//
// GfxFont.cc
//
-// Copyright 1996 Derek B. Noonburg
+// Copyright 1996-2003 Glyph & Cog, LLC
//
//========================================================================
-#ifdef __GNUC__
+#include <aconf.h>
+
+#ifdef USE_GCC_PRAGMAS
#pragma implementation
#endif
-#include <stdlib.h>
-#include <stddef.h>
#include <stdio.h>
+#include <stdlib.h>
#include <string.h>
#include <ctype.h>
-#include "GString.h"
#include "gmem.h"
-#include "gfile.h"
-#include "config.h"
+#include "Error.h"
#include "Object.h"
-#include "Array.h"
#include "Dict.h"
-#include "Error.h"
-#include "Params.h"
+#include "GlobalParams.h"
+#include "CMap.h"
+#include "CharCodeToUnicode.h"
+#include "FontEncodingTables.h"
+#include "BuiltinFontTables.h"
+#include "FoFiType1.h"
+#include "FoFiType1C.h"
+#include "FoFiTrueType.h"
#include "GfxFont.h"
-#include "FontInfo.h"
-#if JAPANESE_SUPPORT
-#include "CMapInfo.h"
-#endif
-
//------------------------------------------------------------------------
-static int CDECL cmpWidthExcep(const void *w1, const void *w2);
-
-//------------------------------------------------------------------------
+struct StdFontMapEntry {
+ char *altName;
+ char *properName;
+};
-static Gushort *defCharWidths[12] = {
- courierWidths,
- courierObliqueWidths,
- courierBoldWidths,
- courierBoldObliqueWidths,
- helveticaWidths,
- helveticaObliqueWidths,
- helveticaBoldWidths,
- helveticaBoldObliqueWidths,
- timesRomanWidths,
- timesItalicWidths,
- timesBoldWidths,
- timesBoldItalicWidths
+// Acrobat 4.0 and earlier substituted Base14-compatible fonts without
+// providing Widths and a FontDescriptor, so we munge the names into
+// the proper Base14 names. This table is from implementation note 44
+// in the PDF 1.4 spec, with some additions based on empirical
+// evidence.
+static StdFontMapEntry stdFontMap[] = {
+ { "Arial", "Helvetica" },
+ { "Arial,Bold", "Helvetica-Bold" },
+ { "Arial,BoldItalic", "Helvetica-BoldOblique" },
+ { "Arial,Italic", "Helvetica-Oblique" },
+ { "Arial-Bold", "Helvetica-Bold" },
+ { "Arial-BoldItalic", "Helvetica-BoldOblique" },
+ { "Arial-BoldItalicMT", "Helvetica-BoldOblique" },
+ { "Arial-BoldMT", "Helvetica-Bold" },
+ { "Arial-Italic", "Helvetica-Oblique" },
+ { "Arial-ItalicMT", "Helvetica-Oblique" },
+ { "ArialMT", "Helvetica" },
+ { "Courier,Bold", "Courier-Bold" },
+ { "Courier,Italic", "Courier-Oblique" },
+ { "Courier,BoldItalic", "Courier-BoldOblique" },
+ { "CourierNew", "Courier" },
+ { "CourierNew,Bold", "Courier-Bold" },
+ { "CourierNew,BoldItalic", "Courier-BoldOblique" },
+ { "CourierNew,Italic", "Courier-Oblique" },
+ { "CourierNew-Bold", "Courier-Bold" },
+ { "CourierNew-BoldItalic", "Courier-BoldOblique" },
+ { "CourierNew-Italic", "Courier-Oblique" },
+ { "CourierNewPS-BoldItalicMT", "Courier-BoldOblique" },
+ { "CourierNewPS-BoldMT", "Courier-Bold" },
+ { "CourierNewPS-ItalicMT", "Courier-Oblique" },
+ { "CourierNewPSMT", "Courier" },
+ { "Helvetica,Bold", "Helvetica-Bold" },
+ { "Helvetica,BoldItalic", "Helvetica-BoldOblique" },
+ { "Helvetica,Italic", "Helvetica-Oblique" },
+ { "Helvetica-BoldItalic", "Helvetica-BoldOblique" },
+ { "Helvetica-Italic", "Helvetica-Oblique" },
+ { "Symbol,Bold", "Symbol" },
+ { "Symbol,BoldItalic", "Symbol" },
+ { "Symbol,Italic", "Symbol" },
+ { "TimesNewRoman", "Times-Roman" },
+ { "TimesNewRoman,Bold", "Times-Bold" },
+ { "TimesNewRoman,BoldItalic", "Times-BoldItalic" },
+ { "TimesNewRoman,Italic", "Times-Italic" },
+ { "TimesNewRoman-Bold", "Times-Bold" },
+ { "TimesNewRoman-BoldItalic", "Times-BoldItalic" },
+ { "TimesNewRoman-Italic", "Times-Italic" },
+ { "TimesNewRomanPS", "Times-Roman" },
+ { "TimesNewRomanPS-Bold", "Times-Bold" },
+ { "TimesNewRomanPS-BoldItalic", "Times-BoldItalic" },
+ { "TimesNewRomanPS-BoldItalicMT", "Times-BoldItalic" },
+ { "TimesNewRomanPS-BoldMT", "Times-Bold" },
+ { "TimesNewRomanPS-Italic", "Times-Italic" },
+ { "TimesNewRomanPS-ItalicMT", "Times-Italic" },
+ { "TimesNewRomanPSMT", "Times-Roman" },
+ { "TimesNewRomanPSMT,Bold", "Times-Bold" },
+ { "TimesNewRomanPSMT,BoldItalic", "Times-BoldItalic" },
+ { "TimesNewRomanPSMT,Italic", "Times-Italic" }
};
//------------------------------------------------------------------------
-// GfxFontEncoding
+// GfxFont
//------------------------------------------------------------------------
-inline int GfxFontEncoding::hash(char *name) {
- int h;
+GfxFont *GfxFont::makeFont(XRef *xref, char *tagA, Ref idA, Dict *fontDict) {
+ GString *nameA;
+ GfxFont *font;
+ Object obj1;
- h = name[0];
- if (name[1])
- h = h * 61 + name[1];
- return h % gfxFontEncHashSize;
-}
+ // get base font name
+ nameA = NULL;
+ fontDict->lookup("BaseFont", &obj1);
+ if (obj1.isName()) {
+ nameA = new GString(obj1.getName());
+ }
+ obj1.free();
-GfxFontEncoding::GfxFontEncoding() {
- int i;
+ // get font type
+ font = NULL;
+ fontDict->lookup("Subtype", &obj1);
+ if (obj1.isName("Type1") || obj1.isName("MMType1")) {
+ font = new Gfx8BitFont(xref, tagA, idA, nameA, fontType1, fontDict);
+ } else if (obj1.isName("Type1C")) {
+ font = new Gfx8BitFont(xref, tagA, idA, nameA, fontType1C, fontDict);
+ } else if (obj1.isName("Type3")) {
+ font = new Gfx8BitFont(xref, tagA, idA, nameA, fontType3, fontDict);
+ } else if (obj1.isName("TrueType")) {
+ font = new Gfx8BitFont(xref, tagA, idA, nameA, fontTrueType, fontDict);
+ } else if (obj1.isName("Type0")) {
+ font = new GfxCIDFont(xref, tagA, idA, nameA, fontDict);
+ } else {
+ error(-1, "Unknown font type: '%s'",
+ obj1.isName() ? obj1.getName() : "???");
+ font = new Gfx8BitFont(xref, tagA, idA, nameA, fontUnknownType, fontDict);
+ }
+ obj1.free();
- encoding = (char **)gmalloc(256 * sizeof(char *));
- freeEnc = gTrue;
- for (i = 0; i < 256; ++i)
- encoding[i] = NULL;
- for (i = 0; i < gfxFontEncHashSize; ++i)
- hashTab[i] = -1;
+ return font;
}
-GfxFontEncoding::GfxFontEncoding(char **encoding1, int encSize) {
- int i;
+GfxFont::GfxFont(char *tagA, Ref idA, GString *nameA) {
+ ok = gFalse;
+ tag = new GString(tagA);
+ id = idA;
+ name = nameA;
+ origName = nameA;
+ embFontName = NULL;
+ extFontFile = NULL;
+}
- encoding = encoding1;
- freeEnc = gFalse;
- for (i = 0; i < gfxFontEncHashSize; ++i)
- hashTab[i] = -1;
- for (i = 0; i < encSize; ++i) {
- if (encoding[i])
- addChar1(i, encoding[i]);
+GfxFont::~GfxFont() {
+ delete tag;
+ if (origName && origName != name) {
+ delete origName;
+ }
+ if (name) {
+ delete name;
+ }
+ if (embFontName) {
+ delete embFontName;
+ }
+ if (extFontFile) {
+ delete extFontFile;
}
}
-void GfxFontEncoding::addChar(int code, char *name) {
- int h, i;
+void GfxFont::readFontDescriptor(XRef *xref, Dict *fontDict) {
+ Object obj1, obj2, obj3, obj4;
+ double t;
+ int i;
+
+ // assume Times-Roman by default (for substitution purposes)
+ flags = fontSerif;
- // replace character associated with code
- if (encoding[code]) {
- h = hash(encoding[code]);
- for (i = 0; i < gfxFontEncHashSize; ++i) {
- if (hashTab[h] == code) {
- hashTab[h] = -2;
- break;
+ embFontID.num = -1;
+ embFontID.gen = -1;
+ missingWidth = 0;
+
+ if (fontDict->lookup("FontDescriptor", &obj1)->isDict()) {
+
+ // get flags
+ if (obj1.dictLookup("Flags", &obj2)->isInt()) {
+ flags = obj2.getInt();
+ }
+ obj2.free();
+
+ // get name
+ obj1.dictLookup("FontName", &obj2);
+ if (obj2.isName()) {
+ embFontName = new GString(obj2.getName());
