Suggest removal of object limit for get_bucket()

[stevecondylios]

To make get_bucket() more reliable in automation, I suggest removing the default limit of 1000 objects, given this is low for a function whose performance is fast, and that the function is likely to be used in automated processes which can be painful to debug when they go down. Also, if being used interactively, the user can always cancel the function at any time.

A quick test of 15k objects takes ~30 seconds. I would suggest a change to Inf would be ideal, given most users probably won't have buckets with objects in the millions, hence the risk of starting a long processes accidentally is low, especially relative to the benefits of improved reliability (since users may not know of the limit unless they've hit it before)