diff --git a/app/src/main/java/io/legado/app/help/JsExtensions.kt b/app/src/main/java/io/legado/app/help/JsExtensions.kt index f59c9a81d..ef8a51b42 100644 --- a/app/src/main/java/io/legado/app/help/JsExtensions.kt +++ b/app/src/main/java/io/legado/app/help/JsExtensions.kt @@ -775,7 +775,7 @@ interface JsExtensions : JsEncodeUtils { } font ?: return null qTTF = QueryTTF(font) - CacheManager.put(key, qTTF) + CacheManager.put(key, qTTF) // debug注释掉 return qTTF } catch (e: Exception) { AppLog.put("获取字体处理类出错", e) @@ -798,15 +798,21 @@ interface JsExtensions : JsEncodeUtils { val contentArray = text.toStringArray() //这里不能用toCharArray,因为有些文字占多个字节 contentArray.forEachIndexed { index, s -> val oldCode = s.codePointAt(0) - val glyf = errorQueryTTF.getGlyfByCode(oldCode) - val code = correctQueryTTF.getCodeByGlyf(glyf) + // 忽略正常的空白字符 + if (errorQueryTTF.isBlankUnicode(oldCode)) { + return@forEachIndexed + } + val glyf = errorQueryTTF.getGlyfByUnicode(oldCode) + // 删除轮廓数据不存在的字符 + if (filter && (glyf == null)) { + contentArray[index] = "" + return@forEachIndexed + } + // 使用轮廓数据反查Unicode + val code = correctQueryTTF.getUnicodeByGlyf(glyf) if (code != 0) { contentArray[index] = code.toChar().toString() } - if (glyf.equals("") && filter) { - // 删除轮廓数据为空的字符 - contentArray[index] = "" - } } return contentArray.joinToString("") } diff --git a/app/src/main/java/io/legado/app/model/analyzeRule/QueryTTF.java b/app/src/main/java/io/legado/app/model/analyzeRule/QueryTTF.java index a5ec8af6f..6a7ee39fd 100644 --- a/app/src/main/java/io/legado/app/model/analyzeRule/QueryTTF.java +++ b/app/src/main/java/io/legado/app/model/analyzeRule/QueryTTF.java @@ -1,44 +1,68 @@ package io.legado.app.model.analyzeRule; -import android.util.Log; - import java.nio.ByteBuffer; import java.nio.ByteOrder; -import java.util.Arrays; -import java.util.HashMap; -import java.util.Map; import java.nio.charset.StandardCharsets; +import java.util.HashMap; import java.util.LinkedList; -import java.util.List; -import java.util.Objects; -import java.util.concurrent.ConcurrentHashMap; -import java.util.concurrent.ExecutorService; -import java.util.concurrent.Executors; -import java.util.concurrent.TimeUnit; @SuppressWarnings({"FieldCanBeLocal", "unused"}) public class QueryTTF { + /** + * 文件头 + * + * @url Microsoft opentype 字体文档 + */ private static class Header { - public int majorVersion; - public int minorVersion; - public int numOfTables; + /** + * uint32 字体版本 0x00010000 (ttf) + */ + public long sfntVersion; + /** + * uint16 Number of tables. + */ + public int numTables; + /** + * uint16 + */ public int searchRange; + /** + * uint16 + */ public int entrySelector; + /** + * uint16 + */ public int rangeShift; } + /** + * 数据表目录 + */ private static class Directory { - public String tag; // table name - public int checkSum; // Check sum - public int offset; // Offset from beginning of file - public int length; // length of the table in bytes + /** + * uint32 (表标识符) + */ + public String tableTag; + /** + * uint32 (该表的校验和) + */ + public int checkSum; + /** + * uint32 (TTF文件 Bytes 数据索引 0 开始的偏移地址) + */ + public int offset; + /** + * uint32 (该表的长度) + */ + public int length; } private static class NameLayout { public int format; public int count; public int stringOffset; - public List records = new LinkedList<>(); + public LinkedList records = new LinkedList<>(); } private static class NameRecord { @@ -50,105 +74,433 @@ public class QueryTTF { public int offset; // 名称字符串相对于stringOffset的字节偏移量 } + /** + * Font Header Table + */ private static class HeadLayout { + /** + * uint16 + */ public int majorVersion; + /** + * uint16 + */ public int minorVersion; + /** + * uint16 + */ public int fontRevision; + /** + * uint32 + */ public int checkSumAdjustment; + /** + * uint32 + */ public int magicNumber; + /** + * uint16 + */ public int flags; + /** + * uint16 + */ public int unitsPerEm; + /** + * long + */ public long created; + /** + * long + */ public long modified; + /** + * int16 + */ public short xMin; + /** + * int16 + */ public short yMin; + /** + * int16 + */ public short xMax; + /** + * int16 + */ public short yMax; + /** + * uint16 + */ public int macStyle; + /** + * uint16 + */ public int lowestRecPPEM; + /** + * int16 + */ public short fontDirectionHint; - public short indexToLocFormat; // <0:loca是2字节数组, 1:loca是4字节数组> + /** + * int16 + *

