Module linera_views::test_utils

Helper types for tests.

Modules

• performance
  Functions for computing the performance of stores.
• test_views
  Some Views that are easy to use with test cases.

Functions

• add_prefix
  Adds a prefix to a list of key-values
• big_read_multi_values
  Reading many keys at a time could trigger an error. This needs to be tested.
• get_random_byte_vector
  Takes a random number generator, a key_prefix and extends it by n random bytes.
• get_random_key_prefix
  Returns a random key prefix used for tests
• get_random_key_value_operations
  A bunch of puts and some deletes.
• get_random_key_values
  Takes a random number generator rng, a number n and returns n random (key, value) which are all distinct with key and value being of length 8.
• get_random_key_values2
  Generates a list of random key-values with no duplicates
• get_random_key_values_prefix
  Takes a random number generator, a key_prefix and generates pairs (key, value) with key obtained by appending 8 bytes at random to key_prefix and value obtained by appending 8 bytes to the trivial vector. We return n such (key, value) pairs which are all distinct
• get_random_kset
  Builds a random k element subset of n
• get_random_test_scenarios
  We build a number of scenarios for testing the reads.
• get_small_key_space
  Appends a small value to a key making collisions likely.
• namespace_admin_test
  Exercises the namespace functionalities of the AdminKeyValueStore. This tests everything except the delete_all which would interact with other namespaces.
• root_key_admin_test
  Tests listing the root keys.
• run_big_write_read
  DynamoDB has limits at 1 MB (for pagination), 4 MB (for write) Let us go right past them at 20 MB of data with writing and then reading it. And 20 MB is not huge by any mean. All KeyValueStore must handle that.
• run_reads
  This test starts with a collection of key/values being inserted into the code which is then followed by a number of reading tests. The functionalities being tested are all the reading functionalities:
• run_test_batch_from_blank
  Writes and then reads data under a prefix, and verifies the result.
• run_writes_from_blank
  Run many operations on batches always starting from a blank state.
• run_writes_from_state
  Run some deterministic and random batches operation and check their correctness
• span_random_reordering_put_delete
  A random reordering of the puts and deletes. For something like MapView it should get us the same result whatever way we are calling.
• tombstone_triggering_test
  That test is especially challenging for ScyllaDB. In its default settings, Scylla has a limitation to 10000 tombstones. A tombstone is an indication that the data has been deleted. That is thus a trie data structure for checking whether a requested key is deleted or not.