Panic when sending batch of raw queries

[jdiamond]

I now know batches of raw queries are not supported yet. The application I'm working on is using them and it's making the proxy panic.

When we send a batch of raw queries, we see an error like this for every statement in the batch:

{"level":"error","time":"2024-10-22T18:06:56.787Z","caller":"proxy/proxy.go:1028","msg":"Item is not of type [16]byte","stacktrace":"github.com/ollionorg/cassandra-to-spanner-proxy/third_party/datastax/proxy.(*client).handleBatch\n\t/go/src/cassandra-to-spanner-proxy/third_party/datastax/proxy/proxy.go:1028\ngithub.com/ollionorg/cassandra-to-spanner-proxy/third_party/datastax/proxy.(*client).Receive\n\t/go/src/cassandra-to-spanner-proxy/third_party/datastax/proxy/proxy.go:643\ngithub.com/ollionorg/cassandra-to-spanner-proxy/third_party/datastax/proxycore.(*Conn).read\n\t/go/src/cassandra-to-spanner-proxy/third_party/datastax/proxycore/conn.go:114"}

Then we see a panic like this and the proxy has to be restarted:

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x48 pc=0x11a6fe6]

goroutine 116 [running]:
github.com/ollionorg/cassandra-to-spanner-proxy/spanner.(*SpannerClient).filterBatch(0x1889698?, {0x185c898, 0xc0006fa960}, 0xc000631e40?, {0xc0006f7860, 0xc, 0x0?})
	/go/src/cassandra-to-spanner-proxy/spanner/spanner.go:981 +0x146
github.com/ollionorg/cassandra-to-spanner-proxy/spanner.(*SpannerClient).FilterAndExecuteBatch.func1({0x185c898, 0xc0006fa960}, 0x13a12a0?)
	/go/src/cassandra-to-spanner-proxy/spanner/spanner.go:713 +0x5e
cloud.google.com/go/spanner.(*ReadWriteTransaction).runInTransaction(0xc0005f26e0, {0x185c898, 0xc0006fa450}, 0xc000573838)
	/go/pkg/mod/cloud.google.com/go/[email protected]/transaction.go:1708 +0x69
cloud.google.com/go/spanner.(*Client).rwTransaction.func2({0x185c898, 0xc0006fa450})
	/go/pkg/mod/cloud.google.com/go/[email protected]/client.go:860 +0x817
cloud.google.com/go/spanner.runWithRetryOnAbortedOrFailedInlineBeginOrSessionNotFound.func1({0x185c898, 0xc0006fa450})
	/go/pkg/mod/cloud.google.com/go/[email protected]/retry.go:91 +0x64
cloud.google.com/go/spanner.runWithRetryOnAbortedOrFailedInlineBeginOrSessionNotFound({0x185c898, 0xc0006fa450}, 0xc0002a3698)
	/go/pkg/mod/cloud.google.com/go/[email protected]/retry.go:131 +0x1ba
cloud.google.com/go/spanner.(*Client).rwTransaction(0xc0005e4c80, {0x185c898, 0xc0006fa450}, 0xc0002a3838, {{0x0, 0xc0003772a0}, {0x0, 0x0}, 0x0, 0x0, ...})
	/go/pkg/mod/cloud.google.com/go/[email protected]/client.go:816 +0x1bd
cloud.google.com/go/spanner.(*Client).ReadWriteTransactionWithOptions(0x60?, {0x185c8d0?, 0xc0004052c0?}, 0xc0002a3860?, {{0x0, 0xc0003772a0}, {0x0, 0x0}, 0x0, 0x0, ...})
	/go/pkg/mod/cloud.google.com/go/[email protected]/client.go:798 +0x10c
github.com/ollionorg/cassandra-to-spanner-proxy/spanner.(*SpannerClient).FilterAndExecuteBatch(0xc0003b7d70, {0x185c8d0, 0xc0004052c0}, {0xc0006f7860, 0xc, 0xc})
	/go/src/cassandra-to-spanner-proxy/spanner/spanner.go:712 +0x10f
github.com/ollionorg/cassandra-to-spanner-proxy/third_party/datastax/proxy.(*client).handleBatch(0xc0006f8f00, 0xc000448b40, 0xc0006fa3c0)
	/go/src/cassandra-to-spanner-proxy/third_party/datastax/proxy/proxy.go:1088 +0x13a8
github.com/ollionorg/cassandra-to-spanner-proxy/third_party/datastax/proxy.(*client).Receive(0xc0006f8f00, {0x183f260?, 0xc0006f7680?})
	/go/src/cassandra-to-spanner-proxy/third_party/datastax/proxy/proxy.go:643 +0x625
github.com/ollionorg/cassandra-to-spanner-proxy/third_party/datastax/proxycore.(*Conn).read(0xc0006f7740)
	/go/src/cassandra-to-spanner-proxy/third_party/datastax/proxycore/conn.go:114 +0x33
created by github.com/ollionorg/cassandra-to-spanner-proxy/third_party/datastax/proxycore.(*Conn).Start in goroutine 80
	/go/src/cassandra-to-spanner-proxy/third_party/datastax/proxycore/conn.go:107 +0x56

I get that we are trying to do something that is documented as not being supported, but I think the DX could be improved with a helpful error message instead of just panicking. Took us over an hour of debugging to figure out what was happening.