The unit-tests binary will often silently terminate part-way through a run - in this case it does *not* return an error to the console.
I'm not aware of any particular test which is causing this; it does not consistently terminate at any one test. Some of the more common outputs are:
…
[ RUN ] AndroidInputSortedVector.count_items_at_removed_index_are_not_found
[ OK ] AndroidInputSortedVector.count_items_at_removed_index_are_not_found (0 ms)
[ RUN ] AndroidInputSortedVector.remove_beyond_end_fails
[ OK ] AndroidInputSortedVector.remove_beyond_end_fails (0 ms)
[----------] 13 tests from AndroidInputSortedVector (1 ms total)%
>
and
…
[ OK ] DrawPatternsTest.checkered_bad_pixel_formats (0 ms)
[----------] 6 tests from DrawPatternsTest (1 ms total)
[----------] Global test environment tear-down
[==========] 361 tests from 79 test cases ran. (427 ms total)%
>
As you can see, it often crashes between printing out the total number of tests run and the number of tests which passed.
First try:
make test
because that will print a summary including what failed in unit-tests.
Second try:
valgrind bin/unit-tests