0 表示短偏移 (Offset16)，1 表示长偏移 (Offset32)。 + */ + public short indexToLocFormat; + /** + * int16 + */ public short glyphDataFormat; } + /** + * Maximum Profile + */ private static class MaxpLayout { - public int majorVersion; - public int minorVersion; - public int numGlyphs; // 字体中的字形数量 + /** + * uint32 高16位表示整数，低16位表示小数 + */ + public int version; + /** + * uint16 字体中的字形数量 + */ + public int numGlyphs; + /** + * uint16 非复合字形中包含的最大点数。点是构成字形轮廓的基本单位。 + */ public int maxPoints; + /** + * uint16 非复合字形中包含的最大轮廓数。轮廓是由一系列点连接形成的封闭曲线。 + */ public int maxContours; + /** + * uint16 复合字形中包含的最大点数。复合字形是由多个简单字形组合而成的。 + */ public int maxCompositePoints; + /** + * uint16 复合字形中包含的最大轮廓数。 + */ public int maxCompositeContours; + /** + * uint16 + */ public int maxZones; + /** + * uint16 + */ public int maxTwilightPoints; + /** + * uint16 + */ public int maxStorage; + /** + * uint16 + */ public int maxFunctionDefs; + /** + * uint16 + */ public int maxInstructionDefs; + /** + * uint16 + */ public int maxStackElements; + /** + * uint16 + */ public int maxSizeOfInstructions; + /** + * uint16 任何复合字形在“顶层”引用的最大组件数。 + */ public int maxComponentElements; + /** + * uint16 递归的最大层数；简单组件为1。 + */ public int maxComponentDepth; } + /** + * 字符到字形索引映射表 + */ private static class CmapLayout { + /** + * uint16 + */ public int version; + /** + * uint16 后面的编码表的数量 + */ public int numTables; - public List records = new LinkedList<>(); - public Map tables = new HashMap<>(); + public LinkedList records = new LinkedList<>(); + public HashMap tables = new HashMap<>(); } + /** + * Encoding records and encodings + */ private static class CmapRecord { - public int platformID; // UInt16 - public int platformSpecificID; // UInt16 - public int offset; // UInt32 + /** + * uint16 Platform ID. + *

0、Unicode + *

1、Macintosh + *

2、ISO + *

3、Windows + *

4、Custom + */ + public int platformID; + /** + * uint16 Platform-specific encoding ID. + *

platform ID = 3 + *

0、Symbol + *

1、Unicode BMP + *

2、ShiftJIS + *

3、PRC + *

4、Big5 + *

5、Wansung + *

6、Johab + *

7、Reserved + *

8、Reserved + *

9、Reserved + *

10、Unicode full repertoire + */ + public int encodingID; + /** + * uint32 从 cmap 表开头到子表的字节偏移量 + */ + public int offset; } private static class CmapFormat { + /** + * uint16 + *

cmapFormat 子表的格式类型 + */ public int format; + /** + * uint16 + *

这个 Format 表的长度（以字节为单位） + */ public int length; + /** + * uint16 + *

仅 platformID=1 时有效 + */ public int language; - public byte[] glyphIdArray; - } - - private static class CmapFormat4 extends CmapFormat { + /** + * uint16[256] + *

仅 Format=2 + *

将高字节映射到 subHeaders 的数组：值为 subHeader 索引x8 + */ + public int[] subHeaderKeys; + /** + * uint16[] + *

仅 Format=2 + *

subHeader 子标头的可变长度数组 + *

其结构为 uint16[][4]{ {uint16,uint16,int16,uint16}, ... } + */ + public int[] subHeaders; + /** + * uint16 segCount x2 + *

仅 Format=4 + *

seg段计数乘以 2。这是因为每个段用两个字节表示，所以这个值是实际段数的两倍。 + */ public int segCountX2; + /** + * uint16 + *

仅 Format=4 + *

小于或等于段数的最大二次幂，再乘以 2。这是为二分查找优化搜索过程。 + */ public int searchRange; + /** + * uint16 + *

仅 Format=4 + *

等于 log2(searchRange/2)，这是最大二次幂的对数。 + */ public int entrySelector; + /** + * uint16 + *

仅 Format=4 + *

segCount * 2 - searchRange 用于调整搜索范围的偏移。 + */ public int rangeShift; + /** + * uint16[segCount] + *

仅 Format=4 + *

每个段的结束字符码，最后一个是 0xFFFF，表示 Unicode 范围的结束。 + */ public int[] endCode; + /** + * uint16 + *

仅 Format=4 + *

固定设置为 0，用于填充保留位以保持数据对齐。 + */ public int reservedPad; + /** + * uint16[segCount] + *

仅 Format=4 + *

每个段的起始字符码。 + */ public int[] startCode; - public short[] idDelta; - public int[] idRangeOffset; - public int[] glyphIdArray; - } - - private static class CmapFormat6 extends CmapFormat { + /** + * int16[segCount] + *

仅 Format=4 + *

用于计算字形索引的偏移值。该值被加到从 startCode 到 endCode 的所有字符码上，得到相应的字形索引。 + */ + public int[] idDelta; + /** + * uint16[segCount] + *

仅 Format=4 + *

偏移到 glyphIdArray 中的起始位置，如果没有额外的字形索引映射，则为 0。 + */ + public int[] idRangeOffsets; + /** + * uint16 + *

仅 Format=6 + *

子范围的第一个字符代码。这是连续字符代码范围的起始点。 + */ public int firstCode; + /** + * uint16 + *

仅 Format=6 + *

子范围中字符代码的数量。这表示从 firstCode 开始，连续多少个字符代码被包含 + */ public int entryCount; + /** + * 字形索引数组 + *

Format=0 为 bye[256]数组 + *

Format>0 为 uint16[] 数组 + *

Format>12 为 uint32[] 数组 + *

@url Microsoft cmap文档 + */ public int[] glyphIdArray; } - private static class CmapFormat12 extends CmapFormat { - public int reserved; - public int length; - public int language; - public int numGroups; - public List groups; + /** + * 字形轮廓数据表 + */ + private static class GlyfLayout { + /** + * int16 非负值为简单字形的轮廓数,负值表示为复合字形 + */ + public short numberOfContours; + /** + * int16 Minimum x for coordinate data. + */ + public short xMin; + /** + * int16 Minimum y for coordinate data. + */ + public short yMin; + /** + * int16 Maximum x for coordinate data. + */ + public short xMax; + /** + * int16 Maximum y for coordinate data. + */ + public short yMax; + /** + * 简单字形数据 + */ + public GlyphTableBySimple glyphSimple; + /** + * 复合字形数据 + */ + public LinkedList glyphComponent; } - private static class GlyfLayout { - public short numberOfContours; // 非负值为简单字型,负值为复合字型 - public short xMin; - public short yMin; - public short xMax; - public short yMax; - public int[] endPtsOfContours; // length=numberOfContours - public int instructionLength; - public byte[] instructions; // length=instructionLength - public byte[] flags; - public short[] xCoordinates; // length = flags.length - public short[] yCoordinates; // length = flags.length + /** + * 简单字形数据表 + */ + private static class GlyphTableBySimple { + /** + * uint16[numberOfContours] + */ + int[] endPtsOfContours; + /** + * uint16 + */ + int instructionLength; + /** + * uint8[instructionLength] + */ + int[] instructions; + /** + * uint8[variable] + *

