Performance improvement: replace memcpy with loop.

In the pb_istream_from_buffer and pb_ostream_from_buffer, memcpy was
used to transfer values to the buffer. For the common case of
count = 1-10 bytes, a simple loop is faster.

2 files changed, 9 insertions, 3 deletions

diff --git a/pb_decode.c b/pb_decode.c
index 7c68b5a2..6ddde773 100644
--- a/pb_decode.c
+++ b/pb_decode.c
@@ -39,11 +39,14 @@ static const pb_decoder_t PB_DECODERS[PB_LTYPES_COUNT] = {
 static bool checkreturn buf_read(pb_istream_t *stream, uint8_t *buf, size_t count)
 {
     uint8_t *source = (uint8_t*)stream->state;
+    stream->state = source + count;
     
     if (buf != NULL)
-        memcpy(buf, source, count);
+    {
+        while (count--)
+            *buf++ = *source++;
+    }
     
-    stream->state = source + count;
     return true;
 }
 
diff --git a/pb_encode.c b/pb_encode.c
index 5d1965d1..95223cb6 100644
--- a/pb_encode.c
+++ b/pb_encode.c
@@ -37,8 +37,11 @@ static const pb_encoder_t PB_ENCODERS[PB_LTYPES_COUNT] = {
 static bool checkreturn buf_write(pb_ostream_t *stream, const uint8_t *buf, size_t count)
 {
     uint8_t *dest = (uint8_t*)stream->state;
-    memcpy(dest, buf, count);
     stream->state = dest + count;
+    
+    while (count--)
+        *dest++ = *buf++;
+    
     return true;
 }