basefont.java

处理PDF

            catch (Exception e) {
                throw new DocumentException(e);
            }
        }
    }
    
    /**
     *Creates new BaseFont
     */
    protected BaseFont() {
    }
    
    /**
     * Creates a new font. This will always be the default Helvetica font (not embedded).
     * This method is introduced because Helvetica is used in many examples.
     * @return	a BaseFont object (Helvetica, Winansi, not embedded)
     * @throws	IOException			This shouldn't occur ever
     * @throws	DocumentException	This shouldn't occur ever
     * @since	2.1.1 
     */
    public static BaseFont createFont() throws DocumentException, IOException {
    	return createFont(BaseFont.HELVETICA, BaseFont.WINANSI, BaseFont.NOT_EMBEDDED);
    }
    
    /**
     * Creates a new font. This font can be one of the 14 built in types,
     * a Type1 font referred to by an AFM or PFM file, a TrueType font (simple or collection) or a CJK font from the
     * Adobe Asian Font Pack. TrueType fonts and CJK fonts can have an optional style modifier
     * appended to the name. These modifiers are: Bold, Italic and BoldItalic. An
     * example would be "STSong-Light,Bold". Note that this modifiers do not work if
     * the font is embedded. Fonts in TrueType collections are addressed by index such as "msgothic.ttc,1".
     * This would get the second font (indexes start at 0), in this case "MS PGothic".
     * <P>
     * The fonts are cached and if they already exist they are extracted from the cache,
     * not parsed again.
     * <P>
     * Besides the common encodings described by name, custom encodings 
     * can also be made. These encodings will only work for the single byte fonts
     * Type1 and TrueType. The encoding string starts with a '#'
     * followed by "simple" or "full". If "simple" there is a decimal for the first character position and then a list
     * of hex values representing the Unicode codes that compose that encoding.<br>
     * The "simple" encoding is recommended for TrueType fonts
     * as the "full" encoding risks not matching the character with the right glyph
     * if not done with care.<br>
     * The "full" encoding is specially aimed at Type1 fonts where the glyphs have to be
     * described by non standard names like the Tex math fonts. Each group of three elements
     * compose a code position: the one byte code order in decimal or as 'x' (x cannot be the space), the name and the Unicode character
     * used to access the glyph. The space must be assigned to character position 32 otherwise
     * text justification will not work.
     * <P>
     * Example for a "simple" encoding that includes the Unicode
     * character space, A, B and ecyrillic:
     * <PRE>
     * "# simple 32 0020 0041 0042 0454"
     * </PRE>
     * <P>
     * Example for a "full" encoding for a Type1 Tex font:
     * <PRE>
     * "# full 'A' nottriangeqlleft 0041 'B' dividemultiply 0042 32 space 0020"
     * </PRE>
     * <P>
     * This method calls:<br>
     * <PRE>
     * createFont(name, encoding, embedded, true, null, null);
     * </PRE>
     * @param name the name of the font or it's location on file
     * @param encoding the encoding to be applied to this font
     * @param embedded true if the font is to be embedded in the PDF
     * @return returns a new font. This font may come from the cache
     * @throws DocumentException the font is invalid
     * @throws IOException the font file could not be read
     */
    public static BaseFont createFont(String name, String encoding, boolean embedded) throws DocumentException, IOException {
        return createFont(name, encoding, embedded, true, null, null);
    }
    
