[Bug python/27682] New: Continue event in Python delivered after thread dies

["enedil at outlook dot com" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=27682

            Bug ID: 27682
           Summary: Continue event in Python delivered after thread dies
           Product: gdb
           Version: HEAD
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: python
          Assignee: unassigned at sourceware dot org
          Reporter: enedil at outlook dot com
  Target Milestone: ---

I have this C++ (this is C++, not C) code:

#include <pthread.h>
#include <stdlib.h>
#include <signal.h>
#include <stdint.h>

void* foo(void* ptr) {
    return NULL;
}

int main() {
    pthread_t th;
    pthread_create(&th, NULL, foo, NULL);
    pthread_kill(pthread_self(), SIGSEGV);
    pthread_join(th, NULL);
}


This is what happens:
/tmp $ /tmp/gdb/bin/gdb -q -n -ex 'pi gdb.events.cont.connect(lambda ev:
gdb.parse_and_eval("2137"))' -ex 'run' ./test
Reading symbols from ./test...
Starting program: /tmp/test 
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
[New Thread 0x7ffff7dae640 (LWP 426580)]
thread.c:72: internal-error: thread_info* inferior_thread(): Assertion
`current_thread_ != nullptr' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)



I compile it in either of these ways:
    gcc -x c++ testcase.cc -g -pthread -o test
    g++ testcase.cc -g -pthread -o test
    clang++ testcase.cc -g -pthread -o test
These however don't trigger the bug:
    gcc -x c testcase.cc -g -pthread -o test
    gcc -x c++ testcase.cc -pthread -o test
    clang -x c testcase.cc -g -pthread -o test

My environment: fully upgraded Fedora 33 Desktop (amd64), with GDB just taken
from master branch (GNU gdb (GDB) 11.0.50.20210331-git) 
Compilers: 
    gcc (GCC) 10.2.1 20201125 (Red Hat 10.2.1-9)
    clang version 11.0.0 (Fedora 11.0.0-2.fc33)

Bug was confirmed on default configuration of Arch Linux.


It should be noted, that if I use GNU gdb (GDB) Fedora 10.1-4.fc33, the doesn't
show up just after `run`, but if I make a `break main`, and later issue
`continue`, it appears once again.

Possibly connected bug: https://sourceware.org/bugzilla/show_bug.cgi?id=22474

-- 
You are receiving this mail because:
You are on the CC list for the bug.