+ }
+ obj2.free();
+
+ // look for embedded font file
+ if (obj1.dictLookupNF("FontFile", &obj2)->isRef()) {
+ if (type == fontType1) {
+ embFontID = obj2.getRef();
+ } else {
+ error(-1, "Mismatch between font type and embedded font file");
}
- if (++h == gfxFontEncHashSize)
- h = 0;
}
- gfree(encoding[code]);
- }
+ obj2.free();
+ if (embFontID.num == -1 &&
+ obj1.dictLookupNF("FontFile2", &obj2)->isRef()) {
+ if (type == fontTrueType || type == fontCIDType2) {
+ embFontID = obj2.getRef();
+ } else {
+ error(-1, "Mismatch between font type and embedded font file");
+ }
+ }
+ obj2.free();
+ if (embFontID.num == -1 &&
+ obj1.dictLookupNF("FontFile3", &obj2)->isRef()) {
+ if (obj2.fetch(xref, &obj3)->isStream()) {
+ obj3.streamGetDict()->lookup("Subtype", &obj4);
+ if (obj4.isName("Type1")) {
+ if (type == fontType1) {
+ embFontID = obj2.getRef();
+ } else {
+ error(-1, "Mismatch between font type and embedded font file");
+ }
+ } else if (obj4.isName("Type1C")) {
+ if (type == fontType1) {
+ type = fontType1C;
+ embFontID = obj2.getRef();
+ } else if (type == fontType1C) {
+ embFontID = obj2.getRef();
+ } else {
+ error(-1, "Mismatch between font type and embedded font file");
+ }
+ } else if (obj4.isName("TrueType")) {
+ if (type == fontTrueType) {
+ embFontID = obj2.getRef();
+ } else {
+ error(-1, "Mismatch between font type and embedded font file");
+ }
+ } else if (obj4.isName("CIDFontType0C")) {
+ if (type == fontCIDType0) {
+ type = fontCIDType0C;
+ embFontID = obj2.getRef();
+ } else {
+ error(-1, "Mismatch between font type and embedded font file");
+ }
+ } else {
+ error(-1, "Unknown embedded font type '%s'",
+ obj4.isName() ? obj4.getName() : "???");
+ }
+ obj4.free();
+ }
+ obj3.free();
+ }
+ obj2.free();
+
+ // look for MissingWidth
+ obj1.dictLookup("MissingWidth", &obj2);
+ if (obj2.isNum()) {
+ missingWidth = obj2.getNum();
+ }
+ obj2.free();
+
+ // get Ascent and Descent
+ obj1.dictLookup("Ascent", &obj2);
+ if (obj2.isNum()) {
+ t = 0.001 * obj2.getNum();
+ // some broken font descriptors set ascent and descent to 0
+ if (t != 0) {
+ ascent = t;
+ }
+ }
+ obj2.free();
+ obj1.dictLookup("Descent", &obj2);
+ if (obj2.isNum()) {
+ t = 0.001 * obj2.getNum();
+ // some broken font descriptors set ascent and descent to 0
+ if (t != 0) {
+ descent = t;
+ }
+ // some broken font descriptors specify a positive descent
+ if (descent > 0) {
+ descent = -descent;
+ }
+ }
+ obj2.free();
- // associate name with code
- encoding[code] = name;
+ // font FontBBox
+ if (obj1.dictLookup("FontBBox", &obj2)->isArray()) {
+ for (i = 0; i < 4 && i < obj2.arrayGetLength(); ++i) {
+ if (obj2.arrayGet(i, &obj3)->isNum()) {
+ fontBBox[i] = 0.001 * obj3.getNum();
+ }
+ obj3.free();
+ }
+ }
+ obj2.free();
- // insert name in hash table
- addChar1(code, name);
+ }
+ obj1.free();
}
-void GfxFontEncoding::addChar1(int code, char *name) {
- int h, i, code2;
-
- // insert name in hash table
- h = hash(name);
- for (i = 0; i < gfxFontEncHashSize; ++i) {
- code2 = hashTab[h];
- if (code2 < 0) {
- hashTab[h] = code;
- break;
- } else if (encoding[code2] && !strcmp(encoding[code2], name)) {
- // keep the highest code for each char -- this is needed because
- // X won't display chars with codes < 32
- if (code > code2)
- hashTab[h] = code;
- break;
- }
- if (++h == gfxFontEncHashSize)
- h = 0;
+CharCodeToUnicode *GfxFont::readToUnicodeCMap(Dict *fontDict, int nBits,
+ CharCodeToUnicode *ctu) {
+ GString *buf;
+ Object obj1;
+ int c;
+
+ if (!fontDict->lookup("ToUnicode", &obj1)->isStream()) {
+ obj1.free();
+ return NULL;
}
+ buf = new GString();
+ obj1.streamReset();
+ while ((c = obj1.streamGetChar()) != EOF) {
+ buf->append(c);
+ }
+ obj1.streamClose();
+ obj1.free();
+ if (ctu) {
+ ctu->mergeCMap(buf, nBits);
+ } else {
+ ctu = CharCodeToUnicode::parseCMap(buf, nBits);
+ }
+ delete buf;
+ return ctu;
}
-GfxFontEncoding::~GfxFontEncoding() {
- int i;
+void GfxFont::findExtFontFile() {
+ static char *type1Exts[] = { ".pfa", ".pfb", ".ps", "", NULL };
+ static char *ttExts[] = { ".ttf", NULL };
- if (freeEnc) {
- for (i = 0; i < 256; ++i) {
- if (encoding[i])
- gfree(encoding[i]);
+ if (name) {
+ if (type == fontType1) {
+ extFontFile = globalParams->findFontFile(name, type1Exts);
+ } else if (type == fontTrueType) {
+ extFontFile = globalParams->findFontFile(name, ttExts);
}
- gfree(encoding);
}
}
-int GfxFontEncoding::getCharCode(char *name) {
- int h, i, code;
+char *GfxFont::readExtFontFile(int *len) {
+ FILE *f;
+ char *buf;
- h = hash(name);
- for (i = 0; i < gfxFontEncHashSize; ++i) {
- code = hashTab[h];
- if (code == -1 ||
- (code > 0 && encoding[code] && !strcmp(encoding[code], name)))
- return code;
- if (++h >= gfxFontEncHashSize)
- h = 0;
+ if (!(f = fopen(extFontFile->getCString(), "rb"))) {
+ error(-1, "External font file '%s' vanished", extFontFile->getCString());
+ return NULL;
}
- return -1;
+ fseek(f, 0, SEEK_END);
+ *len = (int)ftell(f);
+ fseek(f, 0, SEEK_SET);
+ buf = (char *)gmalloc(*len);
+ if ((int)fread(buf, 1, *len, f) != *len) {
+ error(-1, "Error reading external font file '%s'",
+ extFontFile->getCString());
+ }
+ fclose(f);
+ return buf;
+}
+
+char *GfxFont::readEmbFontFile(XRef *xref, int *len) {
+ char *buf;
+ Object obj1, obj2;
+ Stream *str;
+ int c;
+ int size, i;
+
+ obj1.initRef(embFontID.num, embFontID.gen);
+ obj1.fetch(xref, &obj2);
+ if (!obj2.isStream()) {
+ error(-1, "Embedded font file is not a stream");
+ obj2.free();
+ obj1.free();
+ embFontID.num = -1;
+ return NULL;
+ }
+ str = obj2.getStream();
+
+ buf = NULL;
+ i = size = 0;
+ str->reset();
+ while ((c = str->getChar()) != EOF) {
+ if (i == size) {
+ size += 4096;
+ buf = (char *)grealloc(buf, size);
+ }
+ buf[i++] = c;
+ }
+ *len = i;
+ str->close();
+
+ obj2.free();
+ obj1.free();
+
+ return buf;
}
//------------------------------------------------------------------------
-// GfxFont
+// Gfx8BitFont
//------------------------------------------------------------------------
-GfxFont::GfxFont(char *tag1, Ref id1, Dict *fontDict) {
+Gfx8BitFont::Gfx8BitFont(XRef *xref, char *tagA, Ref idA, GString *nameA,
+ GfxFontType typeA, Dict *fontDict):
+ GfxFont(tagA, idA, nameA)
+{
BuiltinFont *builtinFont;
- char buf[256];
+ char **baseEnc;
+ GBool baseEncFromFontFile;
+ char *buf;
+ int len;
+ FoFiType1 *ffT1;
+ FoFiType1C *ffT1C;
+ int code, code2;
+ char *charName;
+ GBool missing, hex;
+ Unicode toUnicode[256];
+ CharCodeToUnicode *utu, *ctu2;
+ Unicode uBuf[8];
+ double mul;
+ int firstChar, lastChar;
+ Gushort w;
Object obj1, obj2, obj3;
- char *p1, *p2;
- int i;
+ int n, i, a, b, m;
- // get font tag and ID
- tag = new GString(tag1);
- id = id1;
+ type = typeA;
+ ctu = NULL;
- // get base font name
- name = NULL;
- fontDict->lookup("BaseFont", &obj1);
- if (obj1.isName())
- name = new GString(obj1.getName());
- obj1.free();
+ // do font name substitution for various aliases of the Base 14 font
+ // names
+ if (name) {
+ a = 0;
+ b = sizeof(stdFontMap) / sizeof(StdFontMapEntry);
+ // invariant: stdFontMap[a].altName <= name < stdFontMap[b].altName
+ while (b - a > 1) {
+ m = (a + b) / 2;
+ if (name->cmp(stdFontMap[m].altName) >= 0) {
+ a = m;
+ } else {
+ b = m;
+ }
+ }
+ if (!name->cmp(stdFontMap[a].altName)) {
+ name = new GString(stdFontMap[a].properName);
+ }
+ }
// is it a built-in font?
