feat(编辑能力): 修复字体选择与裁剪边界问题

ofdgo_editor_font.go

@@ -16,13 +16,167 @@ package ofdgo
 
 import (
 	"bytes"
+	"encoding/binary"
 	"fmt"
+	"maps"
 	"slices"
+	"strconv"
+	"strings"
 
 	"github.com/tdewolff/font"
 )
 
-// subsetFonts 收集新增字体实际使用的字形，仅裁剪保存结果，不修改编辑资源或原文档字体。
+// editorFontUsage 汇总原文档字体的字符、显式字形及无法确定的引用
+type editorFontUsage struct {
+	chars  map[rune]bool
+	glyphs map[uint16]bool
+	unsafe bool
+}
+
+// newEditorFontUsage 创建字体用字记录
+// 返回: *editorFontUsage 用字记录
+func newEditorFontUsage() *editorFontUsage {
+	return &editorFontUsage{chars: make(map[rune]bool), glyphs: make(map[uint16]bool)}
+}
+
+// fontUsage 获取规范化资源标识对应的字体用字记录
+// 入参: id 字体资源标识
+// 返回: *editorFontUsage 用字记录
+func (r *editorResourceRefs) fontUsage(id string) *editorFontUsage {
+	if r.fonts == nil {
+		r.fonts = make(map[string]*editorFontUsage)
+	}
+	id = editorResourceID(id)
+	if r.fonts[id] == nil {
+		r.fonts[id] = newEditorFontUsage()
+	}
+	return r.fonts[id]
+}
+
+// text 收集TextCode字符和CGTransform显式字形，异常编号使该字体保持原样
+// 入参: key 元素名称, value 元素文本
+func (u *editorFontUsage) text(key, value string) {
+	switch key {
+	case "TextCode":
+		for _, char := range textCodeRunes(value) {
+			u.chars[char] = true
+		}
+	case "Glyphs":
+		for _, value := range strings.Fields(strings.ReplaceAll(value, ",", " ")) {
+			id, err := strconv.ParseUint(value, 10, 16)
+			if err != nil {
+				u.unsafe = true
+				return
+			}
+			u.glyphs[uint16(id)] = true
+		}
+	}
+}
+
+// merge 合并指向同一字体文件的全部资源用字
+// 入参: other 用字记录, unsafe 是否存在无法确定的引用
+func (u *editorFontUsage) merge(other *editorFontUsage, unsafe bool) {
+	u.unsafe = u.unsafe || unsafe
+	if other != nil {
+		u.unsafe = u.unsafe || other.unsafe
+		maps.Copy(u.chars, other.chars)
+		maps.Copy(u.glyphs, other.glyphs)
+	}
+}
+
+// subsetSourceFont 裁剪静态TrueType原有字体，保留字形编号、复合依赖、度量及提示指令
+// 集合、字形替换、可变、彩色和未知表保持原样，不对不完整的引用或异常字体猜测修复
+// 入参: data 字体数据, usage 全包用字记录
+// 返回: []byte 更小的字体子集，无确定收益时为空
+func subsetSourceFont(data []byte, usage *editorFontUsage) []byte {
+	if usage.unsafe || len(usage.chars)+len(usage.glyphs) == 0 || bytes.HasPrefix(data, []byte("ttcf")) {
+		return nil
+	}
+	sfnt, err := font.ParseSFNT(data, 0)
+	if err != nil || !sfnt.IsTrueType {
+		return nil
+	}
+	for tag := range sfnt.Tables {
+		switch tag {
+		case "cmap", "head", "hhea", "hmtx", "maxp", "OS/2", "post", "name", "glyf", "loca", "cvt ", "fpgm", "prep", "gasp", "kern", "vhea", "vmtx", "hdmx", "LTSH", "VDMX", "GDEF", "GPOS", "DSIG", "FFTM":
+		default:
+			return nil
+		}
+	}
+	var unicodeTables []uint16
+	subtables := make(map[uint32]uint16)
+	for i, record := range sfnt.Cmap.EncodingRecords {
+		if record.Format == 14 {
+			return nil
+		}
+		offset := binary.BigEndian.Uint32(sfnt.Tables["cmap"][8+8*i:])
+		index, ok := subtables[offset]
+		if !ok {
+			index = uint16(len(subtables))
+			subtables[offset] = index
+		}
+		if record.PlatformID == 0 || record.PlatformID == 3 && (record.EncodingID == 1 || record.EncodingID == 10) {
+			unicodeTables = append(unicodeTables, index)
+		}
+	}
+	if len(unicodeTables) == 0 {
+		return nil
+	}
+	glyphs := maps.Clone(usage.glyphs)
+	glyphs[0] = true
+	mapping := make(map[rune]uint16, len(usage.chars))
+	for char := range usage.chars {
+		id := sfnt.GlyphIndex(char)
+		for _, index := range unicodeTables {
+			if alternate, ok := sfnt.Cmap.Subtables[index].Get(char); ok && alternate != id {
+				return nil
+			}
+		}
+		glyphs[id] = true
+		if id != 0 {
+			mapping[char] = id
+		}
+	}
+	for id := range glyphs {
+		if id >= sfnt.NumGlyphs() {
+			return nil
+		}
+		dependencies, err := sfnt.Glyf.Dependencies(id)
+		if err != nil {
+			return nil
+		}
+		for _, dependency := range dependencies {
+			glyphs[dependency] = true
+		}
+	}
+	var glyf, loca []byte
+	for id := 0; id <= int(sfnt.NumGlyphs()); id++ {
+		if sfnt.Head.IndexToLocFormat == 0 {
+			loca = binary.BigEndian.AppendUint16(loca, uint16(len(glyf)/2))
+		} else {
+			loca = binary.BigEndian.AppendUint32(loca, uint32(len(glyf)))
+		}
+		if id < int(sfnt.NumGlyphs()) && glyphs[uint16(id)] {
+			glyf = append(glyf, sfnt.Glyf.Get(uint16(id))...)
+			if len(glyf)%2 != 0 {
+				glyf = append(glyf, 0)
+			}
+		}
+	}
+	tables := maps.Clone(sfnt.Tables)
+	tables["glyf"], tables["loca"] = glyf, loca
+	tables["cmap"] = buildCmapTable(uint16(sfnt.NumGlyphs()), mapping)
+	tables["head"] = bytes.Clone(tables["head"])
+	clear(tables["head"][8:12])
+	delete(tables, "DSIG")
+	result, err := serializeOTF(tables)
+	if err != nil || len(result) >= len(data) {
+		return nil
+	}
+	return result
+}
+
+// subsetFonts 收集新增字体实际使用的字形，仅裁剪保存结果，不修改编辑资源或原文档字体
 // 返回: map[string][]byte 包内字体子集, error 错误信息
 func (e *Editor) subsetFonts() (map[string][]byte, error) {
 	used := make(map[string]map[uint16]bool)
@@ -77,7 +231,7 @@ func (e *Editor) subsetFonts() (map[string][]byte, error) {
 	return result, nil
 }
 
-// subsetEditorFont 裁剪TrueType字形及复合依赖，保留字符映射、名称、度量和提示指令。
+// subsetEditorFont 裁剪TrueType字形及复合依赖，保留字符映射、名称、度量和提示指令
 // 入参: data 字体数据, glyphs 已排序的字形编号，包含0
 // 返回: []byte 字体子集, error 错误信息
 func subsetEditorFont(data []byte, glyphs []uint16) ([]byte, error) {