From b09d90c79c4821063f9151daf82c4266f73f9c9e Mon Sep 17 00:00:00 2001
From: Fabian Reinartz <fab.reinartz@gmail.com>
Date: Fri, 8 Sep 2017 15:15:12 +0200
Subject: [PATCH] Add decoding method to retrieve unsafe strings

When decoding data from mmaped blocks, we would like to retrieve
a string backed by the mmaped region. As the underlying byte slice
never changes, this is safe.
---
 encoding_helpers.go | 16 ++++++++++++++++
 index.go            | 10 +++-------
 2 files changed, 19 insertions(+), 7 deletions(-)

diff --git a/encoding_helpers.go b/encoding_helpers.go
index 9aa4ba409..17c3ff081 100644
--- a/encoding_helpers.go
+++ b/encoding_helpers.go
@@ -77,6 +77,22 @@ func (d *decbuf) uvarint32() uint32 { return uint32(d.uvarint64()) }
 func (d *decbuf) be32int() int      { return int(d.be32()) }
 func (d *decbuf) be64int64() int64  { return int64(d.be64()) }
 
+// uvarintTempStr decodes like uvarintStr but the returned string is
+// not safe to use if the underyling buffer changes.
+func (d *decbuf) uvarintTempStr() string {
+	l := d.uvarint64()
+	if d.e != nil {
+		return ""
+	}
+	if len(d.b) < int(l) {
+		d.e = errInvalidSize
+		return ""
+	}
+	s := yoloString(d.b[:l])
+	d.b = d.b[l:]
+	return s
+}
+
 func (d *decbuf) uvarintStr() string {
 	l := d.uvarint64()
 	if d.e != nil {
diff --git a/index.go b/index.go
index ddc2c4f52..e65f3fec1 100644
--- a/index.go
+++ b/index.go
@@ -335,10 +335,6 @@ func (w *indexWriter) AddSymbols(sym map[string]struct{}) error {
 
 	for _, s := range symbols {
 		w.symbols[s] = uint32(w.pos) + headerSize + uint32(w.buf2.len())
-
-		// NOTE: len(s) gives the number of runes, not the number of bytes.
-		// Therefore the read-back length for strings with unicode characters will
-		// be off when not using putUvarintStr.
 		w.buf2.putUvarintStr(s)
 	}
 
@@ -636,7 +632,7 @@ func (r *indexReader) readOffsetTable(off uint64) (map[string]uint32, error) {
 		keys := make([]string, 0, keyCount)
 
 		for i := 0; i < keyCount; i++ {
-			keys = append(keys, d2.uvarintStr())
+			keys = append(keys, d2.uvarintTempStr())
 		}
 		res[strings.Join(keys, sep)] = uint32(d2.uvarint())
 
@@ -673,7 +669,7 @@ func (r *indexReader) section(o uint32) (byte, []byte, error) {
 func (r *indexReader) lookupSymbol(o uint32) (string, error) {
 	d := r.decbufAt(int(o))
 
-	s := d.uvarintStr()
+	s := d.uvarintTempStr()
 	if d.err() != nil {
 		return "", errors.Wrapf(d.err(), "read symbol at %d", o)
 	}
@@ -688,7 +684,7 @@ func (r *indexReader) Symbols() (map[string]struct{}, error) {
 	sym := make(map[string]struct{}, count)
 
 	for ; count > 0; count-- {
-		s := d2.uvarintStr()
+		s := d2.uvarintTempStr()
 		sym[s] = struct{}{}
 	}