It is convenient that test-run uses a tarantool / a tarantoolctl executables found in PATH, when we want to test a module (say, vshard). However it is undesirable behaviour, when we test tarantool itself: it is not always obvious what is going on when I forgot to build tarantool prior to run tests.
We already detect tarantool source directory:
|
def setenv(): |
|
"""Find where is tarantool dir by check_file""" |
|
check_file = 'src/trivia/util.h' |
|
path = os.path.abspath('../') |
|
while path != '/': |
|
if os.path.isfile('%s/%s' % (path, check_file)): |
|
os.putenv('TARANTOOL_SRC_DIR', path) |
|
break |
|
path = os.path.abspath(os.path.join(path, '../')) |
So it should be easy to forbid using of an external tarantool in the case.
It is convenient that test-run uses a tarantool / a tarantoolctl executables found in PATH, when we want to test a module (say, vshard). However it is undesirable behaviour, when we test tarantool itself: it is not always obvious what is going on when I forgot to build tarantool prior to run tests.
We already detect tarantool source directory:
test-run/lib/__init__.py
Lines 14 to 22 in e843552
So it should be easy to forbid using of an external tarantool in the case.