builtinFont = NULL;
if (name) {
- for (i = 0; i < numBuiltinFonts; ++i) {
- if (!strcmp(builtinFonts[i].name, name->getCString())) {
+ for (i = 0; i < nBuiltinFonts; ++i) {
+ if (!name->cmp(builtinFonts[i].name)) {
builtinFont = &builtinFonts[i];
break;
}
}
}
- // get font type
- type = fontUnknownType;
- fontDict->lookup("Subtype", &obj1);
- if (obj1.isName("Type1"))
- type = fontType1;
- else if (obj1.isName("Type3"))
- type = fontType3;
- else if (obj1.isName("TrueType"))
- type = fontTrueType;
- else if (obj1.isName("Type0"))
- type = fontType0;
- obj1.free();
- is16 = gFalse;
+ // default ascent/descent values
+ if (builtinFont) {
+ ascent = 0.001 * builtinFont->ascent;
+ descent = 0.001 * builtinFont->descent;
+ fontBBox[0] = 0.001 * builtinFont->bbox[0];
+ fontBBox[1] = 0.001 * builtinFont->bbox[1];
+ fontBBox[2] = 0.001 * builtinFont->bbox[2];
+ fontBBox[3] = 0.001 * builtinFont->bbox[3];
+ } else {
+ ascent = 0.95;
+ descent = -0.35;
+ fontBBox[0] = fontBBox[1] = fontBBox[2] = fontBBox[3] = 0;
+ }
// get info from font descriptor
- // for flags: assume Times-Roman (or TimesNewRoman), but
- // explicitly check for Arial and CourierNew -- certain PDF
- // generators apparently don't include FontDescriptors for Arial,
- // TimesNewRoman, and CourierNew
- flags = fontSerif; // assume Times-Roman by default
- if (type == fontTrueType && !name->cmp("Arial"))
- flags = 0;
- else if (type == fontTrueType && !name->cmp("CourierNew"))
- flags = fontFixedWidth;
- embFontID.num = -1;
- embFontID.gen = -1;
- embFontName = NULL;
- extFontFile = NULL;
- fontDict->lookup("FontDescriptor", &obj1);
- if (obj1.isDict()) {
+ readFontDescriptor(xref, fontDict);
- // flags
- obj1.dictLookup("Flags", &obj2);
- if (obj2.isInt())
- flags = obj2.getInt();
- obj2.free();
-
- // embedded Type 1 font file and font name
- if (type == fontType1) {
- obj1.dictLookupNF("FontFile", &obj2);
- if (obj2.isRef()) {
- embFontID = obj2.getRef();
-
- // get font name from the font file itself since font subsets
- // sometimes use the 'AAAAAA+foo' name and sometimes use just 'foo'
- obj2.fetch(&obj3);
- if (obj3.isStream()) {
- obj3.streamReset();
- for (i = 0; i < 64; ++i) {
- obj3.streamGetLine(buf, sizeof(buf));
- if (!strncmp(buf, "/FontName", 9)) {
- if ((p1 = strchr(buf+9, '/'))) {
- ++p1;
- for (p2 = p1; *p2 && !isspace(*p2); ++p2) ;
- embFontName = new GString(p1, p2 - p1);
- }
- break;
- }
- }
- }
- obj3.free();
- obj2.free();
-
- // couldn't find font name so just use the one in the PDF font
- // descriptor
- if (!embFontName) {
- obj1.dictLookup("FontName", &obj2);
- if (obj2.isName())
- embFontName = new GString(obj2.getName());
- }
- }
- obj2.free();
-
- // embedded TrueType font file
- } else if (type == fontTrueType) {
- obj1.dictLookupNF("FontFile2", &obj2);
- if (obj2.isRef())
- embFontID = obj2.getRef();
- obj2.free();
- }
- }
- obj1.free();
+ // look for an external font file
+ findExtFontFile();
// get font matrix
fontMat[0] = fontMat[3] = 1;
fontMat[1] = fontMat[2] = fontMat[4] = fontMat[5] = 0;
if (fontDict->lookup("FontMatrix", &obj1)->isArray()) {
for (i = 0; i < 6 && i < obj1.arrayGetLength(); ++i) {
- if (obj1.arrayGet(i, &obj2)->isNum())
+ if (obj1.arrayGet(i, &obj2)->isNum()) {
fontMat[i] = obj2.getNum();
+ }
obj2.free();
}
}
obj1.free();
- // get encoding and character widths
- if (type == fontType0) {
- getType0EncAndWidths(fontDict);
- } else if (builtinFont) {
- makeEncoding(fontDict, builtinFont->encoding);
- makeWidths(fontDict, builtinFont->encoding, builtinFont->widths);
- } else {
- makeEncoding(fontDict, NULL);
- makeWidths(fontDict, NULL, NULL);
+ // get Type 3 bounding box, font definition, and resources
+ if (type == fontType3) {
+ if (fontDict->lookup("FontBBox", &obj1)->isArray()) {
+ for (i = 0; i < 4 && i < obj1.arrayGetLength(); ++i) {
+ if (obj1.arrayGet(i, &obj2)->isNum()) {
+ fontBBox[i] = obj2.getNum();
+ }
+ obj2.free();
+ }
+ }
+ obj1.free();
+ if (!fontDict->lookup("CharProcs", &charProcs)->isDict()) {
+ error(-1, "Missing or invalid CharProcs dictionary in Type 3 font");
+ charProcs.free();
+ }
+ if (!fontDict->lookup("Resources", &resources)->isDict()) {
+ resources.free();
+ }
}
-}
-
-GfxFont::~GfxFont() {
- delete tag;
- if (name)
- delete name;
- if (!is16 && encoding)
- delete encoding;
- if (embFontName)
- delete embFontName;
- if (extFontFile)
- delete extFontFile;
- if (is16)
- gfree(widths16.exceps);
-}
-
-double GfxFont::getWidth(GString *s) {
- double w;
- int i;
- w = 0;
- for (i = 0; i < s->getLength(); ++i)
- w += widths[s->getChar(i) & 0xff];
- return w;
-}
+ //----- build the font encoding -----
+
+ // Encodings start with a base encoding, which can come from
+ // (in order of priority):
+ // 1. FontDict.Encoding or FontDict.Encoding.BaseEncoding
+ // - MacRoman / MacExpert / WinAnsi / Standard
+ // 2. embedded or external font file
+ // 3. default:
+ // - builtin --> builtin encoding
+ // - TrueType --> MacRomanEncoding
+ // - others --> StandardEncoding
+ // and then add a list of differences (if any) from
+ // FontDict.Encoding.Differences.
