Added more font tests

pdf/model/font_test.go

@@ -16,6 +16,7 @@ import (
 	"github.com/unidoc/unidoc/pdf/core"
 	"github.com/unidoc/unidoc/pdf/model"
 	"github.com/unidoc/unidoc/pdf/model/fonts"
+	"github.com/unidoc/unidoc/pdf/model/textencoding"
 )
 
 func init() {
@@ -252,6 +253,50 @@ var charcodeBytesToUnicodeTest = []fontFragmentTest{
 			177, 151, 178, 179, 183, 185, 188, 205, 184, 189},
 		"‘ł’ “Ł” Ø `o´ it's ˝ˆ˜¯˘˙¨˚ˇªº‹ı›—–—†‡•„…˛¸‰",
 	},
+	fontFragmentTest{"base glyphs′",
+		"../../contrib/testdata/font/cover.txt", 11,
+		[]byte{44, 45, 46, 48, 49, 50, 51, 53, 54, 55, 56, 58, 59,
+			65, 66, 67, 68, 69, 70, 71, 72,
+			84, 85,
+			97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 108, 109, 110, 111,
+			114, 115, 116, 117},
+		",-.01235678:;ABCDEFGHTUabcdefghijlmnorstu",
+	},
+	fontFragmentTest{"tex glyphs 48->′",
+		"../../contrib/testdata/font/noise-contrast.txt", 36,
+		[]byte{33, 48, 65, 104, 149, 253},
+		"!′Ah•ý",
+	},
+	fontFragmentTest{"tex2 glyphs ",
+		"../../contrib/testdata/font/Weil.txt", 30,
+		[]byte{55, 0, 1, 2, 20, 24, 33, 50, 102, 103, 104, 105},
+		"↦−·×≤∼→∈{}⟨⟩",
+	},
+	fontFragmentTest{"additional glyphs",
+		"../../contrib/testdata/font/noise-contrast.txt", 34,
+		[]byte{32, 40, 48, 64, 80, 88, 65, 104, 149, 253},
+		"({∑∑h•ý",
+	},
+	fontFragmentTest{".notdef glyphs",
+		"../../contrib/testdata/font/lec10.txt", 6,
+		[]byte{59, 66},
+		string([]rune{textencoding.MissingCodeRune, textencoding.MissingCodeRune}),
+	},
+	fontFragmentTest{"Numbered glyphs pattern 1",
+		"../../contrib/testdata/font/v14.txt", 14,
+		[]byte{24, 25, 26, 27, 29},
+		" ffifflfffi",
+	},
+	fontFragmentTest{"Glyph aliases",
+		"../../contrib/testdata/font/townes.txt", 10,
+		[]byte{2, 3, 4, 5, 6, 7, 1, 8, 9, 5, 1, 10, 9, 5, 48},
+		"Townes van Zan…",
+	},
+	fontFragmentTest{"Glyph `.` extensions. e.g. `integral.disp`",
+		"../../contrib/testdata/font/preview.txt", 156,
+		[]byte{83, 0, 4, 67, 62, 64, 100, 65},
+		"∫=≈≥∈<d>",
+	},
 	fontFragmentTest{"A potpourri of glyph naming conventions",
 		"../../contrib/testdata/font/Ingmar.txt", 144,
 		[]byte{18, 20, 10, 11, 13, 14, 15, 16, 21, 22, 23, 25, 26, 27, 28, 29, 30,

pdf/model/fonts/ttfparser.go

@@ -374,7 +374,7 @@ func (t *ttfParser) parseCmapSubtable10(offset10 int64) error {
 		length = t.ReadULong()
 		language = t.ReadULong()
 	}
-	common.Log.Debug("parseCmapSubtable10: format=%d length=%d language=%d",
+	common.Log.Trace("parseCmapSubtable10: format=%d length=%d language=%d",
 		format, length, language)
 
 	if format != 0 {
@@ -402,7 +402,7 @@ func (t *ttfParser) ParseCmap() error {
 	if err := t.Seek("cmap"); err != nil {
 		return err
 	}
-	common.Log.Debug("ParseCmap")
+	common.Log.Trace("ParseCmap")
 	t.ReadUShort() // version is ignored.
 	numTables := int(t.ReadUShort())
 	offset10 := int64(0)

