Adds a test to verify that async_receive (v1) is cancelled on reconnection
Adds a test to verify that a reconnection is triggered when the receive adapter generates an error
Makes the unsubscribe and push adapter tests live only in test_conn_push2, since they test functionality common to async_receive and async_receive2
Migrates all tests to lightweight_test
Entails no functional change
Adds request::{subscribe, unsubscribe, psubscribe, punsubscribe}. When requests created with these functions are executed successfully, the created subscriptions are tracked and restore on re-connection.
close#367
Adds the concept of a "temporary working area" to flat_tree. Nodes in this area belong to a partially parsed message, and are hidden from the user. Now flat_tree can be used as the receive response without explicitly handling partial messages.
Changes flat_tree::get_view return type from const vector& to span.
Adds flat_tree::capacity.
Splits generic_flat_response tests to a separate file and adds extra cases.
close#369
* Fixes a problem that caused passing ranges containing tuples into `request::push_range` to generate invalid commands.
* Adds test_serialization
* Updates request reference docs to reflect the requirements of the types passed to push and push_range
close#360
* Makes flat_tree implementation use a custom buffer. This allows:
* Never dangling nodes (previously, node values could dangle after calling reserve() or if notify_done() wasn't called).
* Reduced memory consumption
* Increased runtime speed
* Changes flat_tree assignment to the usual signature and semantics
* Fixes a bug causing an assertion to trigger when copy-constructing an empty flat_tree.
* Changes basic_node operator== and operator!= return type
* Adds generic_flat_response, basic_tree, tree, view_tree, flat_tree to the reference page.
* Adds a missing resp3:: qualifier to all names in the reference page that belong to the resp3 namespace.
* Adds reference documentation to flat_tree.
* Mentions generic_flat_response in the discussion.
* Adds operator!= for basic_node to basic_node's reference page.
* Adds test_flat_tree.
close#357close#354close#352
Changes request default constructor to set cancel_on_connection_lost to false, matching request::config's default initializer
Overwrites this flag to true for the setup request
Adds unit tests for the latter
* Removes logging all the reader actions, and logs specific messages inside the reader_fsm instead
* Adds constructors to reader actions
* Makes reader_fsm use connection_state
* Refactors reader_fsm tests
* Moves exec_fsm action printing to test code
* Implements async_run as a FSM and adds tests
* Places all sans-io variables in connection_impl in a connection_state struct
Entails no functional change.
* Changes cancel_on_connection_lost default to false
* Deprecates cancel_on_connection_lost and cancel_if_not_connected
* Fixes a TODO in test_conn_exec_cancel
close#323
* Refactors the writer task into a FSM and adds unit tests.
* Adds a testing utility to check logging.
Entails no functional change (other than cosmetic word fixes to the logs).
* Adds support for asio::cancel_after in connection::{async_exec, async_run}
* Adds cancel_after tests
* Adds an example on using asio::cancel_after
* Adds a discussion page on timeouts and the `cancel_if_unresponded` flag
close#226
* Adds support for terminal and partial cancellation to async_run.
* Makes basic_connection::cancel use per-operation cancellation under the hood.
* Fixes a number of race conditions during cancellation which could cause the cancellation to be ignored. This could happen if the cancellation is delivered while an async handler is pending execution.
* Deprecates operation::{resolve, connect, ssl_handshake, reconnection, health_check}, in favor of operation::run. Calling basic_connection::cancel with these values (excepting reconnection) is now equivalent to using operation::run.
* Fixes a problem in the health checker that caused ping timeouts to be reported as cancellations.
* Sanitizes how the parallel group computes its final error code.
* Simplifies the reader, writer and health checker to not care about connection cancellation. This is now the responsibility of the parallel group.
* Removes an unnecessary setup_cancellation action in the reader FSM.
* Adds documentation regarding per-operation cancellation to async_receive.
* Adds additional health checker tests.
* Adds async_run per-operation cancellation tests.
* Adds reader FSM cancellation tests.
* Makes test_conn_exec_retry tests more resilient.
* Removes leftovers in the UNIX and TLS reconnection tests. These were required due to race conditions that have already been fixed.
close#318close#319
Data rotation in the read buffer creates latency, we know it
is preset but so far its magnitude was unknown. This PR adds
it as a new field to the usage struct. For example, the
test_conn_echo_stress outputs now
Commands sent: 780,002
Bytes sent: 32,670,085
Responses received: 780,001
Pushes received: 750,001
Bytes received (response): 3,210,147
Bytes received (push): 32,250,036
Bytes rotated: 3,109,190,184
In total approximately 34Mb are received but 3Gb are
rotated.
Splits test_conn_health_check into a test to verify that the health checker works and another to verify that MONITOR is handled correctly.
Tests no longer wait or use CLIENT PAUSE
