Frens Jan Rumph

@methane perhaps more importantly, the main reason why I filed this issue is _not_ because of performance in terms of throughput, but memory footprint. The #395 implementation uses `O(2n)` memory...

Finally, I'm also curious about the memory behaviour of `del self._buffer[:n]` which both the implementation above as #395 use. I've witnessed `MemoryError`s on this, so I suspect it to copy...

@methane it would remove a copy, so that would be great! The code in https://github.com/python/asyncio/blob/master/asyncio/selector_events.py#L666 doesn't look that difficult, but requires the protocol to expose a buffer or a method...

(afterburner: what a horrific idea that data is copied 3+ times before being returned from the current readexactly)

@socketpair @methane @1st1 A first draft for `readinto` and `readinto_exactly`: https://github.com/TargetHolding/asyncio/commit/bc13eb7ba8b5ab1be8f849f0809ff670dd09b7dd Performance is a bit better than the existing `b''.join(blocks)` based implementation and a bit worse from the suggested implementation...

@methane yes, I shouldn't have used Big-O here; both implementations have a maximum memory footprint which is linear to the message size. @methane indeed, this is an issue for bigger...

`\b` would be better here indeed.

It's on my 'long-list' of interesting features, but I have some other fish to fry at the moment.

I've fried some other fish 😂 We've been implementing a SPARQL layer over our existing data-platform (with statement like data in HBase and entity-centric documents in Elasticsearch). I already had...

@JervenBolleman, could you help me with a more thorough understanding of [SEP-0007](https://github.com/w3c/sparql-dev/blob/main/SEP/SEP-0007/sep-0007.md) and how it relates to the introduction of `LATERAL` in RDF4J? I get the basic gist of [inject](https://github.com/w3c/sparql-dev/blob/main/SEP/SEP-0007/sep-0007.md#an-improved-substitute-operation)...