In some cases ByteBuffers do not provide access to an underlying
byte array without incurring copies, this is notably the case
for off-heap direct bytebuffers.
This patch provides a way to call update on streaming XXHASH instances
against a ByteBuffer, which avoids the extra copy needed into a byte
array.
I'm putting this one up to gauge interest. If present, I can adapt the Java safe and unsafe variants to also support the method, I first wanted some feedback on the general idea.
I'm also a bit unclear on what to do should we reach a case where the buffer supports neither .isDirect nor .hasArray, in this case it feels as though the only way forward is to allocate a byte array to copy the ByteBuffer contents into
Hi, a quick update on this one, it has been running fine for over a year in production. Is there any interest in seeing this get in or should I close the PR?
In some cases ByteBuffers do not provide access to an underlying byte array without incurring copies, this is notably the case for off-heap direct bytebuffers.
This patch provides a way to call update on streaming XXHASH instances against a ByteBuffer, which avoids the extra copy needed into a byte array.
I'm putting this one up to gauge interest. If present, I can adapt the Java safe and unsafe variants to also support the method, I first wanted some feedback on the general idea.
I'm also a bit unclear on what to do should we reach a case where the buffer supports neither
.isDirect
nor.hasArray
, in this case it feels as though the only way forward is to allocate a byte array to copy the ByteBuffer contents into