Build the mbuf directly instead of copying in web_proxy
No sense doing a malloc followed by a memcpy for each packet. Instead,
we know how big the message will be, so allocate that much memory and
build directly into it.
Change-Id: Ie2361ced5403ff481dd10b30e8d9f106071c506b
diff --git a/aos/network/web_proxy_utils.h b/aos/network/web_proxy_utils.h
index cfed19d..19b496d 100644
--- a/aos/network/web_proxy_utils.h
+++ b/aos/network/web_proxy_utils.h
@@ -14,6 +14,10 @@
flatbuffers::FlatBufferBuilder *fbb, const Context &context,
int channel_index, int packet_index);
+// Returns the size that the overall packed message packed by PackMessage will
+// be for the provided packet index.
+size_t PackedMessageSize(const Context &context, int packet_index);
+
// Packs the provided raw data into a series of MessageHeader's of the
// appropriate length.
std::vector<FlatbufferDetachedBuffer<MessageHeader>> PackBuffer(