[Libreoffice-commits] .: filter/inc filter/source
Caolán McNamara
caolan at kemper.freedesktop.org
Mon Aug 13 07:12:44 PDT 2012
filter/inc/filter/msfilter/util.hxx | 18 +++++
filter/source/msfilter/util.cxx | 127 ++++++++++++++++++++++++++++++++++++
2 files changed, 145 insertions(+)
New commits:
commit 1a8d78656a5773d285a430f6f88602beda173a8e
Author: Caolán McNamara <caolanm at redhat.com>
Date: Mon Aug 13 15:07:13 2012 +0100
The missing docs as to the ranges of the MS text categories
especially the definition of Complex Scripts
Change-Id: I693936f1dd150aefd4a632530439acf8edcbdd74
diff --git a/filter/inc/filter/msfilter/util.hxx b/filter/inc/filter/msfilter/util.hxx
index d772a24..607566a 100644
--- a/filter/inc/filter/msfilter/util.hxx
+++ b/filter/inc/filter/msfilter/util.hxx
@@ -70,6 +70,24 @@ MSFILTER_DLLPUBLIC rtl::OString DateTimeToOString( const DateTime& rDateTime );
MSFILTER_DLLPUBLIC sal_Unicode bestFitOpenSymbolToMSFont(sal_Unicode cBullet,
rtl_TextEncoding& r_ioChrSet, rtl::OUString& r_ioFontName, bool bDisableUnicodeSupport = false);
+
+enum TextCategory
+{
+ latin, //Latin
+ cs, //Complex Script
+ ea, //East Asian
+ sym //Symbol
+};
+
+/** Categorize codepoints according to how MS seems to do it.
+
+ It's been bugging me for ages as to what codepoint MS considers in
+ what category. Tom Jebo has a post suggesting the criteria used here
+ and indicating its been submitting to the standards working group
+ as a proposed resolution.
+*/
+MSFILTER_DLLPUBLIC TextCategory categorizeCodePoint(sal_uInt32 codePoint, const rtl::OUString &rBcp47LanguageTag);
+
}
}
diff --git a/filter/source/msfilter/util.cxx b/filter/source/msfilter/util.cxx
index ddb9f14..db8aa6c 100644
--- a/filter/source/msfilter/util.cxx
+++ b/filter/source/msfilter/util.cxx
@@ -179,6 +179,133 @@ sal_Unicode bestFitOpenSymbolToMSFont(sal_Unicode cChar,
return cChar;
}
+/*
+ http://social.msdn.microsoft.com/Forums/hu-HU/os_openXML-ecma/thread/1bf1f185-ee49-4314-94e7-f4e1563b5c00
+
+ The following information is being submitted to the standards working group as
+ a proposed resolution to a defect report and is not yet part of ISO 29500-1.
+ ...
+ For each Unicode character in DrawingML text, the font face can be any of four
+ font “slotsâ€: latin (§21.1.2.3.7), cs (§21.1.2.3.1), ea (§21.1.2.3.3), or sym
+ (§21.1.2.3.10), as specified in the following table. For all ranges not
+ explicitly called out below, the ea font shall be used.
+
+ U+0000–U+007F Use latin font
+ U+0080–U+00A6 Use latin font
+ U+00A9–U+00AF Use latin font
+ U+00B2–U+00B3 Use latin font
+ U+00B5–U+00D6 Use latin font
+ U+00D8–U+00F6 Use latin font
+ U+00F8–U+058F Use latin font
+ U+0590–U+074F Use cs font
+ U+0780–U+07BF Use cs font
+ U+0900–U+109F Use cs font
+ U+10A0–U+10FF Use latin font
+ U+1200–U+137F Use latin font
+ U+13A0–U+177F Use latin font
+ U+1D00–U+1D7F Use latin font
+ U+1E00–U+1FFF Use latin font
+ U+1780–U+18AF Use cs font
+ U+2000–U+200B Use latin font
+ U+200C–U+200F Use cs font
+ U+2010–U+2029 Use latin font Except, for the quote characters in the range
+ 2018 – 201E, use ea font if the text has one of the following language
+ identifiers: ii-CN, ja-JP, ko-KR, zh-CN, zh-HK, zh-MO, zh-SG, zh-TW.
