optimizte reading alligned unsized strings

Cargo.toml

@@ -13,6 +13,7 @@ travis-ci = { repository = "icewind1991/bitstream_reader" }
 [dependencies]
 num-traits = "0.2"
 bitstream_reader_derive = { version = "0.6", path = "bitstream_reader_derive" }
+bytecount = "0.5"
 
 [dev-dependencies]
 maplit = "1.0.1"

benches/bench.rs

@@ -184,7 +184,7 @@ fn perf_bytes_be(b: &mut Bencher) {
 
 #[bench]
 fn perf_bytes_le(b: &mut Bencher) {
-    let buffer = BitBuffer::new(get_string_buffer(), BigEndian);
+    let buffer = BitBuffer::new(get_string_buffer(), LittleEndian);
 
     b.iter(|| {
         let mut pos = 0;
@@ -205,7 +205,7 @@ fn perf_bytes_be_unaligned(b: &mut Bencher) {
     let buffer = BitBuffer::new(get_string_buffer(), BigEndian);
 
     b.iter(|| {
-        let mut pos = 0;
+        let mut pos = 3;
         let len = buffer.bit_len();
         loop {
             if pos + (128 * 8) > len {
@@ -220,7 +220,7 @@ fn perf_bytes_be_unaligned(b: &mut Bencher) {
 
 #[bench]
 fn perf_bytes_le_unaligned(b: &mut Bencher) {
-    let buffer = BitBuffer::new(get_string_buffer(), BigEndian);
+    let buffer = BitBuffer::new(get_string_buffer(), LittleEndian);
 
     b.iter(|| {
         let mut pos = 3;

src/buffer.rs

@@ -435,7 +435,23 @@ where
         }
     }
 
+    #[inline]
+    fn find_null_byte(&self, byte_index: usize) -> usize {
+        self.bytes
+            .iter()
+            .enumerate()
+            .skip(byte_index)
+            .find(|(_, byte)| **byte == 0)
+            .map(|(i, _)| i)
+            .unwrap_or(self.byte_len())
+    }
+
+    #[inline]
     fn read_string_bytes(&self, position: usize) -> Result<Vec<u8>> {
+        if false && position & 7 == 0 {
+            let byte_index = position / 8;
+            Ok(self.bytes[byte_index..self.find_null_byte(byte_index)].to_vec())
+        } else {
             let mut acc = Vec::with_capacity(32);
             let mut pos = position;
             loop {
@@ -470,6 +486,7 @@ where
                 pos += read;
             }
         }
+    }
 
     /// Read a sequence of bits from the buffer as float
     ///