protocol utilities

Author: Snawoot

proto/escape.go

@@ -0,0 +1,50 @@
+package proto
+
+func fromhex(ch byte) int8 {
+	if ch >= '0' && ch <= '9' {
+		return int8(ch - '0')
+	}
+	if ch >= 'a' && ch <= 'f' {
+		return int8(ch - 'a' + 10)
+	}
+	if ch >= 'A' && ch <= 'F' {
+		return int8(ch - 'A' + 10)
+	}
+	return -1
+}
+
+// RFC1738Unescape is a port of Squid rfc1738_unescape function.
+// It does unescaping in-place and returns slice pointing to boundaries of
+// unescaped bytes.
+func RFC1738Unescape(s []byte) []byte {
+	var i, j int /* i is write, j is read */
+	for ; j < len(s); i, j = i+1, j+1 {
+		s[i] = s[j]
+		if s[j] != '%' {
+			/* normal case, nothing more to do */
+		} else if j+1 < len(s) && s[j+1] == '%' { /* %% case */
+			j++ /* Skip % */
+		} else {
+			if j+2 >= len(s) {
+				continue
+			}
+			/* decode */
+			var v1, v2 int8
+			var x byte
+			v1 = fromhex(s[j+1])
+			if v1 < 0 {
+				continue /* non-hex */
+			}
+			v2 = fromhex(s[j+2])
+			if v2 < 0 {
+				continue /* non-hex */
+			}
+			x = byte(v1)<<4 | byte(v2)
+			if x > 0 && x <= 255 {
+				s[i] = x
+				j += 2
+			}
+		}
+	}
+	return s[:i]
+}

proto/escape_test.go

@@ -0,0 +1,75 @@
+package proto
+
+import (
+	"bytes"
+	"fmt"
+	"testing"
+)
+
+func TestRFC1738Unescape(t *testing.T) {
+	testcases := []struct {
+		input  string
+		output string
+	}{
+		{
+			"%2Fdata%2Fsource%2Fpath",
+			"/data/source/path",
+		},
+		{
+			"http://foo.invalid%2Fdata%2Fsource%2Fpath",
+			"http://foo.invalid/data/source/path",
+		},
+		{
+			"w%0Ard",
+			"w\nrd",
+		},
+		{
+			"w%rd",
+			"w%rd",
+		},
+		{
+			"w%%rd",
+			"w%rd",
+		},
+		{
+			"w%%%rd",
+			"w%%rd",
+		},
+		{
+			"Bad String %1",
+			"Bad String %1",
+		},
+		{
+			"Bad String %1A%3",
+			"Bad String \032%3",
+		},
+		{
+			"Good String %1A",
+			"Good String \032",
+		},
+		{
+			"w%00rd",
+			"w%00rd",
+		},
+		{
+			"w%0rd",
+			"w%0rd",
+		},
+		{
+			"w%%00%rd",
+			"w%00%rd",
+		},
+		{
+			"w%%%00%rd",
+			"w%%00%rd",
+		},
+	}
+	for i, tc := range testcases {
+		t.Run(fmt.Sprintf("Testcase#%d", i+1), func(t *testing.T) {
+			output := RFC1738Unescape([]byte(tc.input))
+			if bytes.Compare(output, []byte(tc.output)) != 0 {
+				t.Errorf("got: %x, want: %x", output, []byte(tc.output))
+			}
+		})
+	}
+}