+ U+202A–U+202F Use cs font
+ U+2030–U+2046 Use latin font
+ U+204A–U+245F Use latin font
+ U+2670–U+2671 Use cs font
+ U+27C0–U+2BFF Use latin font
+ U+3099–U+309A Use ea font
+ U+D835 Use latin font
+ U+F000–U+F0FF Symbol, use sym font
+ U+FB00–U+FB17 Use latin font
+ U+FB1D–U+FB4F Use cs font
+ U+FE50–U+FE6F Use latin font
+ Otherwise Use ea font
+*/
+TextCategory categorizeCodePoint(sal_uInt32 codePoint, const rtl::OUString &rBcp47LanguageTag)
+{
+ TextCategory eRet = ea;
+ if (codePoint <= 0x007F)
+ eRet = latin;
+ else if (0x0080 <= codePoint && codePoint <= 0x00A6)
+ eRet = latin;
+ else if (0x00A9 <= codePoint && codePoint <= 0x00AF)
+ eRet = latin;
+ else if (0x00B2 <= codePoint && codePoint <= 0x00B3)
+ eRet = latin;
+ else if (0x00B5 <= codePoint && codePoint <= 0x00D6)
+ eRet = latin;
+ else if (0x00D8 <= codePoint && codePoint <= 0x00F6)
+ eRet = latin;
+ else if (0x00F8 <= codePoint && codePoint <= 0x058F)
+ eRet = latin;
+ else if (0x0590 <= codePoint && codePoint <= 0x074F)
+ eRet = cs;
+ else if (0x0780 <= codePoint && codePoint <= 0x07BF)
+ eRet = cs;
+ else if (0x0900 <= codePoint && codePoint <= 0x109F)
+ eRet = cs;
+ else if (0x10A0 <= codePoint && codePoint <= 0x10FF)
+ eRet = latin;
+ else if (0x1200 <= codePoint && codePoint <= 0x137F)
+ eRet = latin;
+ else if (0x13A0 <= codePoint && codePoint <= 0x177F)
+ eRet = latin;
+ else if (0x1D00 <= codePoint && codePoint <= 0x1D7F)
+ eRet = latin;
+ else if (0x1E00 <= codePoint && codePoint <= 0x1FFF)
+ eRet = latin;
+ else if (0x1780 <= codePoint && codePoint <= 0x18AF)
+ eRet = cs;
+ else if (0x2000 <= codePoint && codePoint <= 0x200B)
+ eRet = latin;
+ else if (0x200C <= codePoint && codePoint <= 0x200F)
+ eRet = cs;
+ else if (0x2010 <= codePoint && codePoint <= 0x2029)
+ {
+ eRet = latin;
+ if (0x2018 <= codePoint && codePoint <= 0x201E)
+ {
+ if (rBcp47LanguageTag == "ii-CN" ||
+ rBcp47LanguageTag == "ja-JP" ||
+ rBcp47LanguageTag == "ko-KR" ||
+ rBcp47LanguageTag == "zh-CN" ||
+ rBcp47LanguageTag == "zh-HK" ||
+ rBcp47LanguageTag == "zh-MO" ||
+ rBcp47LanguageTag == "zh-SG" ||
+ rBcp47LanguageTag == "zh-TW")
+ {
+ eRet = ea;
+ }
+ }
+ }
+ else if (0x202A <= codePoint && codePoint <= 0x202F)
+ eRet = cs;
+ else if (0x2030 <= codePoint && codePoint <= 0x2046)
+ eRet = latin;
+ else if (0x204A <= codePoint && codePoint <= 0x245F)
+ eRet = latin;
+ else if (0x2670 <= codePoint && codePoint <= 0x2671)
+ eRet = latin;
+ else if (0x27C0 <= codePoint && codePoint <= 0x2BFF)
+ eRet = latin;
+ else if (0x3099 <= codePoint && codePoint <= 0x309A)
+ eRet = ea;
+ else if (0xD835 == codePoint)
+ eRet = latin;
+ else if (0xF000 <= codePoint && codePoint <= 0xF0FF)
+ eRet = sym;
+ else if (0xFB00 <= codePoint && codePoint <= 0xFB17)
+ eRet = latin;
+ else if (0xFB1D <= codePoint && codePoint <= 0xFB4F)
+ eRet = cs;
+ else if (0xFE50 <= codePoint && codePoint <= 0xFE6F)
+ eRet = latin;
+ return eRet;
+}
+
}
}
More information about the Libreoffice-commits
mailing list