On 10 November 2014 13:11, Pekka Paalanen <[email protected]> wrote:
> On Fri, 7 Nov 2014 21:56:48 +0200 > Giulio Camuffo <[email protected]> wrote: > > > I have a couple of comments below, otherwise it looks good. > > > > > > -- > > Giulio > > > > > > 2014-09-24 15:37 GMT+03:00 Marek Chalupa <[email protected]>: > > > Add test_set_timeout() function that allows the test to > > > set timeout for its completition. Any other call to the function > > > re-sets the timeout to the new value. The timeouts can be turned off > > > (usefull when debugging) by setting evironment variable > > > NO_TIMEOUTS. > > > > I would namespace the env var a bit, something like > WAYLAND_TESTS_NO_TIMEOUT. > This follows the naming of LEAK_CHECK_ENABLED. I'll rename it and then send follow-up patch that will rename LEAK_CHECK_ENABLED, so that it'll be consistent. > > > > > > > > Signed-off-by: Marek Chalupa <[email protected]> > > > --- > > > tests/test-runner.c | 46 > ++++++++++++++++++++++++++++++++++++++++++++++ > > > tests/test-runner.h | 7 +++++++ > > > 2 files changed, 53 insertions(+) > > > > > > diff --git a/tests/test-runner.c b/tests/test-runner.c > > > index af80d2b..c4687b7 100644 > > > --- a/tests/test-runner.c > > > +++ b/tests/test-runner.c > > > @@ -44,6 +44,11 @@ static void* (*sys_calloc)(size_t, size_t); > > > > > > int leak_check_enabled; > > > > > > +/* when this var is set to 0, every call to test_set_timeout() is > > > + * suppressed - handy when debugging the test. Can be set by > > > + * NO_TIMEOUTS evnironment var */ > > > +static int timeouts_enabled = 1; > > > + > > > extern const struct test __start_test_section, __stop_test_section; > > > > > > __attribute__ ((visibility("default"))) void * > > > @@ -110,14 +115,54 @@ usage(const char *name, int status) > > > exit(status); > > > } > > > > > > +void > > > +test_set_timeout(int to) > > > +{ > > > + int re; > > > + > > > + if (!timeouts_enabled) { > > > + fprintf(stderr, "Timeouts suppressed.\n"); > > > + return; > > > + } > > > + > > > + re = alarm(to); > > > > alarm() takes an 'unsigned', test_set_timeout should probably do the > > same, or anyway catch <0 timeouts. > Yes > > > > > + fprintf(stderr, "Timeout was %sset", re ? "re-" : ""); > > > + > > > + if (to != 0) > > > + fprintf(stderr, " to %d second from now.\n", to); > > > + else > > > + fprintf(stderr, " off.\n"); > > > +} > > > + > > > +static void > > > +sigalrm_handler(int signum) > > > +{ > > > + fprintf(stderr, "Test timeouted.\n"); > > > > nitpick: "timeouted"? Timed out sonds better to me. ;) > true > > > > > + abort(); > > > +} > > > + > > > static void > > > run_test(const struct test *t) > > > { > > > int cur_alloc = num_alloc; > > > int cur_fds, num_fds; > > > + struct sigaction sa; > > > > > > cur_fds = count_open_fds(); > > > + > > > + if (timeouts_enabled) { > > > + sa.sa_handler = sigalrm_handler; > > > + sa.sa_flags = 0; > > > + sigemptyset(&sa.sa_mask); > > > + assert(sigaction(SIGALRM, &sa, NULL) == 0); > > > + } > > > + > > > t->run(); > > > + > > > + /* turn off timeout (if any) after test completition */ > > > + if (timeouts_enabled) > > > + alarm(0); > > > + > > > if (leak_check_enabled) { > > > if (cur_alloc != num_alloc) { > > > fprintf(stderr, "Memory leak detected in test. > " > > > @@ -189,6 +234,7 @@ int main(int argc, char *argv[]) > > > sys_free = dlsym(RTLD_NEXT, "free"); > > > > > > leak_check_enabled = !getenv("NO_ASSERT_LEAK_CHECK"); > > > + timeouts_enabled = !getenv("NO_TIMEOUTS"); > > > > > > if (argc == 2 && strcmp(argv[1], "--help") == 0) > > > usage(argv[0], EXIT_SUCCESS); > > > diff --git a/tests/test-runner.h b/tests/test-runner.h > > > index 707504c..522cb84 100644 > > > --- a/tests/test-runner.h > > > +++ b/tests/test-runner.h > > > @@ -37,4 +37,11 @@ count_open_fds(void); > > > void > > > exec_fd_leak_check(int nr_expected_fds); /* never returns */ > > > > > > +/* > > > + * set/reset the timeout in seconds. The timeout starts > > > + * at the point of invoking this function > > > + */ > > > +void > > > +test_set_timeout(int); > > > + > > > #endif > > Yup, I completely agree with Giulio here. > > > Thanks, > pq > Will fix it, Thanks, Marek
_______________________________________________ wayland-devel mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/wayland-devel
