[PATCH libevdev] test: detect if we're running inside gdb and disable forking

[David Herrmann]

Hi

On Sun, Dec 29, 2013 at 10:59 AM, Peter Hutterer
<peter.hutterer at who-t.net> wrote:
> On 28/12/2013 01:50 , David Herrmann wrote:
>>
>> Hi Peter
>>
>> On Mon, Dec 23, 2013 at 11:44 PM, Peter Hutterer
>> <peter.hutterer at who-t.net> wrote:
>>>
>>> The Check test framework forks by default which is annoying when running
>>> gdb.
>>> Try to detect whether we're inside gdb by ptracing ourselves. If that
>>> works,
>>> we're not inside a debugger. If it doesn't, then assume we're inside a
>>> debugger and set CK_FORK to "no".
>>>
>>> Signed-off-by: Peter Hutterer <peter.hutterer at who-t.net>
>>> ---
>>>   libevdev/libevdev.h |  5 +++--
>>>   test/test-main.c    | 37 +++++++++++++++++++++++++++++++++++++
>>>   2 files changed, 40 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/libevdev/libevdev.h b/libevdev/libevdev.h
>>> index c11c845..c5d12a4 100644
>>> --- a/libevdev/libevdev.h
>>> +++ b/libevdev/libevdev.h
>>> @@ -177,8 +177,9 @@ extern "C" {
>>>    *     git grep "suite_create"
>>>    *     git grep "tcase_create"
>>>    *
>>> - * By default, check forks, making debugging harder. Run gdb as below to
>>> avoid
>>> - * forking.
>>> + * By default, Check forks, making debugging harder. The test suite
>>> tries to detect
>>> + * if it is running inside gdb and disable forking. If that doesn't work
>>> for
>>> + * some reason, run gdb as below to avoid forking.
>>>    *
>>>    *     sudo CK_FORK=no CK_RUN_TEST="test case name" gdb
>>> ./test/test-libevdev
>>
>>
>> You should also change this line then.
>>
>>>    *
>>> diff --git a/test/test-main.c b/test/test-main.c
>>> index 73d79c9..91128c3 100644
>>> --- a/test/test-main.c
>>> +++ b/test/test-main.c
>>> @@ -22,6 +22,10 @@
>>>
>>>   #include <config.h>
>>>   #include <check.h>
>>> +#include <stdlib.h>
>>> +#include <unistd.h>
>>> +#include <sys/ptrace.h>
>>> +#include <sys/wait.h>
>>>
>>>   extern Suite *event_name_suite(void);
>>>   extern Suite *event_code_suite(void);
>>> @@ -31,9 +35,42 @@ extern Suite *libevdev_has_event_test(void);
>>>   extern Suite *libevdev_events(void);
>>>   extern Suite *uinput_suite(void);
>>>
>>> +static int
>>> +is_debugger_attached(void)
>>> +{
>>> +       int status;
>>> +       int rc;
>>> +       int pid = fork();
>>> +
>>> +       if (pid == -1)
>>> +               return 0;
>>> +
>>> +       if (pid == 0) {
>>> +               int ppid = getppid();
>>> +               if (ptrace(PTRACE_ATTACH, ppid, NULL, NULL) == 0) {
>>> +                       waitpid(ppid, NULL, 0);
>>
>>
>> I guess PTRACE_ATTACH stops the tracee or what are you waiting here for?
>>
>>> +                       ptrace(PTRACE_CONT, NULL, NULL);
>>> +                       ptrace(PTRACE_DETACH, ppid, NULL, NULL);
>>> +                       rc = 0;
>>> +               } else
>>> +                       rc = 1;
>>> +               _exit(rc);
>>> +       } else {
>>> +               waitpid(pid, &status, 0);
>>> +               rc = WEXITSTATUS(status);
>>> +       }
>>> +
>>> +       return rc;
>>> +}
>>> +
>>> +
>>>   int main(int argc, char **argv)
>>>   {
>>>          int failed;
>>> +
>>> +       if (is_debugger_attached())
>>
>>
>> Can you add: && !getenv("CK_FORK")
>> Allows to set CK_FORK explicitly on the command-line.
>
>
> setenv("foo", "bar", 0) only overwrites foo when it isn't set, otherwise
> it'll leave it untouched (see the man page) so we're good here.

Oh, I missed that. Fine with me then.

Thanks
David

>
>
>
>>
>> Thanks
>> David
>>
>>> +               setenv("CK_FORK", "no", 0);
>>> +
>>>          Suite *s = libevdev_has_event_test();
>>>          SRunner *sr = srunner_create(s);
>>>          srunner_add_suite(sr, libevdev_events());
>>> --
>>> 1.8.4.2
>>>
>>> _______________________________________________
>>> Input-tools mailing list
>>> Input-tools at lists.freedesktop.org
>>> http://lists.freedesktop.org/mailman/listinfo/input-tools
>
>