+
+ // check FontDict for base encoding
+ hasEncoding = gFalse;
+ usesMacRomanEnc = gFalse;
+ baseEnc = NULL;
+ baseEncFromFontFile = gFalse;
+ fontDict->lookup("Encoding", &obj1);
+ if (obj1.isDict()) {
+ obj1.dictLookup("BaseEncoding", &obj2);
+ if (obj2.isName("MacRomanEncoding")) {
+ hasEncoding = gTrue;
+ usesMacRomanEnc = gTrue;
+ baseEnc = macRomanEncoding;
+ } else if (obj2.isName("MacExpertEncoding")) {
+ hasEncoding = gTrue;
+ baseEnc = macExpertEncoding;
+ } else if (obj2.isName("WinAnsiEncoding")) {
+ hasEncoding = gTrue;
+ baseEnc = winAnsiEncoding;
+ } else if (obj2.isName("StandardEncoding")) {
+ hasEncoding = gTrue;
+ baseEnc = standardEncoding;
+ }
+ obj2.free();
+ } else if (obj1.isName("MacRomanEncoding")) {
+ hasEncoding = gTrue;
+ usesMacRomanEnc = gTrue;
+ baseEnc = macRomanEncoding;
+ } else if (obj1.isName("MacExpertEncoding")) {
+ hasEncoding = gTrue;
+ baseEnc = macExpertEncoding;
+ } else if (obj1.isName("WinAnsiEncoding")) {
+ hasEncoding = gTrue;
+ baseEnc = winAnsiEncoding;
+ } else if (obj1.isName("StandardEncoding")) {
+ hasEncoding = gTrue;
+ baseEnc = standardEncoding;
+ }
-double GfxFont::getWidth16(int c) {
- double w;
- int a, b, m;
-
- w = widths16.defWidth;
- a = -1;
- b = widths16.numExceps;
- // invariant: widths16.exceps[a].last < c < widths16.exceps[b].first
- while (b - a > 1) {
- m = (a + b) / 2;
- if (widths16.exceps[m].last < c) {
- a = m;
- } else if (c < widths16.exceps[m].first) {
- b = m;
+ // check embedded or external font file for base encoding
+ // (only for Type 1 fonts - trying to get an encoding out of a
+ // TrueType font is a losing proposition)
+ ffT1 = NULL;
+ ffT1C = NULL;
+ buf = NULL;
+ if (type == fontType1 && (extFontFile || embFontID.num >= 0)) {
+ if (extFontFile) {
+ ffT1 = FoFiType1::load(extFontFile->getCString());
+ } else {
+ buf = readEmbFontFile(xref, &len);
+ ffT1 = FoFiType1::make(buf, len);
+ }
+ if (ffT1) {
+ if (ffT1->getName()) {
+ if (embFontName) {
+ delete embFontName;
+ }
+ embFontName = new GString(ffT1->getName());
+ }
+ if (!baseEnc) {
+ baseEnc = ffT1->getEncoding();
+ baseEncFromFontFile = gTrue;
+ }
+ }
+ } else if (type == fontType1C && (extFontFile || embFontID.num >= 0)) {
+ if (extFontFile) {
+ ffT1C = FoFiType1C::load(extFontFile->getCString());
} else {
- w = widths16.exceps[m].width;
- break;
+ buf = readEmbFontFile(xref, &len);
+ ffT1C = FoFiType1C::make(buf, len);
+ }
+ if (ffT1C) {
+ if (ffT1C->getName()) {
+ if (embFontName) {
+ delete embFontName;
+ }
+ embFontName = new GString(ffT1C->getName());
+ }
+ if (!baseEnc) {
+ baseEnc = ffT1C->getEncoding();
+ baseEncFromFontFile = gTrue;
+ }
}
}
- return w;
-}
-
-double GfxFont::getWidth16(GString *s) {
- double w;
- int c;
- int i;
-
- w = 0;
- for (i = 0; i < s->getLength(); i += 2) {
- c = (s->getChar(i) << 8) + s->getChar(i+1);
- w += getWidth16(c);
+ if (buf) {
+ gfree(buf);
}
- return w;
-}
-void GfxFont::makeEncoding(Dict *fontDict, GfxFontEncoding *builtinEncoding) {
- GfxFontEncoding *baseEnc;
- Object obj1, obj2, obj3;
- char *charName;
- int code, i;
+ // get default base encoding
+ if (!baseEnc) {
+ if (builtinFont) {
+ baseEnc = builtinFont->defaultBaseEnc;
+ hasEncoding = gTrue;
+ } else if (type == fontTrueType) {
+ baseEnc = winAnsiEncoding;
+ } else {
+ baseEnc = standardEncoding;
+ }
+ }
- // start with empty encoding
- encoding = new GfxFontEncoding();
+ // copy the base encoding
+ for (i = 0; i < 256; ++i) {
+ enc[i] = baseEnc[i];
+ if ((encFree[i] = baseEncFromFontFile) && enc[i]) {
+ enc[i] = copyString(baseEnc[i]);
+ }
+ }
- // get encoding from font dict
- fontDict->lookup("Encoding", &obj1);
+ // some Type 1C font files have empty encodings, which can break the
+ // T1C->T1 conversion (since the 'seac' operator depends on having
+ // the accents in the encoding), so we fill in any gaps from
+ // StandardEncoding
+ if (type == fontType1C && (extFontFile || embFontID.num >= 0) &&
+ baseEncFromFontFile) {
+ for (i = 0; i < 256; ++i) {
+ if (!enc[i] && standardEncoding[i]) {
+ enc[i] = standardEncoding[i];
+ encFree[i] = gFalse;
+ }
+ }
+ }
- // encoding specified by dictionary
+ // merge differences into encoding
if (obj1.isDict()) {
- obj1.dictLookup("BaseEncoding", &obj2);
- baseEnc = makeEncoding1(obj2, fontDict, builtinEncoding);
- obj2.free();
obj1.dictLookup("Differences", &obj2);
if (obj2.isArray()) {
+ hasEncoding = gTrue;
code = 0;
for (i = 0; i < obj2.arrayGetLength(); ++i) {
obj2.arrayGet(i, &obj3);
if (obj3.isInt()) {
code = obj3.getInt();
} else if (obj3.isName()) {
- if (code < 256)
- encoding->addChar(code, copyString(obj3.getName()));
+ if (code >= 0 && code < 256) {
+ if (encFree[code]) {
+ gfree(enc[code]);
+ }
+ enc[code] = copyString(obj3.getName());
+ encFree[code] = gTrue;
+ }
++code;
} else {
error(-1, "Wrong type in font encoding resource differences (%s)",
}
}
obj2.free();
-
- // encoding specified by name or null
- } else {
- baseEnc = makeEncoding1(obj1, fontDict, builtinEncoding);
}
-
- // free the font dict encoding
obj1.free();
+ if (ffT1) {
+ delete ffT1;
+ }
+ if (ffT1C) {
+ delete ffT1C;
+ }
+
+ //----- build the mapping to Unicode -----
- // merge base encoding and differences;
+ // pass 1: use the name-to-Unicode mapping table
+ missing = hex = gFalse;
for (code = 0; code < 256; ++code) {
- if (!encoding->getCharName(code)) {
- if ((charName = baseEnc->getCharName(code)))
- encoding->addChar(code, copyString(charName));
+ if ((charName = enc[code])) {
+ if (!(toUnicode[code] = globalParams->mapNameToUnicode(charName)) &&
+ strcmp(charName, ".notdef")) {
+ // if it wasn't in the name-to-Unicode table, check for a
+ // name that looks like 'Axx' or 'xx', where 'A' is any letter
+ // and 'xx' is two hex digits
+ if ((strlen(charName) == 3 &&
+ isalpha(charName[0]) &&
+ isxdigit(charName[1]) && isxdigit(charName[2]) &&
+ ((charName[1] >= 'a' && charName[1] <= 'f') ||
+ (charName[1] >= 'A' && charName[1] <= 'F') ||
+ (charName[2] >= 'a' && charName[2] <= 'f') ||
+ (charName[2] >= 'A' && charName[2] <= 'F'))) ||
+ (strlen(charName) == 2 &&
+ isxdigit(charName[0]) && isxdigit(charName[1]) &&
+ ((charName[0] >= 'a' && charName[0] <= 'f') ||
+ (charName[0] >= 'A' && charName[0] <= 'F') ||
+ (charName[1] >= 'a' && charName[1] <= 'f') ||
+ (charName[1] >= 'A' && charName[1] <= 'F')))) {
+ hex = gTrue;
+ }
+ missing = gTrue;
+ }
+ } else {
+ toUnicode[code] = 0;
}
}
-}
-GfxFontEncoding *GfxFont::makeEncoding1(Object obj, Dict *fontDict,
- GfxFontEncoding *builtinEncoding) {
- GfxFontEncoding *enc;
- GBool haveEncoding;
- Object obj1, obj2;
- char **path;
- FILE *f;
- FileStream *str;
+ // pass 2: try to fill in the missing chars, looking for names of
+ // the form 'Axx', 'xx', 'Ann', 'ABnn', or 'nn', where 'A' and 'B'
+ // are any letters, 'xx' is two hex digits, and 'nn' is 2-4
+ // decimal digits
+ if (missing && globalParams->getMapNumericCharNames()) {
+ for (code = 0; code < 256; ++code) {
+ if ((charName = enc[code]) && !toUnicode[code] &&
+ strcmp(charName, ".notdef")) {
+ n = strlen(charName);
+ code2 = -1;
+ if (hex && n == 3 && isalpha(charName[0]) &&
+ isxdigit(charName[1]) && isxdigit(charName[2])) {
+ sscanf(charName+1, "%x", &code2);
+ } else if (hex && n == 2 &&
+ isxdigit(charName[0]) && isxdigit(charName[1])) {
+ sscanf(charName, "%x", &code2);
+ } else if (!hex && n >= 2 && n <= 4 &&
+ isdigit(charName[0]) && isdigit(charName[1])) {
+ code2 = atoi(charName);
+ } else if (n >= 3 && n <= 5 &&
+ isdigit(charName[1]) && isdigit(charName[2])) {
+ code2 = atoi(charName+1);
+ } else if (n >= 4 && n <= 6 &&
+ isdigit(charName[2]) && isdigit(charName[3])) {
+ code2 = atoi(charName+2);
+ }
+ if (code2 >= 0 && code2 <= 0xff) {
+ toUnicode[code] = (Unicode)code2;
+ }
+ }
+ }
+ }
- // MacRoman, WinAnsi, or Standard encoding
- if (obj.isName("MacRomanEncoding")) {
- enc = &macRomanEncoding;
- } else if (obj.isName("WinAnsiEncoding")) {
- enc = &winAnsiEncoding;
- } else if (obj.isName("StandardEncoding")) {
- enc = &standardEncoding;
+ // construct the char code -> Unicode mapping object
+ ctu = CharCodeToUnicode::make8BitToUnicode(toUnicode);
- // use the built-in font encoding if possible
- } else if (builtinEncoding) {
- enc = builtinEncoding;
+ // merge in a ToUnicode CMap, if there is one -- this overwrites
+ // existing entries in ctu, i.e., the ToUnicode CMap takes
+ // precedence, but the other encoding info is allowed to fill in any
+ // holes
+ readToUnicodeCMap(fontDict, 8, ctu);
- // check font type
- } else {
+ // look for a Unicode-to-Unicode mapping
+ if (name && (utu = globalParams->getUnicodeToUnicode(name))) {
+ for (i = 0; i < 256; ++i) {
+ toUnicode[i] = 0;
+ }
+ ctu2 = CharCodeToUnicode::make8BitToUnicode(toUnicode);
+ for (i = 0; i < 256; ++i) {
+ n = ctu->mapToUnicode((CharCode)i, uBuf, 8);
+ if (n >= 1) {
+ n = utu->mapToUnicode((CharCode)uBuf[0], uBuf, 8);
+ if (n >= 1) {
+ ctu2->setMapping((CharCode)i, uBuf, n);
+ }
+ }
+ }
+ utu->decRefCnt();
+ delete ctu;
+ ctu = ctu2;
+ }
- // Type 1 font: try to get encoding from font file
- if (type == fontType1) {
+ //----- get the character widths -----
- // default to using standard encoding
- enc = &standardEncoding;
-
- // is there an external font file?
