Fix errors with reload lua scripts

[novln]

This refactoring started when I had an issue testing if the reloading was safe for concurrent use, using SCRIPT FLUSH command on a Redis instance.

Reusing the benchmark provided and adding the flag -race with multiple goroutines, I've detected that there was several issues.

The first one was with the second execution of EvalSha, the args... were missing, producing an error on the Lua increment script because there was no argument given.

limiter/drivers/store/redis/store.go

Line 179 in aa44c77

cmd = store.client.EvalSha(ctx, sha, keys)

The second one was with an data race detected because multiple goroutines could read (without protection) when one could write to modify store.luaIncrSHA. Better safe than sorry, I've added a method to ensure that there is no data race, even if the impact is pretty minimal.

limiter/drivers/store/redis/store.go

Line 114 in aa44c77

store.luaIncrSHA = incrLuaSHA

limiter/drivers/store/redis/store.go

Line 123 in aa44c77

cmd := store.evalSHA(ctx, store.luaIncrSHA, []string{key}, 1, rate.Period.Milliseconds())

One thing led to another and I was down the rabbit hole so, I've done some refactoring.

@git-hulk Feel free to take a look.

[git-hulk]

it looks good to me