tdf#95706: RTF import: Use fontname suffixes to detect encoding
Font names like "Arial CE", "Times New Roman Cyr" are not special
fonts. They are classical Arial, Times New Roman... And these
suffixes can be used to detect encoding used for RTF text.
Most interesting: for MS Word these suffixes have priority:
{\f34\cpg1253\fcharset161 Arial Baltic;} will have cp1257
and not cp1253.
Looks like compatibility issue came from dark ages.
Change-Id: Ife8e781d5d04c3f6a8c11fcf604357c74bf33055
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/132584
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmik...@collabora.com>
diff --git a/sw/qa/extras/rtfexport/data/tdf95706.rtf
b/sw/qa/extras/rtfexport/data/tdf95706.rtf
new file mode 100644
index 000000000000..64c97930441d
--- /dev/null
+++ b/sw/qa/extras/rtfexport/data/tdf95706.rtf
@@ -0,0 +1,36 @@
+{\rtf\ansi
+{\fonttbl
+{\f1 Arial Baltic;}
+{\f2 Arial CE;}
+{\f3 Arial Cyr;}
+{\f4 Arial Greek;}
+{\f5 Arial Tur;}
+{\f6 Arial (Hebrew);}
+{\f7 Arial (Arabic);}
+{\f8 Arial (Vietnamese);}
+{\f9 Arial BlaBlaBla;}
+
+{\f10\cpg1253\fcharset161 Arial;}
+{\f11\fcharset161 Arial;}
+{\f12\cpg1253 Arial;}
+{\f13\cpg1253\fcharset161 Arial Baltic;}
+{\f14 Arial Baltic;\cpg1253\fcharset161}
+
+}
+\pard Font name suffixes:\par
+\pard\f1\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f2\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f3\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f4\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f5\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f6\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f7\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f8\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f9\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard Font entry charset values:\par
+\pard\f10\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f11\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f12\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f13\fs26 \'c0\'c1\'c2\'c3\'c4\par
+\pard\f14\fs26 \'c0\'c1\'c2\'c3\'c4\par
+}
diff --git a/sw/qa/extras/rtfexport/rtfexport4.cxx
b/sw/qa/extras/rtfexport/rtfexport4.cxx
index 8f81b1565695..2d036538d6fe 100644
--- a/sw/qa/extras/rtfexport/rtfexport4.cxx
+++ b/sw/qa/extras/rtfexport/rtfexport4.cxx
@@ -492,6 +492,67 @@ CPPUNIT_TEST_FIXTURE(Test, testClearingBreak)
verify();
}
+DECLARE_RTFEXPORT_TEST(testTdf95706, "tdf95706.rtf")
+{
+ uno::Reference<text::XTextRange> xRun2
+ = getRun(getParagraph(2), 1, u"\u0104\u012e\u0100\u0106\u00c4");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun2,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun3
+ = getRun(getParagraph(3), 1, u"\u0154\u00c1\u00c2\u0102\u00c4");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun3,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun4
+ = getRun(getParagraph(4), 1, u"\u0410\u0411\u0412\u0413\u0414");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun4,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun5
+ = getRun(getParagraph(5), 1, u"\u0390\u0391\u0392\u0393\u0394");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun5,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun6
+ = getRun(getParagraph(6), 1, u"\u00c0\u00c1\u00c2\u00c3\u00c4");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun6,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun7
+ = getRun(getParagraph(7), 1, u"\u05b0\u05b1\u05b2\u05b3\u05b4");
+ // Do not check font for Hebrew: it can be substituted by smth able to
handle these chars
+ //CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun7,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun8
+ = getRun(getParagraph(8), 1, u"\u06c1\u0621\u0622\u0623\u0624");
+ // Do not check font for Arabic: it can be substituted by smth able to
handle these chars
+ //CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun8,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun9
+ = getRun(getParagraph(9), 1, u"\u00c0\u00c1\u00c2\u0102\u00c4");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun9,
"CharFontName"));
+
+ // Ensure strange font remains strange. No reason to check content: in
this case it can vary on locale
+ uno::Reference<text::XTextRange> xRun10 = getRun(getParagraph(10), 1);
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial BlaBlaBla"),
+ getProperty<OUString>(xRun10, "CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun12
+ = getRun(getParagraph(12), 1, u"\u0390\u0391\u0392\u0393\u0394");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun12,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun13
