1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
|
// 首先是基础的读取字节数据
// u32函数，在i位置后面读取32bit，返回uint32
// u32 returns the big-endian uint32 at b[i:].
func u32(b []byte, i int) uint32 {
return uint32(b[i])<<24 | uint32(b[i+1])<<16 | uint32(b[i+2])<<8 | uint32(b[i+3])
}
// u16 returns the big-endian uint16 at b[i:].
func u16(b []byte, i int) uint16 {
return uint16(b[i])<<8 | uint16(b[i+1])
}
// 读取表字节数据
// readTable returns a slice of the TTF data given by a table's directory entry.
func readTable(ttf []byte, offsetLength []byte) ([]byte, error) {
// 偏移量获取
offset := int(u32(offsetLength, 0))
if offset < 0 {
return nil, FormatError(fmt.Sprintf("offset too large: %d", uint32(offset)))
}
// 长度获取
length := int(u32(offsetLength, 4))
if length < 0 {
return nil, FormatError(fmt.Sprintf("length too large: %d", uint32(length)))
}
end := offset + length
if end < 0 || end > len(ttf) {
return nil, FormatError(fmt.Sprintf("offset + length too large: %d", uint32(offset)+uint32(length)))
}
// 返回字节内容
return ttf[offset:end], nil
}
// parse ttf文件解析核心函数，offset初始为0
func parse(ttf []byte, offset int) (font *Font, err error) {
if len(ttf)-offset < 12 {
err = FormatError("TTF data is too short")
return
}
originalOffset := offset
// 获取scaler type
magic, offset := u32(ttf, offset), offset+4
switch magic {
case 0x00010000:
// No-op.
case 0x74746366: // "ttcf" as a big-endian uint32.
if originalOffset != 0 {
err = FormatError("recursive TTC")
return
}
ttcVersion, offset := u32(ttf, offset), offset+4
if ttcVersion != 0x00010000 && ttcVersion != 0x00020000 {
err = FormatError("bad TTC version")
return
}
numFonts, offset := int(u32(ttf, offset)), offset+4
if numFonts <= 0 {
err = FormatError("bad number of TTC fonts")
return
}
if len(ttf[offset:])/4 < numFonts {
err = FormatError("TTC offset table is too short")
return
}
// TODO: provide an API to select which font in a TrueType collection to return,
// not just the first one. This may require an API to parse a TTC's name tables,
// so users of this package can select the font in a TTC by name.
offset = int(u32(ttf, offset))
if offset <= 0 || offset > len(ttf) {
err = FormatError("bad TTC offset")
return
}
return parse(ttf, offset)
default:
err = FormatError("bad TTF version")
return
}
// 读取表个数n
n, offset := int(u16(ttf, offset)), offset+2
offset += 6 // Skip the searchRange, entrySelector and rangeShift.
if len(ttf) < 16*n+offset {
err = FormatError("TTF data is too short")
return
}
f := new(Font)
// Assign the table slices.
for i := 0; i < n; i++ {
x := 16*i + offset
switch string(ttf[x : x+4]) {
// 读取cmap表字节数据
case "cmap":
f.cmap, err = readTable(ttf, ttf[x+8:x+16])
...
}
if err != nil {
return
}
}
...
// 解析cmap表
if err = f.parseCmap(); err != nil {
return
}
...
