Put them in a separate '_generated_v2' subdir, to ease migration.

Attempt to factor generation process for clarity (toward #1482).

It was never actually implemented on the back-end in V1, and has been
dropped altogether in V2.

Addresses:
#1908 (comment)

Generate code for Bigtable v2 protobufs

Processes the 'ReadRowsResponse.CellChunk' state  machine.

Add tests based on JSON acceptance tests:
https://github.com/GoogleCloudPlatform/cloud-bigtable-client/blob/master/bigtable-client-core/src/test/resources/com/google/cloud/bigtable/grpc/scanner/v2/read-rows-acceptance-test.json.

Drop 'Table.rename'.

Indicate their V1 source in their names.

Prepratory to converting to V2 equivalents.

…logic

Add 'ReadRowsResponseV2' wrapper

…ints

Alias Bigtable V1 imports / factories / entry point constants.

Note that {Create,Update,Delete}ColumnFamily messages all collapse to
ModifyColumnFamilies.

Update bigtable.column_family to use V2 protos

Assert that the state is 'ROW_IN_PROGRESS', and check that the completed
rows match the expected results.

Verify that completed, non-error rows match expected results after an
invalid chunk testcase.

…ror-cases

Clarify handling of V2 'ReadRowsResponse' error cases

Folds new ReadRowsResponse logic from #1907, #1915 into table row
handling.

Convert non-instance-admin protos to Bigtable V2

In V2, those operations are on the instance.

Support V2 instance admin

…tos.

Closes #1928.

This is a better hack for #1482, but we still really want #1384.

…1 protos.

This is a better hack for #1482, but we still really want #1384.