- haveEncoding = gFalse;
- if (name) {
- for (path = fontPath; *path; ++path) {
- extFontFile = appendToPath(new GString(*path), name->getCString());
- f = fopen(extFontFile->getCString(), "rb");
- if (!f) {
- extFontFile->append(".pfb");
- f = fopen(extFontFile->getCString(), "rb");
- }
- if (!f) {
- extFontFile->del(extFontFile->getLength() - 4, 4);
- extFontFile->append(".pfa");
- f = fopen(extFontFile->getCString(), "rb");
- }
- if (f) {
- obj1.initNull();
- str = new FileStream(f, 0, -1, &obj1);
- getType1Encoding(str);
- delete str;
- fclose(f);
- haveEncoding = gTrue;
- break;
- }
- delete extFontFile;
- extFontFile = NULL;
+ // initialize all widths
+ for (code = 0; code < 256; ++code) {
+ widths[code] = missingWidth * 0.001;
+ }
+
+ // use widths from font dict, if present
+ fontDict->lookup("FirstChar", &obj1);
+ firstChar = obj1.isInt() ? obj1.getInt() : 0;
+ obj1.free();
+ if (firstChar < 0 || firstChar > 255) {
+ firstChar = 0;
+ }
+ fontDict->lookup("LastChar", &obj1);
+ lastChar = obj1.isInt() ? obj1.getInt() : 255;
+ obj1.free();
+ if (lastChar < 0 || lastChar > 255) {
+ lastChar = 255;
+ }
+ mul = (type == fontType3) ? fontMat[0] : 0.001;
+ fontDict->lookup("Widths", &obj1);
+ if (obj1.isArray()) {
+ flags |= fontFixedWidth;
+ if (obj1.arrayGetLength() < lastChar - firstChar + 1) {
+ lastChar = firstChar + obj1.arrayGetLength() - 1;
+ }
+ for (code = firstChar; code <= lastChar; ++code) {
+ obj1.arrayGet(code - firstChar, &obj2);
+ if (obj2.isNum()) {
+ widths[code] = obj2.getNum() * mul;
+ if (widths[code] != widths[firstChar]) {
+ flags &= ~fontFixedWidth;
}
}
+ obj2.free();
+ }
- // is there an embedded font file?
- // (this has to be checked after the external font because
- // XOutputDev needs the encoding from the external font)
- if (!haveEncoding && embFontID.num >= 0) {
- obj1.initRef(embFontID.num, embFontID.gen);
- obj1.fetch(&obj2);
- if (obj2.isStream())
- getType1Encoding(obj2.getStream());
- obj2.free();
- obj1.free();
+ // use widths from built-in font
+ } else if (builtinFont) {
+ // this is a kludge for broken PDF files that encode char 32
+ // as .notdef
+ if (builtinFont->widths->getWidth("space", &w)) {
+ widths[32] = 0.001 * w;
+ }
+ for (code = 0; code < 256; ++code) {
+ if (enc[code] && builtinFont->widths->getWidth(enc[code], &w)) {
+ widths[code] = 0.001 * w;
}
+ }
- // TrueType font: use Mac encoding
- } else if (type == fontTrueType) {
- enc = &macRomanEncoding;
-
- // not Type 1 or TrueType: just use the standard encoding
+ // couldn't find widths -- use defaults
+ } else {
+ // this is technically an error -- the Widths entry is required
+ // for all but the Base-14 fonts -- but certain PDF generators
+ // apparently don't include widths for Arial and TimesNewRoman
+ if (isFixedWidth()) {
+ i = 0;
+ } else if (isSerif()) {
+ i = 8;
} else {
- enc = &standardEncoding;
+ i = 4;
+ }
+ if (isBold()) {
+ i += 2;
+ }
+ if (isItalic()) {
+ i += 1;
+ }
+ builtinFont = builtinFontSubst[i];
+ // this is a kludge for broken PDF files that encode char 32
+ // as .notdef
+ if (builtinFont->widths->getWidth("space", &w)) {
+ widths[32] = 0.001 * w;
+ }
+ for (code = 0; code < 256; ++code) {
+ if (enc[code] && builtinFont->widths->getWidth(enc[code], &w)) {
+ widths[code] = 0.001 * w;
+ }
}
}
+ obj1.free();
- return enc;
+ ok = gTrue;
}
-void GfxFont::getType1Encoding(Stream *str) {
- char buf[256];
- char *p;
- GBool found;
- int code, i;
+Gfx8BitFont::~Gfx8BitFont() {
+ int i;
- // look for encoding in font file
- str->reset();
- found = gFalse;
- for (i = 0; i < 100; ++i) {
- if (!str->getLine(buf, sizeof(buf)))
- break;
- if (!strncmp(buf, "/Encoding StandardEncoding def", 30))
- break;
- if (!strncmp(buf, "/Encoding 256 array", 19)) {
- found = gTrue;
- break;
- }
- }
-
- // found the encoding, grab it
- if (found) {
- for (i = 0; i < 300; ++i) {
- if (!str->getLine(buf, sizeof(buf)))
- break;
- p = strtok(buf, " \t");
- if (p && !strcmp(p, "dup")) {
- if ((p = strtok(NULL, " \t"))) {
- code = atoi(p);
- if ((p = strtok(NULL, " \t"))) {
- if (p[0] == '/')
- encoding->addChar(code, copyString(p+1));
- }
- }
- }
+ for (i = 0; i < 256; ++i) {
+ if (encFree[i] && enc[i]) {
+ gfree(enc[i]);
}
- //~ look for getinterval/putinterval junk
+ }
+ ctu->decRefCnt();
+ if (charProcs.isDict()) {
+ charProcs.free();
+ }
+ if (resources.isDict()) {
+ resources.free();
}
}
-void GfxFont::makeWidths(Dict *fontDict, GfxFontEncoding *builtinEncoding,
- Gushort *builtinWidths) {
- Object obj1, obj2;
- int firstChar, lastChar;
- int code, code2;
+int Gfx8BitFont::getNextChar(char *s, int len, CharCode *code,
+ Unicode *u, int uSize, int *uLen,
+ double *dx, double *dy, double *ox, double *oy) {
+ CharCode c;
+
+ *code = c = (CharCode)(*s & 0xff);
+ *uLen = ctu->mapToUnicode(c, u, uSize);
+ *dx = widths[c];
+ *dy = *ox = *oy = 0;
+ return 1;
+}
+
+CharCodeToUnicode *Gfx8BitFont::getToUnicode() {
+ ctu->incRefCnt();
+ return ctu;
+}
+
+Gushort *Gfx8BitFont::getCodeToGIDMap(FoFiTrueType *ff) {
+ Gushort *map;
+ int cmapPlatform, cmapEncoding;
+ int unicodeCmap, macRomanCmap, msSymbolCmap, cmap;
+ GBool useMacRoman, useUnicode;
char *charName;
- Gushort *defWidths;
- int index;
- double mult;
+ Unicode u;
+ int code, i, n;
+
+ map = (Gushort *)gmalloc(256 * sizeof(Gushort));
+ for (i = 0; i < 256; ++i) {
+ map[i] = 0;
+ }
- // initialize all widths to zero
- for (code = 0; code < 256; ++code)
- widths[code] = 0;
+ // To match up with the Adobe-defined behaviour, we choose a cmap
+ // like this:
+ // 1. If the PDF font has an encoding:
+ // 1a. If the PDF font specified MacRomanEncoding and the
+ // TrueType font has a Macintosh Roman cmap, use it, and
+ // reverse map the char names through MacRomanEncoding to
+ // get char codes.
+ // 1b. If the TrueType font has a Microsoft Unicode cmap or a
+ // non-Microsoft Unicode cmap, use it, and use the Unicode
+ // indexes, not the char codes.
+ // 1c. If the PDF font is symbolic and the TrueType font has a
+ // Microsoft Symbol cmap, use it, and use char codes
+ // directly (possibly with an offset of 0xf000).
+ // 1d. If the TrueType font has a Macintosh Roman cmap, use it,
+ // as in case 1a.