bit0: 该点位于曲线上 + *

bit1: < 1:xCoordinate为uint8 > + *

bit2: < 1:yCoordinate为uint8 > + *

bit3: < 1:下一个uint8为此条目之后插入的附加逻辑标志条目的数量 > + *

bit4: < bit1=1时表示符号[1.正,0.负]; bit1=0时[1.x坐标重复一次,0.x坐标读为int16] > + *

bit5: < bit2=1时表示符号[1.正,0.负]; bit2=0时[1.y坐标重复一次,0.y坐标读为int16] > + *

bit6: 字形描述中的轮廓可能会重叠 + *

bit7: 保留位,无意义 + */ + int[] flags; + /** + * uint8[] when(flags&0x02==0x02) + * int16[] when(flags&0x12==0x00) + */ + int[] xCoordinates; + /** + * uint8[] when(flags&0x04==0x02) + * int16[] when(flags&0x24==0x00) + */ + int[] yCoordinates; + } + + /** + * 复合字形数据表 + */ + private static class GlyphTableComponent { + /** + * uint16 + *

bit0: < 1:argument是16bit，0:argument是8bit > + *

bit1: < 1:argument是有符号值，0:argument是无符号值 > + *

bit3: 该组件有一个缩放比例，否则比例为1.0 + *

bit5: 表示在此字形之后还有字形 + */ + int flags; + /** + * uint16 + */ + int glyphIndex; + /** + * x-offset + *

uint8 when flags&0x03==0 + *

int8 when flags&0x03==1 + *

uint16 when flags&0x03==2 + *

int16 when flags&0x03==3 + */ + int argument1; + /** + * y-offset + *

uint8 when flags&0x03==0 + *

int8 when flags&0x03==1 + *

uint16 when flags&0x03==2 + *

int16 when flags&0x03==3 + */ + int argument2; + /** + * uint16 + *

值类型为 F2DOT14 的组件缩放X比例值 + */ + float xScale; + /** + * uint16 + *

值类型为 F2DOT14 的2x2变换矩阵01值 + */ + float scale01; + /** + * uint16 + *

值类型为 F2DOT14 的2x2变换矩阵10值 + */ + float scale10; + /** + * uint16 + *

值类型为 F2DOT14 的组件缩放Y比例值 + */ + float yScale; } private static class BufferReader { @@ -168,7 +520,6 @@ public class QueryTTF { return byteBuffer.position(); } - public long ReadUInt64() { return byteBuffer.getLong(); } @@ -198,39 +549,52 @@ public class QueryTTF { } public byte[] ReadByteArray(int len) { - if (len < 0) return null; + assert len >= 0; byte[] result = new byte[len]; byteBuffer.get(result); return result; } - public short[] ReadInt16Array(int len) { - if (len < 0) return null; - var result = new short[len]; + public int[] ReadUInt8Array(int len) { + assert len >= 0; + var result = new int[len]; + for (int i = 0; i < len; ++i) result[i] = byteBuffer.get() & 0xFF; + return result; + } + + public int[] ReadInt16Array(int len) { + assert len >= 0; + var result = new int[len]; for (int i = 0; i < len; ++i) result[i] = byteBuffer.getShort(); return result; } public int[] ReadUInt16Array(int len) { - if (len < 0) return null; + assert len >= 0; var result = new int[len]; for (int i = 0; i < len; ++i) result[i] = byteBuffer.getShort() & 0xFFFF; return result; } + + public int[] ReadInt32Array(int len) { + assert len >= 0; + var result = new int[len]; + for (int i = 0; i < len; ++i) result[i] = byteBuffer.getInt(); + return result; + } } private final Header fileHeader = new Header(); - private final Map directorys = new HashMap<>(); + private final HashMap directorys = new HashMap<>(); private final NameLayout name = new NameLayout(); private final HeadLayout head = new HeadLayout(); private final MaxpLayout maxp = new MaxpLayout(); - private final List loca = new LinkedList<>(); private final CmapLayout Cmap = new CmapLayout(); - private String[] glyf; private final int[][] pps = new int[][]{{3, 10}, {0, 4}, {3, 1}, {1, 0}, {0, 3}, {0, 1}}; private void readNameTable(byte[] buffer) { - var dataTable = Objects.requireNonNull(directorys.get("name")); + var dataTable = directorys.get("name"); + assert dataTable != null; var reader = new BufferReader(buffer, dataTable.offset); name.format = reader.ReadUInt16(); name.count = reader.ReadUInt16(); @@ -248,7 +612,8 @@ public class QueryTTF { } private void readHeadTable(byte[] buffer) { - var dataTable = Objects.requireNonNull(directorys.get("head")); + var dataTable = directorys.get("head"); + assert dataTable != null; var reader = new BufferReader(buffer, dataTable.offset); head.majorVersion = reader.ReadUInt16(); head.minorVersion = reader.ReadUInt16(); @@ -270,90 +635,117 @@ public class QueryTTF { head.glyphDataFormat = reader.ReadInt16(); } + /** + * glyfId到glyphData的索引 + *