+ = getRun(getParagraph(13), 1, u"\u0390\u0391\u0392\u0393\u0394");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun13,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun14 = getRun(getParagraph(14), 1);
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun14,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun15
+ = getRun(getParagraph(15), 1, u"\u0104\u012e\u0100\u0106\u00c4");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun15,
"CharFontName"));
+
+ uno::Reference<text::XTextRange> xRun16
+ = getRun(getParagraph(16), 1, u"\u0104\u012e\u0100\u0106\u00c4");
+ CPPUNIT_ASSERT_EQUAL(OUString("Arial"), getProperty<OUString>(xRun16,
"CharFontName"));
+}
+
DECLARE_RTFEXPORT_TEST(testTdf111851, "tdf111851.rtf")
{
uno::Reference<text::XTextTable> xTable(getParagraphOrTable(1),
uno::UNO_QUERY);
diff --git a/writerfilter/source/rtftok/rtfcharsets.cxx
b/writerfilter/source/rtftok/rtfcharsets.cxx
index f7a03ee17dc9..14d27b5f2deb 100644
--- a/writerfilter/source/rtftok/rtfcharsets.cxx
+++ b/writerfilter/source/rtftok/rtfcharsets.cxx
@@ -9,6 +9,7 @@
#include "rtfcharsets.hxx"
#include <array>
+#include <rtl/textenc.h>
namespace writerfilter::rtftok
{
@@ -50,6 +51,14 @@ RTFEncoding const aRTFEncodings[] = {
int nRTFEncodings = std::size(aRTFEncodings);
+RTFFontNameSuffix const aRTFFontNameSuffixes[] = {
+ { "Baltic", RTL_TEXTENCODING_MS_1257 }, { "CE", RTL_TEXTENCODING_MS_1250
},
+ { "Cyr", RTL_TEXTENCODING_MS_1251 }, { "Greek",
RTL_TEXTENCODING_MS_1253 },
+ { "Tur", RTL_TEXTENCODING_MS_1254 }, { "(Hebrew)",
RTL_TEXTENCODING_MS_1255 },
+ { "(Arabic)", RTL_TEXTENCODING_MS_1256 }, { "(Vietnamese)",
RTL_TEXTENCODING_MS_1258 },
+ { "", RTL_TEXTENCODING_DONTKNOW } // End of array
+};
+
} // namespace writerfilter::rtftok
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */
diff --git a/writerfilter/source/rtftok/rtfcharsets.hxx
b/writerfilter/source/rtftok/rtfcharsets.hxx
index 865a9310289e..826dea271f6b 100644
--- a/writerfilter/source/rtftok/rtfcharsets.hxx
+++ b/writerfilter/source/rtftok/rtfcharsets.hxx
@@ -19,6 +19,19 @@ struct RTFEncoding
};
extern RTFEncoding const aRTFEncodings[];
extern int nRTFEncodings;
+
+/// Font name can contain special suffixes used
+/// to determine encoding for given font table entry
+/// For example "Arial CE" is "Arial" with CP1250 encoding
+/// List of these suffixes is not official and detected in a empirical
+/// way thus may be inexact and incomplete.
+struct RTFFontNameSuffix
+{
+ const char* suffix;
+ int codepage;
+};
+extern RTFFontNameSuffix const aRTFFontNameSuffixes[];
+
} // namespace writerfilter::rtftok
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */
diff --git a/writerfilter/source/rtftok/rtfdocumentimpl.cxx
b/writerfilter/source/rtftok/rtfdocumentimpl.cxx
index da53e2df1f24..5a19ccebb20a 100644
--- a/writerfilter/source/rtftok/rtfdocumentimpl.cxx
+++ b/writerfilter/source/rtftok/rtfdocumentimpl.cxx
@@ -50,6 +50,7 @@
#include "rtfskipdestination.hxx"
#include "rtftokenizer.hxx"
#include "rtflookahead.hxx"
+#include "rtfcharsets.hxx"
using namespace com::sun::star;
@@ -1370,14 +1371,48 @@ void RTFDocumentImpl::text(OUString& rString)
case Destination::FONTTABLE:
case Destination::FONTENTRY:
{
- m_aFontNames[m_nCurrentFontIndex] = aName;
+ // Old documents can contain no encoding information
in fontinfo,
+ // but there can be font name suffixes: Arial CE is
not a special
+ // font, it is ordinal Arial, but with used cp 1250
encoding.
+ // Moreover these suffixes have priority over \cpgN
and \fcharsetN
+ // in MS Word.
+ OUString aFontSuffix;
+ OUString aNameNoSuffix(aName);
+ sal_Int32 nLastSpace = aName.lastIndexOf(' ');
+ if (nLastSpace >= 0)
+ {
+ aFontSuffix = aName.copy(nLastSpace + 1);
+ aNameNoSuffix = aName.copy(0, nLastSpace);
+ sal_Int32 nEncoding = RTL_TEXTENCODING_DONTKNOW;
+ for (int i = 0;
+ aRTFFontNameSuffixes[i].codepage !=
RTL_TEXTENCODING_DONTKNOW; i++)
+ {
+ if
(aFontSuffix.equalsAscii(aRTFFontNameSuffixes[i].suffix))
+ {
+ nEncoding =
aRTFFontNameSuffixes[i].codepage;
+ break;
+ }
+ }
+ if (nEncoding > RTL_TEXTENCODING_DONTKNOW)
+ {
+ m_nCurrentEncoding = nEncoding;
+
m_aStates.top().setCurrentEncoding(m_nCurrentEncoding);
+ }
+ else
+ {
+ // Unknown suffix: looks like it is just a
part of font name, restore it
+ aNameNoSuffix = aName;
+ }
+ }
+
+ m_aFontNames[m_nCurrentFontIndex] = aNameNoSuffix;
if (m_nCurrentEncoding >= 0)
{
m_aFontEncodings[m_nCurrentFontIndex] =
m_nCurrentEncoding;
m_nCurrentEncoding = -1;
}
m_aStates.top().getTableAttributes().set(NS_ooxml::LN_CT_Font_name,
- new
RTFValue(aName));
+ new
RTFValue(aNameNoSuffix));
writerfilter::Reference<Properties>::Pointer_t const
pProp(
new
RTFReferenceProperties(m_aStates.top().getTableAttributes(),
