Changes:
* CHECK_STRIKES macro now takes a comma-delimited list of strike counts for when units have a different number of strikes for different weapons, and updates the existing usage in the attacks tests
* Fixed a typo in the attacks_zero test
* Adds tests for berserk as a weapon special ability.
Berserk's handling differs from attacks as a weapon special ability:
* A value less than 1 is treated as effectively infinite rounds of combat (undocumented)
* The cumulative attribute is handled differently - in some cases it sums the values instead of using the highest single value whereas for attacks it always uses the highest single value
Also now exclude data/tests/ from scons pot-update since it was hitting the argument limit for number of arguments to a script.
This is testing experience=yes and experience=no, with the expectation that
some new values for that attribute will be added afterwards.
Adds a new generic macro ASSERT_UNIT_HP_XP. Looking thorugh the ability tests,
ASSERT_UNIT_HP (without the XP) could be a heavily-used macro, but it wouldn't
be used by this particular commit, so I'm not adding it in this commit.
Turn the 4 side setup of the firststrike_and_laststrike test into a macro
called `COMMON_KEEP_A_B_C_D`for reuse in other tests.
The comment for `GENERIC_UNIT_TEST` gets a couple of extra paragraphs.
All the files get wmlindent run on them, which reindents `FAIL_IF_NOT`.
The main reason for having these is separate files is Git's fuzzy-patch
matching when merging and rebasing. Having large blocks of identical text
in `wml_unit_test_macros.cfg` can mean the wrong section gets patched.
- The t_string type is now a schema built-in type and no longer attempts a regex match.
- You can also specify that non-t_string types may be optionally-translatable; this case supports a regex match on the string (but note that the translation mark is not part of the match).
- Error messages involving keys with very large values ( > 128 characters) will now truncate the value.
- To account for occasional cases where the schema is intentionally violated, the --validate command-line option now automatically defines the SCHEMA_VALIDATION preprocessor define.
A key validates as type t_string if one of the following is true:
- The key is not present
- The key has at least one segment with a translation mark
- The key is blank (an empty string)
Any type other than t_string is not allowed to be translatable by default, unless you specify allow_translatable=yes in the [type] tag.
An optionally-translatable string could also be defined as a union of t_string and some other type.
This adds an additional `test_result` attribute to [endlevel], intended for use with the automated unit tests. This allows for the unit tests to differentiate a pass/fail result separately from scenario victory or defeat, which allows for more accurately determining the outcome of a test as well as addresses the potential, for example, for a scenario to be expect to pass because of the {SUCCEED} macro but instead passes because the scenario ended as a victory through some other method.
Additional unit tests which were the original motivation for this change are also added as part of this. They test, as much as possible, that events are executed at all, and are then also executed in the expected order.
This explains (in the log) why a WML conditional has failed (or succeeded). Currently it is fairly basic, looping through and/or/not and printing out the specific conditional tag that failed as well as, for [variable], the current content of the variable.
It's used in the WML unit tests system, but could also be useful for debugging.
This commit also fixes an issue with passing multiple -a arguments to run_wml_tests.
There is some WML in data/test/scenarios/ that causes wmlindent to
misbehave (bug #22264 and possibly others awaiting a more thorough
diagnosis), so that subdir is not part of this commit.
I initially had linger mode off in the endlevel tags of these
macros, to make sure they didn't block the end of noninteractive
scenarios.
But it turns out its better to have it on. It doesn't block the
tests when they run noninteractively, and it means that if a test
fails, you can boot it up with --showgui, and then run :inspect
from linger mode to see what happened. If you don't have linger
mode, then even with --showgui the failed test closes out too
quickly.