根据定义，索引零指向“丢失的字符”。 + *

loca.length = maxp.numGlyphs + 1; + */ + private int[] loca; + + private void readLocaTable(byte[] buffer) { + var dataTable = directorys.get("loca"); + assert dataTable != null; + var reader = new BufferReader(buffer, dataTable.offset); + if (head.indexToLocFormat == 0) { + loca = reader.ReadUInt16Array(dataTable.length / 2); + } else { + loca = reader.ReadInt32Array(dataTable.length / 4); + } + } + private void readCmapTable(byte[] buffer) { - var dataTable = Objects.requireNonNull(directorys.get("cmap")); + var dataTable = directorys.get("cmap"); + assert dataTable != null; var reader = new BufferReader(buffer, dataTable.offset); Cmap.version = reader.ReadUInt16(); Cmap.numTables = reader.ReadUInt16(); for (int i = 0; i < Cmap.numTables; ++i) { CmapRecord record = new CmapRecord(); record.platformID = reader.ReadUInt16(); - record.platformSpecificID = reader.ReadUInt16(); + record.encodingID = reader.ReadUInt16(); record.offset = reader.ReadUInt32(); Cmap.records.add(record); } - for (int i = 0; i < Cmap.numTables; ++i) { - int fmtOffset = Cmap.records.get(i).offset; - reader.position(dataTable.offset + fmtOffset); - int EndIndex = reader.position(); - int format = reader.ReadUInt16(); + for (var formatTable : Cmap.records) { + int fmtOffset = formatTable.offset; if (Cmap.tables.containsKey(fmtOffset)) continue; - if (format == 0) { - CmapFormat f = new CmapFormat(); - f.format = format; - f.length = reader.ReadUInt16(); - f.language = reader.ReadUInt16(); - f.glyphIdArray = reader.ReadByteArray(f.length - 6); - Cmap.tables.put(fmtOffset, f); - } else if (format == 4) { - CmapFormat4 f = new CmapFormat4(); - f.format = format; - f.length = reader.ReadUInt16(); - f.language = reader.ReadUInt16(); - f.segCountX2 = reader.ReadUInt16(); - int segCount = f.segCountX2 >> 1; - f.searchRange = reader.ReadUInt16(); - f.entrySelector = reader.ReadUInt16(); - f.rangeShift = reader.ReadUInt16(); - f.endCode = reader.ReadUInt16Array(segCount); - f.reservedPad = reader.ReadUInt16(); - f.startCode = reader.ReadUInt16Array(segCount); - f.idDelta = reader.ReadInt16Array(segCount); - f.idRangeOffset = reader.ReadUInt16Array(segCount); - f.glyphIdArray = reader.ReadUInt16Array((EndIndex + f.length - reader.position()) >> 1); - Cmap.tables.put(fmtOffset, f); - } else if (format == 6) { - CmapFormat6 f = new CmapFormat6(); - f.format = format; - f.length = reader.ReadUInt16(); - f.language = reader.ReadUInt16(); - f.firstCode = reader.ReadUInt16(); - f.entryCount = reader.ReadUInt16(); - f.glyphIdArray = reader.ReadUInt16Array(f.entryCount); - Cmap.tables.put(fmtOffset, f); - } else if (format == 12) { - CmapFormat12 f = new CmapFormat12(); - f.format = format; - f.reserved = reader.ReadUInt16(); - f.length = reader.ReadUInt32(); - f.language = reader.ReadUInt32(); - f.numGroups = reader.ReadUInt32(); - f.groups = new LinkedList<>(); - for (int n = 0; n < f.numGroups; ++n) { - f.groups.add(new int[]{reader.ReadUInt32(), reader.ReadUInt32(), reader.ReadUInt32()}); - } - Cmap.tables.put(fmtOffset, f); - } - } - } + reader.position(dataTable.offset + fmtOffset); - private void readLocaTable(byte[] buffer) { - var dataTable = Objects.requireNonNull(directorys.get("loca")); - var reader = new BufferReader(buffer, dataTable.offset); - var locaTableSize = dataTable.length; - if (head.indexToLocFormat == 0) { - for (long i = 0; i < locaTableSize; i += 2) loca.add(reader.ReadUInt16()); - } else { - for (long i = 0; i < locaTableSize; i += 4) loca.add(reader.ReadUInt32()); + CmapFormat f = new CmapFormat(); + f.format = reader.ReadUInt16(); + f.length = reader.ReadUInt16(); + f.language = reader.ReadUInt16(); + switch (f.format) { + case 0: { + f.glyphIdArray = reader.ReadUInt8Array(f.length - 6); + // 记录 unicode->glyphId 映射表 + int unicodeInclusive = 0; + int unicodeExclusive = f.glyphIdArray.length; + for (; unicodeInclusive < unicodeExclusive; unicodeInclusive++) { + unicodeToGlyphId.put(unicodeInclusive, f.glyphIdArray[unicodeInclusive]); + } + break; + } + case 4: { + f.segCountX2 = reader.ReadUInt16(); + int segCount = f.segCountX2 / 2; + f.searchRange = reader.ReadUInt16(); + f.entrySelector = reader.ReadUInt16(); + f.rangeShift = reader.ReadUInt16(); + f.endCode = reader.ReadUInt16Array(segCount); + f.reservedPad = reader.ReadUInt16(); + f.startCode = reader.ReadUInt16Array(segCount); + f.idDelta = reader.ReadInt16Array(segCount); + f.idRangeOffsets = reader.ReadUInt16Array(segCount); + // 一个包含字形索引的数组，其长度是任意的，取决于映射的复杂性和字体中的字符数量。 + int glyphIdArrayLength = (f.length - 16 - (segCount * 8)) / 2; + f.glyphIdArray = reader.ReadUInt16Array(glyphIdArrayLength); + + // 记录 unicode->glyphId 映射表 + for (int segmentIndex = 0; segmentIndex < segCount; segmentIndex++) { + int unicodeInclusive = f.startCode[segmentIndex]; + int unicodeExclusive = f.endCode[segmentIndex]; + int idDelta = f.idDelta[segmentIndex]; + int idRangeOffset = f.idRangeOffsets[segmentIndex]; + for (int unicode = unicodeInclusive; unicode <= unicodeExclusive; unicode++) { + if (idRangeOffset == 0) { + unicodeToGlyphId.put(unicode, (unicode + idDelta) & 0xFFFF); + } else { + int gIndex = (idRangeOffset / 2) + unicode - unicodeInclusive + segmentIndex; + unicodeToGlyphId.put(unicode, gIndex < glyphIdArrayLength ? f.glyphIdArray[gIndex] + idDelta : 0); + } + } + } + break; + } + case 6: { + f.firstCode = reader.ReadUInt16(); + f.entryCount = reader.ReadUInt16(); + // 范围内字符代码的字形索引值数组。 + f.glyphIdArray = reader.ReadUInt16Array(f.entryCount); + + // 记录 unicode->glyphId 映射表 + int unicodeIndex = f.firstCode; + int unicodeCount = f.entryCount; + for (int gIndex = 0; gIndex < unicodeCount; gIndex++) { + unicodeToGlyphId.put(unicodeIndex, f.glyphIdArray[gIndex]); + unicodeIndex++; + } + break; + } + default: + break; + } + Cmap.tables.put(fmtOffset, f); } } private void readMaxpTable(byte[] buffer) { - var dataTable = Objects.requireNonNull(directorys.get("maxp")); + var dataTable = directorys.get("maxp"); + assert dataTable != null; var reader = new BufferReader(buffer, dataTable.offset); - maxp.majorVersion = reader.ReadUInt16(); - maxp.minorVersion = reader.ReadUInt16(); + maxp.version = reader.ReadUInt32(); maxp.numGlyphs = reader.ReadUInt16(); maxp.maxPoints = reader.ReadUInt16(); maxp.maxContours = reader.ReadUInt16(); @@ -370,80 +762,163 @@ public class QueryTTF { maxp.maxComponentDepth = reader.ReadUInt16(); } - private void readGlyfTable(byte[] buffer) { - var dataTable = Objects.requireNonNull(directorys.get("glyf")); - int glyfCount = maxp.numGlyphs; - glyf = new String[glyfCount]; + /** + * 字形轮廓表 数组 + */ + private GlyfLayout[] glyfArray; - int coreCount = Runtime.getRuntime().availableProcessors(); - ExecutorService executor = Executors.newFixedThreadPool(coreCount); - int sliceSize = (coreCount < glyfCount) ? (glyfCount / coreCount) : glyfCount; - for (int blockOffset = 0; blockOffset < glyfCount; blockOffset += sliceSize) { - final int inclusive = blockOffset; - final int exclusive = Math.min(blockOffset + sliceSize, glyfCount); - executor.submit(() -> { - var reader = new BufferReader(buffer, 0); - for (int index = inclusive; index < exclusive; index++) { - int offset = dataTable.offset + loca.get(index); - int glyfLength = (index + 1 < glyfCount ? loca.get(index + 1) : dataTable.offset + dataTable.length) - loca.get(index); - reader.position(offset); - byte[] glyph; - short numberOfContours = reader.ReadInt16(); - if (numberOfContours > 0) { - short g_xMin = reader.ReadInt16(); - short g_yMin = reader.ReadInt16(); - short g_xMax = reader.ReadInt16(); - short g_yMax = reader.ReadInt16(); - int[] endPtsOfContours = reader.ReadUInt16Array(numberOfContours); - int instructionLength = reader.ReadUInt16(); - var instructions = reader.ReadByteArray(instructionLength); - int flagLength = endPtsOfContours[endPtsOfContours.length - 1] + 1; - // 获取轮廓点描述标志 - var flags = new byte[flagLength]; - for (int n = 0; n < flagLength; ++n) { - flags[n] = reader.ReadInt8(); - if ((flags[n] & 0x08) != 0x00) { - for (int m = reader.ReadUInt8(); m > 0; --m) { - flags[++n] = flags[n - 1]; - } - } + private void readGlyfTable(byte[] buffer) { + var dataTable = directorys.get("glyf"); + assert dataTable != null; + int glyfCount = maxp.numGlyphs; + glyfArray = new GlyfLayout[glyfCount + 1]; // 创建容器时，多创建一个作为保留区 + + var reader = new BufferReader(buffer, 0); + for (int index = 1; index <= glyfCount; index++) { + if (loca[index - 1] == loca[index]) continue; // 当前loca与下一个loca相同，表示这个字形不存在 + int offset = dataTable.offset + loca[index - 1]; + // 读GlyphHeaders + var glyph = new GlyfLayout(); + reader.position(offset); + glyph.numberOfContours = reader.ReadInt16(); + glyph.xMin = reader.ReadInt16(); + glyph.yMin = reader.ReadInt16(); + glyph.xMax = reader.ReadInt16(); + glyph.yMax = reader.ReadInt16(); + + // 读Glyph轮廓数据 + if (glyph.numberOfContours > 0) { + // 简单轮廓 + glyph.glyphSimple = new GlyphTableBySimple(); + glyph.glyphSimple.endPtsOfContours = reader.ReadUInt16Array(glyph.numberOfContours); + glyph.glyphSimple.instructionLength = reader.ReadUInt16(); + glyph.glyphSimple.instructions = reader.ReadUInt8Array(glyph.glyphSimple.instructionLength); + int flagLength = glyph.glyphSimple.endPtsOfContours[glyph.glyphSimple.endPtsOfContours.length - 1] + 1; + // 获取轮廓点描述标志 + glyph.glyphSimple.flags = new int[flagLength]; + for (int n = 0; n < flagLength; ++n) { + var glyphSimpleFlag = reader.ReadUInt8(); + glyph.glyphSimple.flags[n] = glyphSimpleFlag; + if ((glyphSimpleFlag & 0x08) == 0x08) { + for (int m = reader.ReadUInt8(); m > 0; --m) { + glyph.glyphSimple.flags[++n] = glyphSimpleFlag; } - // 获取轮廓点描述x,y相对值 - ByteBuffer xyCoordinatesBuffer = ByteBuffer.allocate(flagLength * 4); - // 获取轮廓点描述x轴相对值 - for (int n = 0; n < flagLength; ++n) { - short same = (short) ((flags[n] & 0x10) != 0 ? 1 : -1); - if ((flags[n] & 0x02) != 0) { - xyCoordinatesBuffer.putShort((short) (same * reader.ReadUInt8())); - } else { - xyCoordinatesBuffer.putShort(same == 1 ? (short) 0 : reader.ReadInt16()); - } - } - // 获取轮廓点描述y轴相对值 - for (int n = 0; n < flagLength; ++n) { - short same = (short) ((flags[n] & 0x20) != 0 ? 1 : -1); - if ((flags[n] & 0x04) != 0) { - xyCoordinatesBuffer.putShort((short) (same * reader.ReadUInt8())); - } else { - xyCoordinatesBuffer.putShort(same == 1 ? (short) 0 : reader.ReadInt16()); - } - } - // 保存轮廓点描述x,y相对值ByteArray - glyph = xyCoordinatesBuffer.array(); - } else { - // TODO: 处理复合字形 - glyph = reader.ReadByteArray(glyfLength - 2); } - glyf[index] = getHexFromBytes(glyph); } - }); + // 获取轮廓点描述x轴相对值 + glyph.glyphSimple.xCoordinates = new int[flagLength]; + for (int n = 0; n < flagLength; ++n) { + switch (glyph.glyphSimple.flags[n] & 0x12) { + case 0x02: + glyph.glyphSimple.xCoordinates[n] = -1 * reader.ReadUInt8(); + break; + case 0x12: + glyph.glyphSimple.xCoordinates[n] = reader.ReadUInt8(); + break; + case 0x10: + glyph.glyphSimple.xCoordinates[n] = 0; // 点位数据重复上一次数据，那么相对数据变化量就是0 + break; + case 0x00: + glyph.glyphSimple.xCoordinates[n] = reader.ReadInt16(); + break; + } + } + // 获取轮廓点描述y轴相对值 + glyph.glyphSimple.yCoordinates = new int[flagLength]; + for (int n = 0; n < flagLength; ++n) { + switch (glyph.glyphSimple.flags[n] & 0x24) { + case 0x04: + glyph.glyphSimple.yCoordinates[n] = -1 * reader.ReadUInt8(); + break; + case 0x24: + glyph.glyphSimple.yCoordinates[n] = reader.ReadUInt8(); + break; + case 0x20: + glyph.glyphSimple.yCoordinates[n] = 0; // 点位数据重复上一次数据，那么相对数据变化量就是0 + break; + case 0x00: + glyph.glyphSimple.yCoordinates[n] = reader.ReadInt16(); + break; + } + } + } else { + // 复合轮廓 + glyph.glyphComponent = new LinkedList<>(); + while (true) { + var glyphTableComponent = new GlyphTableComponent(); + glyphTableComponent.flags = reader.ReadUInt16(); + glyphTableComponent.glyphIndex = reader.ReadUInt16(); + switch (glyphTableComponent.flags & 0b11) { + case 0b00: + glyphTableComponent.argument1 = reader.ReadUInt8(); + glyphTableComponent.argument2 = reader.ReadUInt8(); + break; + case 0b10: + glyphTableComponent.argument1 = reader.ReadInt8(); + glyphTableComponent.argument2 = reader.ReadInt8(); + break; + case 0b01: + glyphTableComponent.argument1 = reader.ReadUInt16(); + glyphTableComponent.argument2 = reader.ReadUInt16(); + break; + case 0b11: + glyphTableComponent.argument1 = reader.ReadInt16(); + glyphTableComponent.argument2 = reader.ReadInt16(); + break; + } + switch (glyphTableComponent.flags & 0b11001000) { + case 0b00001000: + // 有单一比例 + glyphTableComponent.yScale = glyphTableComponent.xScale = ((float) reader.ReadUInt16()) / 16384.0f; + break; + case 0b01000000: + // 有X和Y的独立比例 + glyphTableComponent.xScale = ((float) reader.ReadUInt16()) / 16384.0f; + glyphTableComponent.yScale = ((float) reader.ReadUInt16()) / 16384.0f; + break; + case 0b10000000: + // 有2x2变换矩阵 + glyphTableComponent.xScale = ((float) reader.ReadUInt16()) / 16384.0f; + glyphTableComponent.scale01 = ((float) reader.ReadUInt16()) / 16384.0f; + glyphTableComponent.scale10 = ((float) reader.ReadUInt16()) / 16384.0f; + glyphTableComponent.yScale = ((float) reader.ReadUInt16()) / 16384.0f; + break; + } + glyph.glyphComponent.add(glyphTableComponent); + if ((glyphTableComponent.flags & 0x20) == 0) break; + } + } + glyfArray[index] = glyph; // 根据文档 glyfId=0 作为保留区使用，这里赋值从索引1开始 } - executor.shutdown(); - try { - boolean b = executor.awaitTermination(60, TimeUnit.SECONDS); - } catch (InterruptedException e) { - Log.e("queryTTF", "glyf表解析出错: " + e); + } + + /** + * 使用轮廓索引值获取轮廓数据 + * + * @param glyfId 轮廓索引 + * @return 轮廓数据 + */ + public String getGlyfById(int glyfId) { + var glyph = glyfArray[glyfId]; + if (glyph == null) return null; // 过滤不存在的字体轮廓 + String glyphString; + if (glyph.numberOfContours >= 0) { + // 