+ // 2. If the PDF font does not have an encoding:
+ // 2a. If the TrueType font has a Macintosh Roman cmap, use it,
+ // and use char codes directly (possibly with an offset of
+ // 0xf000).
+ // 2b. If the TrueType font has a Microsoft Symbol cmap, use it,
+ // and use char codes directly (possible with an offset of
+ // 0xf000).
+ // 3. If none of these rules apply, use the first cmap and hope for
+ // the best (this shouldn't happen).
+ unicodeCmap = macRomanCmap = msSymbolCmap = -1;
+ for (i = 0; i < ff->getNumCmaps(); ++i) {
+ cmapPlatform = ff->getCmapPlatform(i);
+ cmapEncoding = ff->getCmapEncoding(i);
+ if ((cmapPlatform == 3 && cmapEncoding == 1) ||
+ cmapPlatform == 0) {
+ unicodeCmap = i;
+ } else if (cmapPlatform == 1 && cmapEncoding == 0) {
+ macRomanCmap = i;
+ } else if (cmapPlatform == 3 && cmapEncoding == 0) {
+ msSymbolCmap = i;
+ }
+ }
+ cmap = 0;
+ useMacRoman = gFalse;
+ useUnicode = gFalse;
+ if (hasEncoding) {
+ if (usesMacRomanEnc && macRomanCmap >= 0) {
+ cmap = macRomanCmap;
+ useMacRoman = gTrue;
+ } else if (unicodeCmap >= 0) {
+ cmap = unicodeCmap;
+ useUnicode = gTrue;
+ } else if ((flags & fontSymbolic) && msSymbolCmap >= 0) {
+ cmap = msSymbolCmap;
+ } else if (macRomanCmap >= 0) {
+ cmap = macRomanCmap;
+ useMacRoman = gTrue;
+ }
+ } else {
+ if (macRomanCmap >= 0) {
+ cmap = macRomanCmap;
+ } else if (msSymbolCmap >= 0) {
+ cmap = msSymbolCmap;
+ }
+ }
- // use widths from built-in font
- if (builtinEncoding) {
- code2 = 0; // to make gcc happy
- for (code = 0; code < 256; ++code) {
- if ((charName = encoding->getCharName(code)) &&
- (code2 = builtinEncoding->getCharCode(charName)) >= 0)
- widths[code] = builtinWidths[code2] * 0.001;
+ // reverse map the char names through MacRomanEncoding, then map the
+ // char codes through the cmap
+ if (useMacRoman) {
+ for (i = 0; i < 256; ++i) {
+ if ((charName = enc[i])) {
+ if ((code = globalParams->getMacRomanCharCode(charName))) {
+ map[i] = ff->mapCodeToGID(cmap, code);
+ }
+ }
}
- // get widths from font dict
- } else {
- fontDict->lookup("FirstChar", &obj1);
- firstChar = obj1.isInt() ? obj1.getInt() : 0;
- obj1.free();
- fontDict->lookup("LastChar", &obj1);
- lastChar = obj1.isInt() ? obj1.getInt() : 255;
- obj1.free();
- if (type == fontType3)
- mult = fontMat[0];
- else
- mult = 0.001;
- fontDict->lookup("Widths", &obj1);
- if (obj1.isArray()) {
- for (code = firstChar; code <= lastChar; ++code) {
- obj1.arrayGet(code - firstChar, &obj2);
- if (obj2.isNum())
- widths[code] = obj2.getNum() * mult;
- obj2.free();
+ // map Unicode through the cmap
+ } else if (useUnicode) {
+ for (i = 0; i < 256; ++i) {
+ if ((n = ctu->mapToUnicode((CharCode)i, &u, 1))) {
+ map[i] = ff->mapCodeToGID(cmap, u);
}
- } else {
+ }
- // couldn't find widths -- use defaults
-#if 0
- //~ certain PDF generators apparently don't include widths
- //~ for Arial and TimesNewRoman -- and this error message
- //~ is a nuisance
- error(-1, "No character widths resource for non-builtin font");
-#endif
- if (isFixedWidth())
- index = 0;
- else if (isSerif())
- index = 8;
- else
- index = 4;
- if (isBold())
- index += 2;
- if (isItalic())
- index += 1;
- defWidths = defCharWidths[index];
- code2 = 0; // to make gcc happy
- for (code = 0; code < 256; ++code) {
- if ((charName = encoding->getCharName(code)) &&
- (code2 = standardEncoding.getCharCode(charName)) >= 0)
- widths[code] = defWidths[code2] * 0.001;
+ // map the char codes through the cmap, possibly with an offset of
+ // 0xf000
+ } else {
+ for (i = 0; i < 256; ++i) {
+ if (!(map[i] = ff->mapCodeToGID(cmap, i))) {
+ map[i] = ff->mapCodeToGID(cmap, 0xf000 + i);
}
}
- obj1.free();
}
+
+ // try the TrueType 'post' table to handle any unmapped characters
+ for (i = 0; i < 256; ++i) {
+ if (!map[i] && (charName = enc[i])) {
+ map[i] = (Gushort)(int)ff->mapNameToGID(charName);
+ }
+ }
+
+ return map;
}
-void GfxFont::getType0EncAndWidths(Dict *fontDict) {
- Object obj1, obj2, obj3, obj4, obj5, obj6;
- int excepsSize;
- int i, j, k, n;
+Dict *Gfx8BitFont::getCharProcs() {
+ return charProcs.isDict() ? charProcs.getDict() : (Dict *)NULL;
+}
- fontDict->lookup("DescendantFonts", &obj1);
- if (!obj1.isArray() || obj1.arrayGetLength() != 1) {
- error(-1, "Bad DescendantFonts entry for Type 0 font");
+Object *Gfx8BitFont::getCharProc(int code, Object *proc) {
+ if (enc[code] && charProcs.isDict()) {
+ charProcs.dictLookup(enc[code], proc);
+ } else {
+ proc->initNull();
+ }
+ return proc;
+}
+
+Dict *Gfx8BitFont::getResources() {
+ return resources.isDict() ? resources.getDict() : (Dict *)NULL;
+}
+
+//------------------------------------------------------------------------
+// GfxCIDFont
+//------------------------------------------------------------------------
+
+static int CDECL cmpWidthExcep(const void *w1, const void *w2) {
+ return ((GfxFontCIDWidthExcep *)w1)->first -
+ ((GfxFontCIDWidthExcep *)w2)->first;
+}
+
+static int CDECL cmpWidthExcepV(const void *w1, const void *w2) {
+ return ((GfxFontCIDWidthExcepV *)w1)->first -
+ ((GfxFontCIDWidthExcepV *)w2)->first;
+}
+
+GfxCIDFont::GfxCIDFont(XRef *xref, char *tagA, Ref idA, GString *nameA,
+ Dict *fontDict):
+ GfxFont(tagA, idA, nameA)
+{
+ Dict *desFontDict;
+ GString *collection, *cMapName;
+ Object desFontDictObj;
+ Object obj1, obj2, obj3, obj4, obj5, obj6;
+ int c1, c2;
+ int excepsSize, i, j, k;
+
+ ascent = 0.95;
+ descent = -0.35;
+ fontBBox[0] = fontBBox[1] = fontBBox[2] = fontBBox[3] = 0;
+ cMap = NULL;
+ ctu = NULL;
+ widths.defWidth = 1.0;
+ widths.defHeight = -1.0;
+ widths.defVY = 0.880;
+ widths.exceps = NULL;
+ widths.nExceps = 0;
+ widths.excepsV = NULL;
+ widths.nExcepsV = 0;
+ cidToGID = NULL;
+ cidToGIDLen = 0;
+
+ // get the descendant font
+ if (!fontDict->lookup("DescendantFonts", &obj1)->isArray()) {
+ error(-1, "Missing DescendantFonts entry in Type 0 font");
+ obj1.free();
goto err1;
}
- obj1.arrayGet(0, &obj2);
- if (!obj2.isDict("Font")) {
- error(-1, "Bad descendant font of Type 0 font");
- goto err2;
+ if (!obj1.arrayGet(0, &desFontDictObj)->isDict()) {
+ error(-1, "Bad descendant font in Type 0 font");
+ goto err3;
}
+ obj1.free();
+ desFontDict = desFontDictObj.getDict();
- obj2.dictLookup("CIDSystemInfo", &obj3);
- if (!obj3.isDict()) {
- error(-1, "Bad CIDSystemInfo in Type 0 font descendant");
+ // font type
+ if (!desFontDict->lookup("Subtype", &obj1)) {
+ error(-1, "Missing Subtype entry in Type 0 descendant font");
goto err3;
}
- obj3.dictLookup("Registry", &obj4);
- obj3.dictLookup("Ordering", &obj5);
- if (obj4.isString() && obj5.isString()) {
- if (obj4.getString()->cmp("Adobe") == 0 &&
- obj5.getString()->cmp("Japan1") == 0) {
-#if JAPANESE_SUPPORT
- is16 = gTrue;
