fix(ShardManager): revise shard initial connection flow

[Snazzah]

Some issues with the shard manager:

• The client was checking for a shardConcurrency option for some reason under connect()
• Shards wait until its ready then wait 5s before connecting another shard rather than waiting 5s after sending an identify packet
• The default create timeout being 2 seconds is, ehh, slow? There usually isnt such a large delay between these events
• Waiting for shards to be ready is bound by the client's unavailable guilds rather than it's own list of guilds to wait for

So this PR changes some things:

• Reworks the connection strategy, it will wait 5s after sending the identify packet before connecting another shard in it's bucket, the connectTimeout waits until the next best moment to check over shards again. Also it was just iterating the queue while it could be altered
• Shards now hold their own "awaited guilds" set for those IDs to be expected to come in before it is considered as ready
• guildCreateTimeout: 2 seconds -> 500 ms, this is subjective so feel free to change this if you feel that it's too small