简单字形 + int dataCount = glyph.glyphSimple.flags.length; + String[] coordinateArray = new String[dataCount]; + for (int i = 0; i < dataCount; i++) { + coordinateArray[i] = glyph.glyphSimple.xCoordinates[i] + "," + glyph.glyphSimple.yCoordinates[i]; + } + glyphString = String.join("|", coordinateArray); + } else { + // 复合字形 + LinkedList glyphIdList = new LinkedList<>(); + for (var g : glyph.glyphComponent) { + glyphIdList.add(String.valueOf(g.glyphIndex)); + } + glyphString = "[" + String.join(",", glyphIdList) + "]"; } + return glyphString; } /** @@ -453,196 +928,107 @@ public class QueryTTF { */ public QueryTTF(final byte[] buffer) { var fontReader = new BufferReader(buffer, 0); - Log.i("queryTTF", "读文件头"); - // 获取文件头 - fileHeader.majorVersion = fontReader.ReadUInt16(); - fileHeader.minorVersion = fontReader.ReadUInt16(); - fileHeader.numOfTables = fontReader.ReadUInt16(); +// Log.i("QueryTTF", "读文件头"); // 获取文件头 + fileHeader.sfntVersion = fontReader.ReadUInt32(); + fileHeader.numTables = fontReader.ReadUInt16(); fileHeader.searchRange = fontReader.ReadUInt16(); fileHeader.entrySelector = fontReader.ReadUInt16(); fileHeader.rangeShift = fontReader.ReadUInt16(); // 获取目录 - for (int i = 0; i < fileHeader.numOfTables; ++i) { + for (int i = 0; i < fileHeader.numTables; ++i) { Directory d = new Directory(); - d.tag = new String(fontReader.ReadByteArray(4), StandardCharsets.US_ASCII); + d.tableTag = new String(fontReader.ReadByteArray(4), StandardCharsets.US_ASCII); d.checkSum = fontReader.ReadUInt32(); d.offset = fontReader.ReadUInt32(); d.length = fontReader.ReadUInt32(); - directorys.put(d.tag, d); + directorys.put(d.tableTag, d); } - Log.i("queryTTF", "解析表 name"); // 字体信息,包含版权、名称、作者等... +// Log.i("QueryTTF", "解析表 name"); // 字体信息,包含版权、名称、作者等... readNameTable(buffer); - Log.i("queryTTF", "解析表 head"); // 获取 head.indexToLocFormat +// Log.i("QueryTTF", "解析表 head"); // 获取 head.indexToLocFormat readHeadTable(buffer); - Log.i("queryTTF", "解析表 cmap"); // Unicode编码->轮廓索引 对照表 +// Log.i("QueryTTF", "解析表 cmap"); // Unicode编码->轮廓索引 对照表 readCmapTable(buffer); - Log.i("queryTTF", "解析表 loca"); // 轮廓数据偏移地址表 +// Log.i("QueryTTF", "解析表 loca"); // 轮廓数据偏移地址表 readLocaTable(buffer); - Log.i("queryTTF", "解析表 maxp"); // 获取 maxp.numGlyphs 字体轮廓数量 +// Log.i("QueryTTF", "解析表 maxp"); // 获取 maxp.numGlyphs 字体轮廓数量 readMaxpTable(buffer); - Log.i("queryTTF", "解析表 glyf"); // 字体轮廓数据表,需要解析loca,maxp表后计算 +// Log.i("QueryTTF", "解析表 glyf"); // 字体轮廓数据表,需要解析loca,maxp表后计算 readGlyfTable(buffer); - Log.i("queryTTF", "建立Unicode&Glyph映射表"); - makeHashTable(); - Log.i("queryTTF", "字体处理完成"); - } - - public final ConcurrentHashMap unicodeToGlyph = new ConcurrentHashMap<>(); - public final ConcurrentHashMap glyphToUnicode = new ConcurrentHashMap<>(); - public final ConcurrentHashMap unicodeToGlyphIndex = new ConcurrentHashMap<>(); - - private void makeHashTable() { - int coreCount = Runtime.getRuntime().availableProcessors(); - ExecutorService executor = Executors.newFixedThreadPool(coreCount); - int unicodeCount = 0xFFFF; - int sliceSize = (coreCount < unicodeCount) ? (unicodeCount / coreCount) : unicodeCount; - for (int blockOffset = 0; blockOffset < unicodeCount; blockOffset += sliceSize) { - final int inclusive = blockOffset; - final int exclusive = Math.min(blockOffset + sliceSize, unicodeCount); - executor.submit(() -> { - for (int index = inclusive; index < exclusive; index++) { - int gid = queryGlyfIndex(index); - if (gid < glyf.length) { - unicodeToGlyphIndex.put(index, gid); - var val = glyf[gid]; - unicodeToGlyph.put(index, val); - if (!glyphToUnicode.containsKey(val)) glyphToUnicode.put(val, index); - } - } - }); - } - executor.shutdown(); - try { - boolean b = executor.awaitTermination(Long.MAX_VALUE, TimeUnit.SECONDS); - } catch (InterruptedException e) { - Log.e("queryTTF", "建立Unicode&Glyph映射表出错: " + e); +// Log.i("QueryTTF", "建立Unicode&Glyph映射表"); + int glyfArrayLength = glyfArray.length; + for (var item : unicodeToGlyphId.entrySet()) { + int key = item.getKey(); + int val = item.getValue() + 1; // glyphArray已根据TTF文档将索引0作为保留位，这里从1开始索引 + if (val >= glyfArrayLength) continue; + String glyfString = getGlyfById(val); + unicodeToGlyph.put(key, glyfString); + glyphToUnicode.put(glyfString, key); } +// Log.i("QueryTTF", "字体处理完成"); } -// /** -// * 获取字体信息 (1=字体名称) -// * -// * @param nameId 传入十进制字体信息索引 -// * @return 返回查询结果字符串 -// */ -// public String getNameById(int nameId) { -// fontReader.index = Objects.requireNonNull(directorys.get("name")).offset; -// for (NameRecord record : name.records) { -// if (record.nameID != nameId) continue; -// fontReader.index += name.stringOffset + record.offset; -// return fontReader.ReadStrings(record.length, record.platformID == 1 ? StandardCharsets.UTF_8 : StandardCharsets.UTF_16BE); -// } -// return "error"; -// } + public final HashMap unicodeToGlyph = new HashMap<>(); + public final HashMap glyphToUnicode = new HashMap<>(); + public final HashMap unicodeToGlyphId = new HashMap<>(); - public String getGlyfById(int gid) { - return glyf[gid]; + /** + * 使用 Unicode 值获查询廓索引 + * + * @param unicode 传入 Unicode 值 + * @return 轮廓索引 + */ + public int getGlyfIdByUnicode(int unicode) { + var result = unicodeToGlyphId.get(unicode); + if (result == null) return 0; + return result + 1; // 根据TTF文档，轮廓索引的定义从1开始 } /** - * 使用Unicode值查找轮廓索引 + * 使用 Unicode 值查询轮廓数据 * - * @param unicode 传入Unicode值 - * @return 返回十进制轮廓索引 + * @param unicode 传入 Unicode 值 + * @return 轮廓数据 */ - public int queryGlyfIndex(int unicode) { - if (unicode == 0) return 0; - - int fmtKey = 0; - for (var item : pps) { - for (CmapRecord record : Cmap.records) { - if ((item[0] == record.platformID) && (item[1] == record.platformSpecificID)) { - fmtKey = record.offset; - break; - } - } - if (fmtKey > 0) break; - } - if (fmtKey == 0) return 0; - - int glyfID = 0; - CmapFormat table = Cmap.tables.get(fmtKey); - assert table != null; - int fmt = table.format; - if (fmt == 0) { - if (unicode < table.glyphIdArray.length) glyfID = table.glyphIdArray[unicode] & 0xFF; - } else if (fmt == 4) { - CmapFormat4 tab = (CmapFormat4) table; - if (unicode > tab.endCode[tab.endCode.length - 1]) return 0; - // 二分法查找数值索引 - int start = 0, middle, end = tab.endCode.length - 1; - while (start + 1 < end) { - middle = (start + end) / 2; - if (tab.endCode[middle] <= unicode) start = middle; - else end = middle; - } - if (tab.endCode[start] < unicode) ++start; - if (unicode < tab.startCode[start]) return 0; - if (tab.idRangeOffset[start] != 0) { - glyfID = tab.glyphIdArray[unicode - tab.startCode[start] + (tab.idRangeOffset[start] >> 1) - (tab.idRangeOffset.length - start)]; - } else glyfID = unicode + tab.idDelta[start]; - glyfID &= 0xFFFF; - } else if (fmt == 6) { - CmapFormat6 tab = (CmapFormat6) table; - int index = unicode - tab.firstCode; - if (0 <= index && index < tab.glyphIdArray.length) glyfID = tab.glyphIdArray[index]; - } else if (fmt == 12) { - CmapFormat12 tab = (CmapFormat12) table; - if (unicode > tab.groups.get(tab.numGroups - 1)[1]) return 0; - for (int i = 0; i < tab.numGroups; i++) { - if (tab.groups.get(i)[0] <= unicode && unicode <= tab.groups.get(i)[1]) { - glyfID = tab.groups.get(i)[2] + unicode - tab.groups.get(i)[0]; - break; - } - } - } - return glyfID; + public String getGlyfByUnicode(int unicode) { + return unicodeToGlyph.get(unicode); } /** - * 使用Unicode值获取轮廓索引 + * 使用轮廓数据反查 Unicode 值 * - * @param unicode 传入Unicode值 - * @return 返回十进制轮廓索引 + * @param glyph 传入轮廓数据 + * @return Unicode */ - public int getGlyfIndex(int unicode) { - var glyfIndex = unicodeToGlyphIndex.get(unicode); - if (glyfIndex == null) return 0; - return glyfIndex; + public int getUnicodeByGlyf(String glyph) { + var result = glyphToUnicode.get(glyph); + if (result == null) return 0; + return result; } /** - * 使用Unicode值获取轮廓数据 + * Unicode 空白字符判断 * - * @param unicode 传入Unicode值 - * @return 返回轮廓数组的String值 + * @param unicode 字符的 Unicode 值 + * @return true:是空白字符; false:非空白字符 */ - public String getGlyfByCode(int unicode) { - return unicodeToGlyph.getOrDefault(unicode, ""); - } - - /** - * 使用轮廓数据获取Unicode值 - * - * @param glyph 传入轮廓数组的String值 - * @return 返回Unicode十进制值 - */ - public int getCodeByGlyf(String glyph) { - //noinspection ConstantConditions - return glyphToUnicode.getOrDefault(glyph, 0); - } - - /** - * 字体轮廓数据转Hex字符串 - * - * @param glyph 字体轮廓数据 - * @return 返回轮廓数组的String值 - */ - public String getHexFromBytes(byte[] glyph) { - if (glyph == null) return ""; - StringBuilder hexString = new StringBuilder(); - for (byte b : glyph) hexString.append(String.format("%02X", b)); - return hexString.toString(); + public boolean isBlankUnicode(int unicode) { + return switch (unicode) { + case 0x0009, // 水平制表符 (Horizontal Tab) + 0x0020, // 空格 (Space) + 0x00A0, // 不中断空格 (No-Break Space) + 0x2002, // En空格 (En Space) + 0x2003, // Em空格 (Em Space) + 0x2007, // 刚性空格 (Figure Space) + 0x200A, // 发音修饰字母的连字符 (Hair Space) + 0x200B, // 零宽空格 (Zero Width Space) + 0x200C, // 零宽不连字 (Zero Width Non-Joiner) + 0x200D, // 零宽连字 (Zero Width Joiner) + 0x202F, // 狭窄不中断空格 (Narrow No-Break Space) + 0x205F // 中等数学空格 (Medium Mathematical Space) + -> true; + default -> false; + }; } }