- enc16.charSet = font16AdobeJapan12;
-#else
- error(-1, "Xpdf was compiled without Japanese font support");
- goto err4;
-#endif
- } else {
- error(-1, "Uknown Type 0 character set: %s-%s",
- obj4.getString()->getCString(), obj5.getString()->getCString());
- goto err4;
- }
+ if (obj1.isName("CIDFontType0")) {
+ type = fontCIDType0;
+ } else if (obj1.isName("CIDFontType2")) {
+ type = fontCIDType2;
} else {
- error(-1, "Unknown Type 0 character set");
+ error(-1, "Unknown Type 0 descendant font type '%s'",
+ obj1.isName() ? obj1.getName() : "???");
+ goto err3;
+ }
+ obj1.free();
+
+ // get info from font descriptor
+ readFontDescriptor(xref, desFontDict);
+
+ // look for an external font file
+ findExtFontFile();
+
+ //----- encoding info -----
+
+ // char collection
+ if (!desFontDict->lookup("CIDSystemInfo", &obj1)->isDict()) {
+ error(-1, "Missing CIDSystemInfo dictionary in Type 0 descendant font");
+ goto err3;
+ }
+ obj1.dictLookup("Registry", &obj2);
+ obj1.dictLookup("Ordering", &obj3);
+ if (!obj2.isString() || !obj3.isString()) {
+ error(-1, "Invalid CIDSystemInfo dictionary in Type 0 descendant font");
goto err4;
}
- obj5.free();
- obj4.free();
+ collection = obj2.getString()->copy()->append('-')->append(obj3.getString());
obj3.free();
+ obj2.free();
+ obj1.free();
- obj2.dictLookup("DW", &obj3);
- if (obj3.isInt())
- widths16.defWidth = obj3.getInt() * 0.001;
- else
- widths16.defWidth = 1.0;
- obj3.free();
+ // look for a ToUnicode CMap
+ if (!(ctu = readToUnicodeCMap(fontDict, 16, NULL))) {
+
+ // the "Adobe-Identity" and "Adobe-UCS" collections don't have
+ // cidToUnicode files
+ if (collection->cmp("Adobe-Identity") &&
+ collection->cmp("Adobe-UCS")) {
+
+ // look for a user-supplied .cidToUnicode file
+ if (!(ctu = globalParams->getCIDToUnicode(collection))) {
+ error(-1, "Unknown character collection '%s'",
+ collection->getCString());
+ delete collection;
+ goto err2;
+ }
+ }
+ }
+
+ // encoding (i.e., CMap)
+ //~ need to handle a CMap stream here
+ //~ also need to deal with the UseCMap entry in the stream dict
+ if (!fontDict->lookup("Encoding", &obj1)->isName()) {
+ error(-1, "Missing or invalid Encoding entry in Type 0 font");
+ delete collection;
+ goto err3;
+ }
+ cMapName = new GString(obj1.getName());
+ obj1.free();
+ if (!(cMap = globalParams->getCMap(collection, cMapName))) {
+ error(-1, "Unknown CMap '%s' for character collection '%s'",
+ cMapName->getCString(), collection->getCString());
+ delete collection;
+ delete cMapName;
+ goto err2;
+ }
+ delete collection;
+ delete cMapName;
+
+ // CIDToGIDMap (for embedded TrueType fonts)
+ if (type == fontCIDType2) {
+ desFontDict->lookup("CIDToGIDMap", &obj1);
+ if (obj1.isStream()) {
+ cidToGIDLen = 0;
+ i = 64;
+ cidToGID = (Gushort *)gmalloc(i * sizeof(Gushort));
+ obj1.streamReset();
+ while ((c1 = obj1.streamGetChar()) != EOF &&
+ (c2 = obj1.streamGetChar()) != EOF) {
+ if (cidToGIDLen == i) {
+ i *= 2;
+ cidToGID = (Gushort *)grealloc(cidToGID, i * sizeof(Gushort));
+ }
+ cidToGID[cidToGIDLen++] = (Gushort)((c1 << 8) + c2);
+ }
+ } else if (!obj1.isName("Identity") && !obj1.isNull()) {
+ error(-1, "Invalid CIDToGIDMap entry in CID font");
+ }
+ obj1.free();
+ }
- widths16.exceps = NULL;
- widths16.numExceps = 0;
- obj2.dictLookup("W", &obj3);
- if (obj3.isArray()) {
+ //----- character metrics -----
+
+ // default char width
+ if (desFontDict->lookup("DW", &obj1)->isInt()) {
+ widths.defWidth = obj1.getInt() * 0.001;
+ }
+ obj1.free();
+
+ // char width exceptions
+ if (desFontDict->lookup("W", &obj1)->isArray()) {
excepsSize = 0;
- k = 0;
i = 0;
- while (i+1 < obj3.arrayGetLength()) {
- obj3.arrayGet(i, &obj4);
- obj3.arrayGet(i+1, &obj5);
- if (obj4.isInt() && obj5.isInt()) {
- obj3.arrayGet(i+2, &obj6);
- if (!obj6.isNum()) {
+ while (i + 1 < obj1.arrayGetLength()) {
+ obj1.arrayGet(i, &obj2);
+ obj1.arrayGet(i + 1, &obj3);
+ if (obj2.isInt() && obj3.isInt() && i + 2 < obj1.arrayGetLength()) {
+ if (obj1.arrayGet(i + 2, &obj4)->isNum()) {
+ if (widths.nExceps == excepsSize) {
+ excepsSize += 16;
+ widths.exceps = (GfxFontCIDWidthExcep *)
+ grealloc(widths.exceps,
+ excepsSize * sizeof(GfxFontCIDWidthExcep));
+ }
+ widths.exceps[widths.nExceps].first = obj2.getInt();
+ widths.exceps[widths.nExceps].last = obj3.getInt();
+ widths.exceps[widths.nExceps].width = obj4.getNum() * 0.001;
+ ++widths.nExceps;
+ } else {
error(-1, "Bad widths array in Type 0 font");
- obj6.free();
- obj5.free();
- obj4.free();
- break;
}
- if (k == excepsSize) {
- excepsSize += 16;
- widths16.exceps = (GfxFontWidthExcep *)
- grealloc(widths16.exceps,
- excepsSize * sizeof(GfxFontWidthExcep));
- }
- widths16.exceps[k].first = obj4.getInt();
- widths16.exceps[k].last = obj5.getInt();
- widths16.exceps[k].width = obj6.getNum() * 0.001;
- obj6.free();
- ++k;
+ obj4.free();
i += 3;
- } else if (obj4.isInt() && obj5.isArray()) {
- if (k + obj5.arrayGetLength() >= excepsSize) {
- excepsSize = (k + obj5.arrayGetLength() + 15) & ~15;
- widths16.exceps = (GfxFontWidthExcep *)
- grealloc(widths16.exceps,
- excepsSize * sizeof(GfxFontWidthExcep));
+ } else if (obj2.isInt() && obj3.isArray()) {
+ if (widths.nExceps + obj3.arrayGetLength() > excepsSize) {
+ excepsSize = (widths.nExceps + obj3.arrayGetLength() + 15) & ~15;
+ widths.exceps = (GfxFontCIDWidthExcep *)
+ grealloc(widths.exceps,
+ excepsSize * sizeof(GfxFontCIDWidthExcep));
}
- n = obj4.getInt();
- for (j = 0; j < obj5.arrayGetLength(); ++j) {
- obj5.arrayGet(j, &obj6);
- if (!obj6.isNum()) {
+ j = obj2.getInt();
+ for (k = 0; k < obj3.arrayGetLength(); ++k) {
+ if (obj3.arrayGet(k, &obj4)->isNum()) {
+ widths.exceps[widths.nExceps].first = j;
+ widths.exceps[widths.nExceps].last = j;
+ widths.exceps[widths.nExceps].width = obj4.getNum() * 0.001;
+ ++j;
+ ++widths.nExceps;
+ } else {
error(-1, "Bad widths array in Type 0 font");
- obj6.free();
- break;
}
- widths16.exceps[k].first = widths16.exceps[k].last = n++;
- widths16.exceps[k].width = obj6.getNum() * 0.001;
- obj6.free();
- ++k;
+ obj4.free();
}
i += 2;
} else {
error(-1, "Bad widths array in Type 0 font");
- obj6.free();
- obj5.free();
- obj4.free();
- break;
+ ++i;
}
- obj5.free();
- obj4.free();
+ obj3.free();
+ obj2.free();
}
- widths16.numExceps = k;
- if (k > 0)
- qsort(widths16.exceps, k, sizeof(GfxFontWidthExcep), &cmpWidthExcep);
+ qsort(widths.exceps, widths.nExceps, sizeof(GfxFontCIDWidthExcep),
+ &cmpWidthExcep);
}
- obj3.free();
-
- obj2.free();
obj1.free();
- fontDict->lookup("Encoding", &obj1);
- if (!obj1.isName()) {
- error(-1, "Bad encoding for Type 0 font");
- goto err1;
- }
-#if JAPANESE_SUPPORT
- if (enc16.charSet == font16AdobeJapan12) {
- for (i = 0; gfxFontEnc16Tab[i].name; ++i) {
