fix: non-flexible ElectLeadersRequest V0/V1 encode/decode

[hindessm]

And clean up the misleading layout/comment of the []bytes in the test:

Test layout for V2 []byte was:

	0, 0, // empty tagged fields
	0, 39, 16, 0, // timeout 10000

but, while the bytes were correct the comments were misleading as the encoding is really:

	0,            // empty tagged fields
	0, 0, 39, 16, // timeout 10000
	0, // empty tagged fields

[dnwe]

LGTM, good catch

[dnwe]

Aside: we probably should have a functional test for elect leaders as I feel like this has likely never properly worked?