Also I just wanted to point out that there is a cassandra-drivers slack room on the ASF slack where folks that work on the different drivers interact.
(added to both user and dev list thread) > On Aug 14, 2024, at 5:59 PM, Dinesh Joshi <djo...@apache.org> wrote: > > Hi Vincent, > > This is the Cassandra user's mailing list. You can engage the Cassandra > developers on d...@cassandra.apache.org <mailto:d...@cassandra.apache.org> > mailing list. I have bcc'd the user list and added the dev list. > > To answer your question, the spec may have drifted and may require a fix. > Please feel free to raise a jira and contribute a patch to the documentation. > > On a side note, why are you implementing your own driver? > > Thanks, > > Dinesh > > On Sun, Aug 11, 2024 at 8:29 AM Vincent Rischmann <vinc...@rischmann.fr > <mailto:vinc...@rischmann.fr>> wrote: >> Hello, >> >> this may not be the best place to ask this, feel free to redirect me. >> >> I'm working on writing a Cassandra client in my spare time and am currently >> implementing the framing that has been added in protocol v5. >> >> I followed the spec available here: >> https://github.com/apache/cassandra/blob/trunk/doc/native_protocol_v5.spec#L97 >> but I hit an issue regarding the CRC32, when I tried to decode a frame >> generated by cqlsh (which I captured using wireshark) I couldn't get the >> right checksum. >> >> After debugging for a while I realized that the CRC32 hash is always >> initialized with 4 "magic" bytes: >> https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/net/Crc.java#L38-L54 >> >> Shouldn't this be added to the specification ?
