This should help identify differences in environments, e.g. what the screen
resolution is in `PRINT_MODE:TEXT`. The `scripts/gui/blueprint:render_labels`
test currently fails in an 80x25 resolution, which implies that our CI is using
something else.
to support @BeforeEach and @AfterEach semantics. however, much of our
code is scope-based (lambda-based?) so it seemed to make sense to
use a single wrapper function instead of separate before_each and
after_each functions.
Any library functions written in Lua, like `dfhack.with_finalize()`, would
cause the traceback to end even if there were more stack frames in the test file.
string:split() isn't part of the standard library, but is defined if
`gui/load-screen.lua` is run. This happens in the default dfhack.init-example
and on GitHub Actions, but not on Buildmaster.
This is an intermediate solution to get CI working again. A more complete
solution would be to move string:split() to dfhack.lua, like
string:startswith(), or to take a pass over scripts and make sure they aren't
modifying built-in types like string/table.
- implement navigation function for loading a fortress from the title
screen (requires a fortress save to be ready in region1/).
- ensure we don't try repeatedly to enter a mode that we can't reach
(such as getting back to the title screen from fortress mode). failing
to enter the mode once will skip all remaining tests in that mode.
- provides expect,printerr_match for matching printerr output
- fails tests if printerr is called outside the printerr_match wrapper
- changes api of expect.error_match to mirror the new printerr_match api
This allows tests to test these functions without needing to include the test
wrapper directly (now ci/test.lua, formerly test/main.lua). Hopefully this
location is also more stable, similar to other libraries that are already tested.
lethosor
myk002