    /** Creates a new font. This font can be one of the 14 built in types,
     * a Type1 font referred to by an AFM or PFM file, a TrueType font (simple or collection) or a CJK font from the
     * Adobe Asian Font Pack. TrueType fonts and CJK fonts can have an optional style modifier
     * appended to the name. These modifiers are: Bold, Italic and BoldItalic. An
     * example would be "STSong-Light,Bold". Note that this modifiers do not work if
     * the font is embedded. Fonts in TrueType collections are addressed by index such as "msgothic.ttc,1".
     * This would get the second font (indexes start at 0), in this case "MS PGothic".
     * <P>
     * The fonts may or may not be cached depending on the flag <CODE>cached</CODE>.
     * If the <CODE>byte</CODE> arrays are present the font will be
     * read from them instead of the name. A name is still required to identify
     * the font type.
     * <P>
     * Besides the common encodings described by name, custom encodings 
     * can also be made. These encodings will only work for the single byte fonts
     * Type1 and TrueType. The encoding string starts with a '#'
     * followed by "simple" or "full". If "simple" there is a decimal for the first character position and then a list
     * of hex values representing the Unicode codes that compose that encoding.<br>
     * The "simple" encoding is recommended for TrueType fonts
     * as the "full" encoding risks not matching the character with the right glyph
     * if not done with care.<br>
     * The "full" encoding is specially aimed at Type1 fonts where the glyphs have to be
     * described by non standard names like the Tex math fonts. Each group of three elements
     * compose a code position: the one byte code order in decimal or as 'x' (x cannot be the space), the name and the Unicode character
     * used to access the glyph. The space must be assigned to character position 32 otherwise
     * text justification will not work.
     * <P>
     * Example for a "simple" encoding that includes the Unicode
     * character space, A, B and ecyrillic:
     * <PRE>
     * "# simple 32 0020 0041 0042 0454"
     * </PRE>
     * <P>
     * Example for a "full" encoding for a Type1 Tex font:
     * <PRE>
     * "# full 'A' nottriangeqlleft 0041 'B' dividemultiply 0042 32 space 0020"
     * </PRE>
     * @param name the name of the font or it's location on file
     * @param encoding the encoding to be applied to this font
     * @param embedded true if the font is to be embedded in the PDF
     * @param cached true if the font comes from the cache or is added to
     * the cache if new, false if the font is always created new
     * @param ttfAfm the true type font or the afm in a byte array
     * @param pfb the pfb in a byte array
     * @return returns a new font. This font may come from the cache but only if cached
     * is true, otherwise it will always be created new
     * @throws DocumentException the font is invalid
     * @throws IOException the font file could not be read
     */
    public static BaseFont createFont(String name, String encoding, boolean embedded, boolean cached, byte ttfAfm[], byte pfb[]) throws DocumentException, IOException {
        return createFont(name, encoding, embedded, cached, ttfAfm, pfb, false);
    }
    