font = f
return
}
func (f *Font) parseCmap() error {
const (
cmapFormat4 = 4
cmapFormat12 = 12
languageIndependent = 0
)
// cmap子表解析，返回最佳的使用子表
offset, _, err := parseSubtables(f.cmap, "cmap", 4, 8, nil)
if err != nil {
return err
}
offset = int(u32(f.cmap, offset+4))
if offset <= 0 || offset > len(f.cmap) {
return FormatError("bad cmap offset")
}
// 读取cmap 格式类型
cmapFormat := u16(f.cmap, offset)
switch cmapFormat {
// format4 格式解析
case cmapFormat4:
language := u16(f.cmap, offset+4)
if language != languageIndependent {
return UnsupportedError(fmt.Sprintf("language: %d", language))
}
segCountX2 := int(u16(f.cmap, offset+6))
if segCountX2%2 == 1 {
return FormatError(fmt.Sprintf("bad segCountX2: %d", segCountX2))
}
// 计算段个数
segCount := segCountX2 / 2
offset += 14
f.cm = make([]cm, segCount)
// 读取end数组
for i := 0; i < segCount; i++ {
f.cm[i].end = uint32(u16(f.cmap, offset))
offset += 2
}
// 跳过字段
offset += 2
// 读取start数组
for i := 0; i < segCount; i++ {
f.cm[i].start = uint32(u16(f.cmap, offset))
offset += 2
}
// 读取delta数组
for i := 0; i < segCount; i++ {
f.cm[i].delta = uint32(u16(f.cmap, offset))
offset += 2
}
// 读取offset数组
for i := 0; i < segCount; i++ {
f.cm[i].offset = uint32(u16(f.cmap, offset))
offset += 2
}
// 字形下标数组
f.cmapIndexes = f.cmap[offset:]
return nil
case cmapFormat12:
if u16(f.cmap, offset+2) != 0 {
return FormatError(fmt.Sprintf("cmap format: % x", f.cmap[offset:offset+4]))
}
length := u32(f.cmap, offset+4)
language := u32(f.cmap, offset+8)
if language != languageIndependent {
return UnsupportedError(fmt.Sprintf("language: %d", language))
}
nGroups := u32(f.cmap, offset+12)
if length != 12*nGroups+16 {
return FormatError("inconsistent cmap length")
}
offset += 16
f.cm = make([]cm, nGroups)
for i := uint32(0); i < nGroups; i++ {
f.cm[i].start = u32(f.cmap, offset+0)
f.cm[i].end = u32(f.cmap, offset+4)
f.cm[i].delta = u32(f.cmap, offset+8) - f.cm[i].start
offset += 12
}
return nil
}
return UnsupportedError(fmt.Sprintf("cmap format: %d", cmapFormat))
}
// 子表选择，规定如下
// 优先使用unicode编码表
// 优先使用非BMP限制的unicode表
// 如果有4、12格式的unicode表，则优先使用
func parseSubtables(table []byte, name string, offset, size int, pred func([]byte) bool) (
bestOffset int, bestPID uint32, retErr error) {
if len(table) < 4 {
return 0, 0, FormatError(name + " too short")
}
// 表个数读取
nSubtables := int(u16(table, 2))
if len(table) < size*nSubtables+offset {
return 0, 0, FormatError(name + " too short")
}
ok := false
for i := 0; i < nSubtables; i, offset = i+1, offset+size {
if pred != nil && !pred(table[offset:]) {
continue
}
// We read the 16-bit Platform ID and 16-bit Platform Specific ID as a single uint32.
// All values are big-endian.
pidPsid := u32(table, offset)
// We prefer the Unicode cmap encoding. Failing to find that, we fall
// back onto the Microsoft cmap encoding.
if pidPsid == unicodeEncodingBMPOnly || pidPsid == unicodeEncodingFull {
bestOffset, bestPID, ok = offset, pidPsid>>16, true
break
} else if pidPsid == microsoftSymbolEncoding ||
pidPsid == microsoftUCS2Encoding ||
pidPsid == microsoftUCS4Encoding {
bestOffset, bestPID, ok = offset, pidPsid>>16, true
// We don't break out of the for loop, so that Unicode can override Microsoft.
}
}
if !ok {
return 0, 0, UnsupportedError(name + " encoding")
}
return bestOffset, bestPID, nil
}
// cmap数据使用
// Index returns a Font's index for the given rune.
func (f *Font) Index(x rune) Index {
c := uint32(x)
// 二分查找，找打x所在范围的cm
for i, j := 0, len(f.cm); i < j; {
h := i + (j-i)/2
cm := &f.cm[h]
if c < cm.start {
j = h
} else if cm.end < c {
i = h + 1
} else if cm.offset == 0 {
// offset为0，则直接返回delta+c
return Index(c + cm.delta)
} else {
// offset不是0 ，则offset+ ？ +c在cm里的偏移量
offset := int(cm.offset) + 2*(h-len(f.cm)+int(c-cm.start))
return Index(u16(f.cmapIndexes, offset))
}
}
return 0
}
|