851c909cd3
* Fix #8460 [effect] apply_to=variation Previously the code could apply the variation effects last, so that codes like ``` [effect] apply_to=variation .. [/effect] [effect] apply_to=hitpoints heal_full=yes [/effect] ``` Would not set the unit hitpoints to the new variations hitpoints because the variation effect was applied after the healing effect. In 1.16 this worked because healing was applied a little too often but that lead also to bugs like #8342 * f prev * f prev * f prev * f prev * f prev * Create modification_effect_type_variation.cfg * Update wml_test_schedule |
||
|---|---|---|
| .. | ||
| behavioral_tests | ||
| cve_tests | ||
| lua_tests | ||
| macro_tests | ||
| manual_tests | ||
| test_tests | ||
| wml_tests | ||
| README.md | ||
Test scenarios
This directory contains both the scenarios used by C++ unit tests and those which are WML unit tests.
C++ unit tests
For the C++ unit tests, it is recommended to reuse the same scenario file as much as possible and just inject WML into it.
Injection can be done by adding a config object containing event code and then registering that
manually for game_events.
Manual tests
The manual_tests subdirectory contains scenarios that expect to be run interactively, either by
binding a hotkey for the main menu's "Choose Test Scenario" option, or with the command line
argument -t <testname>.
Many of these are closer to workbenches than tests, allowing developers to do some action that isn't automated, and then to find out whether the result matched the expectation.
Automated WML unit tests
WML unit tests are self-contained scenario files to test a specific area of WML.
The test result is a status code from the unit_test_result enum found in game_launcher.hpp, or
in rare cases tests expect to be timed out by the test runner. They can be run individually with
Wesnoth's -u <testname> command line argument, but are usually run by the run_wml_tests script
based on the list in wml_test_schedule.
They are unlikely to return the same status if run with -t <testname>. Running them with -t can
still be helpful for debugging.
Guidelines for writing automated new tests
Tests are generally implemented with the GENERIC_UNIT_TEST macro, with two leaders called Alice
and Bob on separate keeps. If your test needs them to be adjacent to each other, consider using
COMMON_KEEP_A_B_UNIT_TEST instead, which puts their starting locations next to each other instead
of needing to move them.
Most will expect the result PASS, and new tests should generally be written to result in a PASS.
The testing mechanism supports other expectations too, however the optimisation to run a batch of
tests in a single instance of Wesnoth currently only supports batching for tests that return PASS.
Tests that shouldn't PASS should have a name that makes that expectation obvious. However, the
existing tests don't conform to this goal yet.
Names containing _fail_ or ending _fail are reserved for tests that should not PASS. However,
they may expect a status that is neither PASS nor FAIL.
Names containing break or error might be for tests expected to PASS. Some of these are testing
loops, or testing error-handling that is expected to handle the error.