bug: fix SIGSEGV during shutdown with --nocache (low-priority)

[LarryRuane]

Low-priority bugfix, follow-on to #483 - prevent segmentation fault when lightwalletd exits (for example, receives a control-C):

{"app":"lightwalletd","level":"info","msg":"Starting gRPC server on 127.0.0.1:9067","time":"2025-08-13T14:19:45-06:00"}
^Cpanic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x20 pc=0xa92db3]

goroutine 12 [running]:
github.com/zcash/lightwalletd/common.(*BlockCache).Sync(0x0)
	/sd/g/lightwalletd/common/cache.go:384 +0x13
github.com/zcash/lightwalletd/cmd.startServer.func1()
	/sd/g/lightwalletd/cmd/root.go:305 +0x69
created by github.com/zcash/lightwalletd/cmd.startServer in goroutine 1
	/sd/g/lightwalletd/cmd/root.go:303 +0x1da9
exit status 2

The bug is that if lightwalletd is started with the --nocache configuration, the cache variable isn't set (is nil), but we're not testing it before dereferencing it during shutdown (after receiving a signal).

I verified that this variable, cache, cannot be changing while it's being tested (for nil value), because it's set (or not) earlier in this same function (before the asynchronous goroutine is launched). Although, that would have been a bug previously too.

[nuttycom]

utACK; this looks trivially safe.