- if (!strcmp(obj1.getName(), gfxFontEnc16Tab[i].name))
- break;
+ // default metrics for vertical font
+ if (desFontDict->lookup("DW2", &obj1)->isArray() &&
+ obj1.arrayGetLength() == 2) {
+ if (obj1.arrayGet(0, &obj2)->isNum()) {
+ widths.defVY = obj2.getNum() * 0.001;
}
- if (!gfxFontEnc16Tab[i].name) {
- error(-1, "Unknown encoding '%s' for Adobe-Japan1-2 font",
- obj1.getName());
- goto err1;
+ obj2.free();
+ if (obj1.arrayGet(1, &obj2)->isNum()) {
+ widths.defHeight = obj2.getNum() * 0.001;
}
- enc16.enc = gfxFontEnc16Tab[i].enc;
+ obj2.free();
+ }
+ obj1.free();
+
+ // char metric exceptions for vertical font
+ if (desFontDict->lookup("W2", &obj1)->isArray()) {
+ excepsSize = 0;
+ i = 0;
+ while (i + 1 < obj1.arrayGetLength()) {
+ obj1.arrayGet(i, &obj2);
+ obj1.arrayGet(i+ 1, &obj3);
+ if (obj2.isInt() && obj3.isInt() && i + 4 < obj1.arrayGetLength()) {
+ if (obj1.arrayGet(i + 2, &obj4)->isNum() &&
+ obj1.arrayGet(i + 3, &obj5)->isNum() &&
+ obj1.arrayGet(i + 4, &obj6)->isNum()) {
+ if (widths.nExcepsV == excepsSize) {
+ excepsSize += 16;
+ widths.excepsV = (GfxFontCIDWidthExcepV *)
+ grealloc(widths.excepsV,
+ excepsSize * sizeof(GfxFontCIDWidthExcepV));
+ }
+ widths.excepsV[widths.nExcepsV].first = obj2.getInt();
+ widths.excepsV[widths.nExcepsV].last = obj3.getInt();
+ widths.excepsV[widths.nExcepsV].height = obj4.getNum() * 0.001;
+ widths.excepsV[widths.nExcepsV].vx = obj5.getNum() * 0.001;
+ widths.excepsV[widths.nExcepsV].vy = obj6.getNum() * 0.001;
+ ++widths.nExcepsV;
+ } else {
+ error(-1, "Bad widths (W2) array in Type 0 font");
+ }
+ obj6.free();
+ obj5.free();
+ obj4.free();
+ i += 5;
+ } else if (obj2.isInt() && obj3.isArray()) {
+ if (widths.nExcepsV + obj3.arrayGetLength() / 3 > excepsSize) {
+ excepsSize =
+ (widths.nExcepsV + obj3.arrayGetLength() / 3 + 15) & ~15;
+ widths.excepsV = (GfxFontCIDWidthExcepV *)
+ grealloc(widths.excepsV,
+ excepsSize * sizeof(GfxFontCIDWidthExcepV));
+ }
+ j = obj2.getInt();
+ for (k = 0; k < obj3.arrayGetLength(); k += 3) {
+ if (obj3.arrayGet(k, &obj4)->isNum() &&
+ obj3.arrayGet(k+1, &obj5)->isNum() &&
+ obj3.arrayGet(k+2, &obj6)->isNum()) {
+ widths.excepsV[widths.nExceps].first = j;
+ widths.excepsV[widths.nExceps].last = j;
+ widths.excepsV[widths.nExceps].height = obj4.getNum() * 0.001;
+ widths.excepsV[widths.nExceps].vx = obj5.getNum() * 0.001;
+ widths.excepsV[widths.nExceps].vy = obj6.getNum() * 0.001;
+ ++j;
+ ++widths.nExcepsV;
+ } else {
+ error(-1, "Bad widths (W2) array in Type 0 font");
+ }
+ obj6.free();
+ obj5.free();
+ obj4.free();
+ }
+ i += 2;
+ } else {
+ error(-1, "Bad widths (W2) array in Type 0 font");
+ ++i;
+ }
+ obj3.free();
+ obj2.free();
+ }
+ qsort(widths.excepsV, widths.nExcepsV, sizeof(GfxFontCIDWidthExcepV),
+ &cmpWidthExcepV);
}
-#endif
obj1.free();
+ desFontDictObj.free();
+ ok = gTrue;
return;
err4:
- obj5.free();
- obj4.free();
- err3:
obj3.free();
- err2:
obj2.free();
- err1:
+ err3:
obj1.free();
- makeEncoding(fontDict, NULL);
- makeWidths(fontDict, NULL, NULL);
+ err2:
+ desFontDictObj.free();
+ err1:;
}
-static int CDECL cmpWidthExcep(const void *w1, const void *w2) {
- return ((GfxFontWidthExcep *)w1)->first - ((GfxFontWidthExcep *)w2)->first;
+GfxCIDFont::~GfxCIDFont() {
+ if (cMap) {
+ cMap->decRefCnt();
+ }
+ if (ctu) {
+ ctu->decRefCnt();
+ }
+ gfree(widths.exceps);
+ gfree(widths.excepsV);
+ if (cidToGID) {
+ gfree(cidToGID);
+ }
+}
+
+int GfxCIDFont::getNextChar(char *s, int len, CharCode *code,
+ Unicode *u, int uSize, int *uLen,
+ double *dx, double *dy, double *ox, double *oy) {
+ CID cid;
+ double w, h, vx, vy;
+ int n, a, b, m;
+
+ if (!cMap) {
+ *code = 0;
+ *uLen = 0;
+ *dx = *dy = 0;
+ return 1;
+ }
+
+ *code = (CharCode)(cid = cMap->getCID(s, len, &n));
+ if (ctu) {
+ *uLen = ctu->mapToUnicode(cid, u, uSize);
+ } else {
+ *uLen = 0;
+ }
+
+ // horizontal
+ if (cMap->getWMode() == 0) {
+ w = widths.defWidth;
+ h = vx = vy = 0;
+ if (widths.nExceps > 0 && cid >= widths.exceps[0].first) {
+ a = 0;
+ b = widths.nExceps;
+ // invariant: widths.exceps[a].first <= cid < widths.exceps[b].first
+ while (b - a > 1) {
+ m = (a + b) / 2;
+ if (widths.exceps[m].first <= cid) {
+ a = m;
+ } else {
+ b = m;
+ }
+ }
+ if (cid <= widths.exceps[a].last) {
+ w = widths.exceps[a].width;
+ }
+ }
+
+ // vertical
+ } else {
+ w = 0;
+ h = widths.defHeight;
+ vx = widths.defWidth / 2;
+ vy = widths.defVY;
+ if (widths.nExcepsV > 0 && cid >= widths.excepsV[0].first) {
+ a = 0;
+ b = widths.nExcepsV;
+ // invariant: widths.excepsV[a].first <= cid < widths.excepsV[b].first
+ while (b - a > 1) {
+ m = (a + b) / 2;
+ if (widths.excepsV[m].last <= cid) {
+ a = m;
+ } else {
+ b = m;
+ }
+ }
+ if (cid <= widths.excepsV[a].last) {
+ h = widths.excepsV[a].height;
+ vx = widths.excepsV[a].vx;
+ vy = widths.excepsV[a].vy;
+ }
+ }
+ }
+
+ *dx = w;
+ *dy = h;
+ *ox = vx;
+ *oy = vy;
+
+ return n;
+}
+
+int GfxCIDFont::getWMode() {
+ return cMap ? cMap->getWMode() : 0;
+}
+
+CharCodeToUnicode *GfxCIDFont::getToUnicode() {
+ if (ctu) {
+ ctu->incRefCnt();
+ }
+ return ctu;
+}
+
+GString *GfxCIDFont::getCollection() {
+ return cMap ? cMap->getCollection() : (GString *)NULL;
}
//------------------------------------------------------------------------
// GfxFontDict
//------------------------------------------------------------------------
-GfxFontDict::GfxFontDict(Dict *fontDict) {
+GfxFontDict::GfxFontDict(XRef *xref, Ref *fontDictRef, Dict *fontDict) {
int i;
Object obj1, obj2;
+ Ref r;
numFonts = fontDict->getLength();
fonts = (GfxFont **)gmalloc(numFonts * sizeof(GfxFont *));
for (i = 0; i < numFonts; ++i) {
fontDict->getValNF(i, &obj1);
- obj1.fetch(&obj2);
- if (obj1.isRef() && obj2.isDict("Font")) {
- fonts[i] = new GfxFont(fontDict->getKey(i), obj1.getRef(),
- obj2.getDict());
+ obj1.fetch(xref, &obj2);
+ if (obj2.isDict()) {
+ if (obj1.isRef()) {
+ r = obj1.getRef();
+ } else {
+ // no indirect reference for this font, so invent a unique one
+ // (legal generation numbers are five digits, so any 6-digit
+ // number would be safe)
+ r.num = i;
+ if (fontDictRef) {
+ r.gen = 100000 + fontDictRef->num;
+ } else {
+ r.gen = 999999;
+ }
+ }
+ fonts[i] = GfxFont::makeFont(xref, fontDict->getKey(i),
+ r, obj2.getDict());
+ if (fonts[i] && !fonts[i]->isOk()) {
+ delete fonts[i];
+ fonts[i] = NULL;
+ }
} else {
error(-1, "font resource is not a dictionary");
fonts[i] = NULL;
GfxFontDict::~GfxFontDict() {
int i;
- for (i = 0; i < numFonts; ++i)
- delete fonts[i];
+ for (i = 0; i < numFonts; ++i) {
+ if (fonts[i]) {
+ delete fonts[i];
+ }
+ }
gfree(fonts);
}
int i;
for (i = 0; i < numFonts; ++i) {
- if (fonts[i]->matches(tag))
+ if (fonts[i] && fonts[i]->matches(tag)) {
return fonts[i];
+ }
}
return NULL;
}