pdf/model/textencoding/glyphlist/utils/glyphparser/parse_all_glyphs.go

@@ -98,6 +98,8 @@ func buildAll() error {
 		return err
 	}
 
+	gs.update("elipsis", map[string]rune{"elipsis": 0x2026}) // …
+
 	printAliases(gs.aliases)
 	printGlyphToRuneList(gs.glyphRune)
 	printRuneToGlyphList(gs.glyphRune)
@@ -176,14 +178,9 @@ func printGlyphToRuneList(glyphRune map[string]rune) {
 	keys := sorted(glyphRune)
 
 	fmt.Printf("var glyphlistGlyphToRuneMap = map[string]rune{ // %d entries \n", len(keys))
-	runes := map[rune]bool{}
 	for _, glyph := range keys {
 		r := glyphRune[glyph]
 		fmt.Printf("\t\t%q:\t0x%04x, %s\n", glyph, r, showRune(r))
-		if _, ok := runes[r]; ok {
-			panic("duplicate runes")
-		}
-		runes[r] = true
 	}
 	fmt.Printf("}\n")
 }
@@ -194,16 +191,9 @@ func printRuneToGlyphList(glyphRune map[string]rune) {
 	keys := sorted(glyphRune)
 
 	fmt.Printf("var glyphlistRuneToGlyphMap = map[rune]string{ // %d entries \n", len(keys))
-	runes := map[rune]bool{}
 	for _, glyph := range keys {
 		r := glyphRune[glyph]
 		fmt.Printf("\t\t0x%04x:\t%q, %s\n", r, glyph, showRune(r))
-		if _, ok := runes[r]; ok {
-			panic("duplicate runes")
-		}
-		if r == 0xfffd {
-			fmt.Fprintf(os.Stderr, "$$ %c->%q\n", r, glyph)
-		}
 	}
 	fmt.Printf("}\n")
 }
@@ -250,17 +240,6 @@ func sorted(glyphRune map[string]rune) []string {
 		}
 		return si < sj
 	})
-	if len(keys) != len(glyphRune) {
-		panic("gggk")
-	}
-
-	runes := map[rune]bool{}
-	for _, g := range keys {
-		runes[glyphRune[g]] = true
-	}
-	if len(keys) != len(runes) {
-		panic("h24")
-	}
 	return keys
 }
 
@@ -475,10 +454,6 @@ func parseUnknown(filename string) (map[string]rune, error) {
 			fmt.Fprintf(os.Stderr, "parseUnknown: No match line %d: %q\n", i, ln)
 			return nil, errors.New("bad line")
 		}
-		if strings.Contains(ln, "kafmedial") {
-			fmt.Fprintf(os.Stderr, "##@ %q %q\n", ln, g[1:])
-			// panic("rrr")
-		}
 		groups = append(groups, g[1:])
 	}
 	fmt.Printf("// groups=%d\n", len(groups))
@@ -493,26 +468,13 @@ func parseUnknown(filename string) (map[string]rune, error) {
 		r := rune(n)
 		parts := reSpace.Split(row[1], -1)
 		for _, g := range parts {
-			if g == "" {
+			if g == "" || strings.Contains(g, "000") || strings.Contains(g, "0.0") {
 				continue
 			}
-			if strings.Contains(g, "f_i") {
-				fmt.Fprintf(os.Stderr, "### %q\n", row)
-				// panic("rrr")
-			}
-			if strings.Contains(g, "000") {
-				continue
-			}
-			if strings.Contains(g, " ") {
-				panic("space")
-			}
 			glyph2rune[g] = r
 		}
 	}
 	fmt.Printf("// entries=%d\n", len(glyph2rune))
-	if _, ok := glyph2rune["kafmedial"]; !ok {
-		panic("kafmedial")
-	}
 	return glyph2rune, nil
 }
 
@@ -538,6 +500,7 @@ type glyphRune struct {
 }
 
 var baseNames = []string{"SymbolEncoding", "WinAnsiEncoding", "ZapfDingbatsEncoding"}
+
 var basicEncodings = map[string]map[uint16]glyphRune{
 	"SymbolEncoding": map[uint16]glyphRune{ // 189 entries
 		0x20: {"space", '\u0020'},         //