Merge pull request #247 from dennwc/fonts_minor_2

Minor improvements to the fonts encoders

pdf/internal/textencoding/encoder.go

@@ -6,6 +6,8 @@
 package textencoding
 
 import (
+	"encoding/binary"
+
 	"github.com/unidoc/unidoc/common"
 	"github.com/unidoc/unidoc/pdf/core"
 )
@@ -50,12 +52,13 @@ type TextEncoder interface {
 
 // Convenience functions
 
-// doEncode converts a Go unicode string `raw` to a PDF encoded string using the encoder `enc`.
-func doEncode(enc TextEncoder, raw string) []byte {
-	encoded := []byte{}
+// encodeString8bit converts a Go unicode string `raw` to a PDF encoded string using the encoder `enc`.
+// It expects that character codes will fit into a single byte.
+func encodeString8bit(enc TextEncoder, raw string) []byte {
+	encoded := make([]byte, 0, len(raw))
 	for _, r := range raw {
 		code, found := enc.RuneToCharcode(r)
-		if !found {
+		if !found || code > 0xff {
 			common.Log.Debug("Failed to map rune to charcode for rune 0x%04x", r)
 			continue
 		}
@@ -64,6 +67,27 @@ func doEncode(enc TextEncoder, raw string) []byte {
 	return encoded
 }
 
+// encodeString16bit converts a Go unicode string `raw` to a PDF encoded string using the encoder `enc`.
+// Each character will be encoded as two bytes.
+func encodeString16bit(enc TextEncoder, raw string) []byte {
+	// runes -> character codes -> bytes
+	runes := []rune(raw)
+	encoded := make([]byte, 0, len(runes)*2)
+	for _, r := range runes {
+		code, ok := enc.RuneToCharcode(r)
+		if !ok {
+			common.Log.Debug("Failed to map rune to charcode. rune=%+q", r)
+			continue
+		}
+
+		// Each entry represented by 2 bytes.
+		var v [2]byte
+		binary.BigEndian.PutUint16(v[:], code)
+		encoded = append(encoded, v[:]...)
+	}
+	return encoded
+}
+
 // doRuneToCharcode converts rune `r` to a PDF character code.
 // The bool return flag is true if there was a match, and false otherwise.
 func doRuneToCharcode(enc TextEncoder, r rune) (uint16, bool) {

pdf/internal/textencoding/identity.go

@@ -6,10 +6,10 @@
 package textencoding
 
 import (
-	"bytes"
 	"fmt"
+	"strconv"
+	"strings"
 
-	"github.com/unidoc/unidoc/common"
 	"github.com/unidoc/unidoc/pdf/core"
 )
 
@@ -31,20 +31,7 @@ func (enc IdentityEncoder) String() string {
 
 // Encode converts the Go unicode string `raw` to a PDF encoded string.
 func (enc IdentityEncoder) Encode(raw string) []byte {
-	// runes -> character codes -> bytes
-	var encoded bytes.Buffer
-	for _, r := range raw {
-		code, ok := enc.RuneToCharcode(r)
-		if !ok {
-			common.Log.Debug("Failed to map rune to charcode. rune=%+q", r)
-			continue
-		}
-
-		// Each entry represented by 2 bytes.
-		encoded.WriteByte(byte((code & 0xff00) >> 8))
-		encoded.WriteByte(byte(code & 0xff))
-	}
-	return encoded.Bytes()
+	return encodeString16bit(enc, raw)
 }
 
 // CharcodeToGlyph returns the glyph name matching character code `code`.
@@ -63,17 +50,11 @@ func (enc IdentityEncoder) CharcodeToGlyph(code uint16) (string, bool) {
 // GlyphToCharcode returns the character code matching glyph `glyph`.
 // The bool return flag is true if there was a match, and false otherwise.
 func (enc IdentityEncoder) GlyphToCharcode(glyph string) (uint16, bool) {
-	// String with "uniXXXX" format where XXXX is the hexcode.
-	if len(glyph) == 7 && glyph[0:3] == "uni" {
-		var unicode uint16
-		n, err := fmt.Sscanf(glyph, "uni%X", &unicode)
-		if n == 1 && err == nil {
-			return enc.RuneToCharcode(rune(unicode))
-		}
+	r, ok := enc.GlyphToRune(glyph)
+	if !ok {
+		return 0, false
 	}
-
-	common.Log.Debug("Symbol encoding error: unable to find glyph->charcode entry (%s)", glyph)
-	return 0, false
+	return enc.RuneToCharcode(r)
 }
 
 // RuneToCharcode converts rune `r` to a PDF character code.
@@ -91,7 +72,7 @@ func (enc IdentityEncoder) CharcodeToRune(code uint16) (rune, bool) {
 // RuneToGlyph returns the glyph name for rune `r`.
 // The bool return flag is true if there was a match, and false otherwise.
 func (enc IdentityEncoder) RuneToGlyph(r rune) (string, bool) {
-	if r == 0x20 {
+	if r == ' ' {
 		return "space", true
 	}
 	glyph := fmt.Sprintf("uni%.4X", r)
@@ -102,14 +83,16 @@ func (enc IdentityEncoder) RuneToGlyph(r rune) (string, bool) {
 // The bool return flag is true if there was a match, and false otherwise.
 func (enc IdentityEncoder) GlyphToRune(glyph string) (rune, bool) {
 	// String with "uniXXXX" format where XXXX is the hexcode.
-	if len(glyph) == 7 && glyph[0:3] == "uni" {
-		unicode := uint16(0)
-		n, err := fmt.Sscanf(glyph, "uni%X", &unicode)
-		if n == 1 && err == nil {
-			return rune(unicode), true
-		}
+	if glyph == "space" {
+		return ' ', true
+	} else if !strings.HasPrefix(glyph, "uni") || len(glyph) != 7 {
+		return 0, false
 	}
-	return 0, false
+	r, err := strconv.ParseUint(glyph[3:], 16, 16)
+	if err != nil {
+		return 0, false
+	}
+	return rune(r), true
 }
 
 // ToPdfObject returns a nil as it is not truly a PDF object and should not be attempted to store in file.

pdf/internal/textencoding/simple.go

@@ -117,7 +117,7 @@ func (se SimpleEncoder) String() string {
 
 // Encode converts a Go unicode string `raw` to a PDF encoded string.
 func (se SimpleEncoder) Encode(raw string) []byte {
-	return doEncode(se, raw)
+	return encodeString8bit(se, raw)
 }
 
 // CharcodeToGlyph returns the glyph name for character code `code`.

pdf/internal/textencoding/truetype.go

@@ -6,7 +6,6 @@
 package textencoding
 
 import (
-	"bytes"
 	"fmt"
 	"sort"
 	"strings"
@@ -64,20 +63,7 @@ func (enc TrueTypeFontEncoder) String() string {
 
 // Encode converts the Go unicode string `raw` to a PDF encoded string.
 func (enc TrueTypeFontEncoder) Encode(raw string) []byte {
-	// runes -> character codes -> bytes
-	var encoded bytes.Buffer
-	for _, r := range raw {
-		code, ok := enc.RuneToCharcode(r)
-		if !ok {
-			common.Log.Debug("Failed to map rune to charcode. rune=%+q", r)
-			continue
-		}
-
-		// Each entry represented by 2 bytes.
-		encoded.WriteByte(byte((code & 0xff00) >> 8))
-		encoded.WriteByte(byte(code & 0xff))
-	}
-	return encoded.Bytes()
+	return encodeString16bit(enc, raw)
 }
 
 // CharcodeToGlyph returns the glyph name matching character code `code`.