bpo-31173: Rewrite WSTOPSIG test of test_subprocess #3055
Conversation
The current test_child_terminated_in_stopped_state() function test creates a child process which calls ptrace(PTRACE_TRACEME, 0, 0) and then crash (SIGSEGV). The problem is that calling os.waitpid() in the parent process is not enough to close the process: the child process remains alive and so the unit test leaks a child process in a strange state. Closing the child process requires non-trivial code, maybe platform specific. Remove the functional test and replaces it with an unit test which mocks os.waitpid() using a new _testcapi.W_STOPCODE() function to test the WIFSTOPPED() path.
|
I think your mocking test makes sense here. LGTM. Somewhat unfortunate to lose the integration test that triggers the actual OS behavior that led to the bug and confirms it is fixed, but if it is fragile and leaking resources that is not good. Is there no way to unwedge the stuck killed but stopped for tracing process using a ptrace API? |
The leaking resource is a blocker point for me. When I ran "./python -m test -m test_child_terminated_in_stopped_state -F test_subprocess", my laptop started to die: the total memory usage was close to 100%!
I hate ptrace() because it makes processes enter very weird states. For example, sending SIGKILL to a traced process is not enough to kill it... "traced process": a process which just called ptrace(PTRACE_ME) with no other process attached to it ... I spent some time working on a debugger written in pure Python: https://github.com/haypo/python-ptrace/ I was always amazed by the weird ptrace bugs I had. Sometimes, I just had to reboot my computer. It was impossible to kill some processes, my computer was stuck. So yeah, usually I prefer functional tests over unit tests, but here I think that's an unit test is an acceptable compromise. What do you think? -- To be honest, I didn't try hard to workaround the weird zombie process. I tried two additional os.waitpid() on macOS, but I had another behaviour on Linux... So I just decided to give up and try something completely different ;-) |
|
GH-3070 is a backport of this pull request to the 3.6 branch. |
|
btw for reference, what platform (os and version) were you on where you saw the problem? |
The increasing memory and https://bugs.python.org/issue31173#msg300066 message was on Linux. My attempt to use additional os.waitpid() to workaround the issue was tried on macOS. |
The current test_child_terminated_in_stopped_state() function test creates a child process which calls ptrace(PTRACE_TRACEME, 0, 0) and then crash (SIGSEGV). The problem is that calling os.waitpid() in the parent process is not enough to close the process: the child process remains alive and so the unit test leaks a child process in a strange state. Closing the child process requires non-trivial code, maybe platform specific. Remove the functional test and replaces it with an unit test which mocks os.waitpid() using a new _testcapi.W_STOPCODE() function to test the WIFSTOPPED() path. (cherry picked from commit 7b7c6dc)
The current test_child_terminated_in_stopped_state() function test creates a child process which calls ptrace(PTRACE_TRACEME, 0, 0) and then crash (SIGSEGV). The problem is that calling os.waitpid() in the parent process is not enough to close the process: the child process remains alive and so the unit test leaks a child process in a strange state. Closing the child process requires non-trivial code, maybe platform specific. Remove the functional test and replaces it with an unit test which mocks os.waitpid() using a new _testcapi.W_STOPCODE() function to test the WIFSTOPPED() path. (cherry picked from commit 7b7c6dc)
The current test_child_terminated_in_stopped_state() function test
creates a child process which calls ptrace(PTRACE_TRACEME, 0, 0) and
then crash (SIGSEGV). The problem is that calling os.waitpid() in the
parent process is not enough to close the process: the child process
remains alive and so the unit test leaks a child process in a
strange state. Closing the child process requires non-trivial code,
maybe platform specific.
Remove the functional test and replaces it with an unit test which
mocks os.waitpid() using a new _testcapi.W_STOPCODE() function to
test the WIFSTOPPED() path.
https://bugs.python.org/issue31173