    /** Creates a new font. This font can be one of the 14 built in types,
     * a Type1 font referred to by an AFM or PFM file, a TrueType font (simple or collection) or a CJK font from the
     * Adobe Asian Font Pack. TrueType fonts and CJK fonts can have an optional style modifier
     * appended to the name. These modifiers are: Bold, Italic and BoldItalic. An
     * example would be "STSong-Light,Bold". Note that this modifiers do not work if
     * the font is embedded. Fonts in TrueType collections are addressed by index such as "msgothic.ttc,1".
     * This would get the second font (indexes start at 0), in this case "MS PGothic".
     * <P>
     * The fonts may or may not be cached depending on the flag <CODE>cached</CODE>.
     * If the <CODE>byte</CODE> arrays are present the font will be
     * read from them instead of the name. A name is still required to identify
     * the font type.
     * <P>
     * Besides the common encodings described by name, custom encodings 
     * can also be made. These encodings will only work for the single byte fonts
     * Type1 and TrueType. The encoding string starts with a '#'
     * followed by "simple" or "full". If "simple" there is a decimal for the first character position and then a list
     * of hex values representing the Unicode codes that compose that encoding.<br>
     * The "simple" encoding is recommended for TrueType fonts
     * as the "full" encoding risks not matching the character with the right glyph
     * if not done with care.<br>
     * The "full" encoding is specially aimed at Type1 fonts where the glyphs have to be
     * described by non standard names like the Tex math fonts. Each group of three elements
     * compose a code position: the one byte code order in decimal or as 'x' (x cannot be the space), the name and the Unicode character
     * used to access the glyph. The space must be assigned to character position 32 otherwise
     * text justification will not work.
     * <P>
     * Example for a "simple" encoding that includes the Unicode
     * character space, A, B and ecyrillic:
     * <PRE>
     * "# simple 32 0020 0041 0042 0454"
     * </PRE>
     * <P>
     * Example for a "full" encoding for a Type1 Tex font:
     * <PRE>
     * "# full 'A' nottriangeqlleft 0041 'B' dividemultiply 0042 32 space 0020"
     * </PRE>
     * @param name the name of the font or it's location on file
     * @param encoding the encoding to be applied to this font
     * @param embedded true if the font is to be embedded in the PDF
     * @param cached true if the font comes from the cache or is added to
     * the cache if new, false if the font is always created new
     * @param ttfAfm the true type font or the afm in a byte array
     * @param pfb the pfb in a byte array
     * @param noThrow if true will not throw an exception if the font is not recognized and will return null, if false will throw
     * an exception if the font is not recognized. Note that even if true an exception may be thrown in some circumstances.
     * This parameter is useful for FontFactory that may have to check many invalid font names before finding the right one
     * @return returns a new font. This font may come from the cache but only if cached
     * is true, otherwise it will always be created new
     * @throws DocumentException the font is invalid
     * @throws IOException the font file could not be read
     */
    public static BaseFont createFont(String name, String encoding, boolean embedded, boolean cached, byte ttfAfm[], byte pfb[], boolean noThrow) throws DocumentException, IOException {
        String nameBase = getBaseName(name);
        encoding = normalizeEncoding(encoding);
        boolean isBuiltinFonts14 = BuiltinFonts14.containsKey(name);
        boolean isCJKFont = isBuiltinFonts14 ? false : CJKFont.isCJKFont(nameBase, encoding);
        if (isBuiltinFonts14 || isCJKFont)
            embedded = false;
        else if (encoding.equals(IDENTITY_H) || encoding.equals(IDENTITY_V))
            embedded = true;
        BaseFont fontFound = null;
        BaseFont fontBuilt = null;
        String key = name + "\n" + encoding + "\n" + embedded;
        if (cached) {
            synchronized (fontCache) {
                fontFound = (BaseFont)fontCache.get(key);
            }
            if (fontFound != null)
                return fontFound;
        }
        if (isBuiltinFonts14 || name.toLowerCase().endsWith(".afm") || name.toLowerCase().endsWith(".pfm")) {
            fontBuilt = new Type1Font(name, encoding, embedded, ttfAfm, pfb);
            fontBuilt.fastWinansi = encoding.equals(CP1252);
        }
        else if (nameBase.toLowerCase().endsWith(".ttf") || nameBase.toLowerCase().endsWith(".otf") || nameBase.toLowerCase().indexOf(".ttc,") > 0) {
            if (encoding.equals(IDENTITY_H) || encoding.equals(IDENTITY_V))
                fontBuilt = new TrueTypeFontUnicode(name, encoding, embedded, ttfAfm);
            else {
                fontBuilt = new TrueTypeFont(name, encoding, embedded, ttfAfm);
                fontBuilt.fastWinansi = encoding.equals(CP1252);
            }
        }
        else if (isCJKFont)
            fontBuilt = new CJKFont(name, encoding, embedded);
        else if (noThrow)
            return null;
        else
            throw new DocumentException("Font '" + name + "' with '" + encoding + "' is not recognized.");
        if (cached) {
            synchronized (fontCache) {
                fontFound = (BaseFont)fontCache.get(key);
                if (fontFound != null)
                    return fontFound;
                fontCache.put(key, fontBuilt);
            }
        }
        return fontBuilt;
    }
    
    /**
     * Creates a font based on an existing document font. The created font font may not
     * behave as expected, depending on the encoding or subset.
     * @param fontRef the reference to the document font
     * @return the font
     */    
    public static BaseFont createFont(PRIndirectReference fontRef) {
        return new DocumentFont(fontRef);
    }
    
