read_rows shouldn't extend the user's deadline

[igorbernstein2]

python-bigtable/google/cloud/bigtable/row_data.py

Lines 407 to 413 in a535f99

	# The `timeout` parameter must be somewhat greater than the value
	# contained in `self.retry`, in order to avoid race-like condition and
	# allow registering the first deadline error before invoking the retry.
	# Otherwise there is a risk of entering an infinite loop that resets
	# the timeout counter just before it being triggered. The increment
	# by 1 second here is customary but should not be much less than that.
	self.response_iterator = read_method(request, timeout=self.retry._deadline + 1)

Artificially adds 1 sec to a read rows deadline. This is a bit surprising. We should investigate a way to avoid doing this (while also avoiding the infinite loop the comment mentions)

[meredithslota]

This was added here: #16 and has moved upwards in the file to

python-bigtable/google/cloud/bigtable/row_data.py

Line 154 in a4e7325

# The `timeout` parameter must be somewhat greater than the value

but has the same rationale. I'm classifying this as a type: cleanup instead of a bug unless you feel strongly this is a bug instead of a potentially-now-obsolete feature. 😄