Move string-encoding to an internal/jsonstring package

1 files changed, 128 insertions, 0 deletions

diff --git a/internal/jsonstring/encode_string.go b/internal/jsonstring/encode_string.go
new file mode 100644
index 0000000..f29dc3f
--- /dev/null
+++ b/internal/jsonstring/encode_string.go
@@ -0,0 +1,128 @@
+// Copyright (C) 2022-2023  Luke Shumaker <lukeshu@lukeshu.com>
+//
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+package jsonstring
+
+import (
+	"fmt"
+	"io"
+	"unicode/utf8"
+
+	"git.lukeshu.com/go/lowmemjson/internal/fastio"
+	"git.lukeshu.com/go/lowmemjson/internal/jsonparse"
+)
+
+// BackslashEscapeMode is describe in the main lowmemjson package
+// docs.
+type BackslashEscapeMode uint8
+
+const (
+	BackslashEscapeNone BackslashEscapeMode = iota
+	BackslashEscapeShort
+	BackslashEscapeUnicode
+)
+
+// BackslashEscaper is describe in the main lowmemjson package docs.
+type BackslashEscaper = func(rune, BackslashEscapeMode) BackslashEscapeMode
+
+func writeStringUnicodeEscape(w io.Writer, c rune) (int, error) {
+	buf := [6]byte{
+		'\\',
+		'u',
+		jsonparse.Hex[(c>>12)&0xf],
+		jsonparse.Hex[(c>>8)&0xf],
+		jsonparse.Hex[(c>>4)&0xf],
+		jsonparse.Hex[(c>>0)&0xf],
+	}
+	return w.Write(buf[:])
+}
+
+func writeStringShortEscape(w io.Writer, c rune) (int, error) {
+	var b byte
+	switch c {
+	case '"', '\\', '/':
+		b = byte(c)
+	case '\b':
+		b = 'b'
+	case '\f':
+		b = 'f'
+	case '\n':
+		b = 'n'
+	case '\r':
+		b = 'r'
+	case '\t':
+		b = 't'
+	default:
+		panic(fmt.Errorf("should not happen: writeStringShortEscape called with invalid rune: %q", c))
+	}
+	buf := [2]byte{'\\', b}
+	return w.Write(buf[:])
+}
+
+func WriteStringChar(w fastio.AllWriter, c rune, escape BackslashEscapeMode) (int, error) {
+	switch escape {
+	case BackslashEscapeNone:
+		switch {
+		case c < 0x0020: // override, gotta escape these
+			switch c {
+			case '\b', '\f', '\n', '\r', '\t': // short-escape if possible
+				return writeStringShortEscape(w, c)
+			default:
+				return writeStringUnicodeEscape(w, c)
+			}
+		case c == '"' || c == '\\': // override, gotta escape these
+			return writeStringShortEscape(w, c)
+		default: // obey
+			return w.WriteRune(c)
+		}
+	case BackslashEscapeShort:
+		switch c {
+		case '"', '\\', '/', '\b', '\f', '\n', '\r', '\t': // obey
+			return writeStringShortEscape(w, c)
+		default: // override, can't short-escape these
+			return w.WriteRune(c)
+		}
+	case BackslashEscapeUnicode:
+		switch {
+		case c > 0xFFFF: // override, can't escape these (TODO: unless we use UTF-16 surrogates?)
+			return w.WriteRune(c)
+		default: // obey
+			return writeStringUnicodeEscape(w, c)
+		}
+	default:
+		panic("escaper returned an invalid escape mode")
+	}
+}
+
+func EncodeStringFromString(w fastio.AllWriter, escaper BackslashEscaper, str string) error {
+	if err := w.WriteByte('"'); err != nil {
+		return err
+	}
+	for _, c := range str {
+		if _, err := WriteStringChar(w, c, escaper(c, BackslashEscapeNone)); err != nil {
+			return err
+		}
+	}
+	if err := w.WriteByte('"'); err != nil {
+		return err
+	}
+	return nil
+}
+
+func EncodeStringFromBytes(w fastio.AllWriter, escaper BackslashEscaper, str []byte) error {
+	if err := w.WriteByte('"'); err != nil {
+		return err
+	}
+	for i := 0; i < len(str); {
+		c, size := utf8.DecodeRune(str[i:])
+		if _, err := WriteStringChar(w, c, escaper(c, BackslashEscapeNone)); err != nil {
+			return err
+		}
+		i += size
+	}
+	if err := w.WriteByte('"'); err != nil {
+		return err
+	}
+	return nil
+}