    /**
     * Gets the name without the modifiers Bold, Italic or BoldItalic.
     * @param name the full name of the font
     * @return the name without the modifiers Bold, Italic or BoldItalic
     */
    protected static String getBaseName(String name) {
        if (name.endsWith(",Bold"))
            return name.substring(0, name.length() - 5);
        else if (name.endsWith(",Italic"))
            return name.substring(0, name.length() - 7);
        else if (name.endsWith(",BoldItalic"))
            return name.substring(0, name.length() - 11);
        else
            return name;
    }
    
    /**
     * Normalize the encoding names. "winansi" is changed to "Cp1252" and
     * "macroman" is changed to "MacRoman".
     * @param enc the encoding to be normalized
     * @return the normalized encoding
     */
    protected static String normalizeEncoding(String enc) {
        if (enc.equals("winansi") || enc.equals(""))
            return CP1252;
        else if (enc.equals("macroman"))
            return MACROMAN;
        else
            return enc;
    }
    
    /**
     * Creates the <CODE>widths</CODE> and the <CODE>differences</CODE> arrays
     */
    protected void createEncoding() {
        if (encoding.startsWith("#")) {
            specialMap = new IntHashtable();
            StringTokenizer tok = new StringTokenizer(encoding.substring(1), " ,\t\n\r\f");
            if (tok.nextToken().equals("full")) {
                while (tok.hasMoreTokens()) {
                    String order = tok.nextToken();
                    String name = tok.nextToken();
                    char uni = (char)Integer.parseInt(tok.nextToken(), 16);
                    int orderK;
                    if (order.startsWith("'"))
                        orderK = order.charAt(1);
                    else
                        orderK = Integer.parseInt(order);
                    orderK %= 256;
                    specialMap.put(uni, orderK);
                    differences[orderK] = name;
                    unicodeDifferences[orderK] = uni;
                    widths[orderK] = getRawWidth(uni, name);
                    charBBoxes[orderK] = getRawCharBBox(uni, name);
                }
            }
            else {
                int k = 0;
                if (tok.hasMoreTokens())
                    k = Integer.parseInt(tok.nextToken());
                while (tok.hasMoreTokens() && k < 256) {
                    String hex = tok.nextToken();
                    int uni = Integer.parseInt(hex, 16) % 0x10000;
                    String name = GlyphList.unicodeToName(uni);
                    if (name != null) {
                        specialMap.put(uni, k);
                        differences[k] = name;
                        unicodeDifferences[k] = (char)uni;
                        widths[k] = getRawWidth(uni, name);
                        charBBoxes[k] = getRawCharBBox(uni, name);
                        ++k;
                    }
                }
            }
            for (int k = 0; k < 256; ++k) {
                if (differences[k] == null) {
                    differences[k] = notdef;
                }
            }
        }
        else if (fontSpecific) {
            for (int k = 0; k < 256; ++k) {
                widths[k] = getRawWidth(k, null);
                charBBoxes[k] = getRawCharBBox(k, null);
            }
        }
        else {
            String s;
            String name;
            char c;
            byte b[] = new byte[1];
            for (int k = 0; k < 256; ++k) {
                b[0] = (byte)k;
                s = PdfEncodings.convertToString(b, encoding);
                if (s.length() > 0) {
                    c = s.charAt(0);
                }
                else {
                    c = '?';
                }
                name = GlyphList.unicodeToName(c);
                if (name == null)
                    name = notdef;
                differences[k] = name;
                unicodeDifferences[k] = c;
                widths[k] = getRawWidth(c, name);
                charBBoxes[k] = getRawCharBBox(c, name);
            }
        }
    }
    
    /**
     * Gets the width from the font according to the Unicode char <CODE>c</CODE>
     * or the <CODE>name</CODE>. If the <CODE>name</CODE> is null it's a symbolic font.
     * @param c the unicode char
     * @param name the glyph name
     * @return the width of the char
     */
    abstract int getRawWidth(int c, String name);
    
    /**
     * Gets the kerning between two Unicode chars.
     * @param char1 the first char
     * @param char2 the second char