Call getJobs in thread: PyEval_RestoreThread: NULL tstate

[anxuae]

Calling the function getJobs in a thread crash the application. It seems that some functions are not thread-safe. How can I handle it?

Exception in thread Thread-1:
Traceback (most recent call last):
  File "/usr/lib/python3.7/threading.py", line 917, in _bootstrap_inner
    self.run()
  File "/usr/lib/python3.7/threading.py", line 865, in run
    self._target(*self._args, **self._kwargs)
  File "test3.py", line 19, in print_jobs
    "job-uri", "job-state"])))
cups.IPPError: (1280, 'No such file or directory')

Fatal Python error: PyEval_RestoreThread: NULL tstate

Current thread 0x76f6cad0 (most recent call first):
  File "test3.py", line 34 in <module>
Abandon

I'm using the following script to reproduce it:

import threading
import time
import cups
import os.path as osp

FILENAME = "test3.py"
RUN = True

def print_jobs():
    time.sleep(1)
    while RUN == True:
        time.sleep(0.1)
        print("Jobs count:", len(conn.getJobs(my_jobs=True, requested_attributes=["job-id", "job-name",
                                                                                  "job-uri", "job-state"])))

THREAD = threading.Thread(target=print_jobs)
THREAD.daemon = True

# Create a CUPS connection
conn = cups.Connection()
name = conn.getDefault()

try:
    THREAD.start()
    while True:
        time.sleep(0.1)
        conn.printFile(name, FILENAME, osp.basename(FILENAME), {})

except KeyboardInterrupt:
    print("Stop")

finally:
    print("Cancel tasks")
    RUN = False
    conn.cancelAllJobs(name)
    THREAD.join()

[zdohnal]

Hi,

I've tried to reproduce the issue with your script, but with no luck... are you able to get me C backtraces (the pycups is a module for python written in C) for both issues you see in pibooth?

The first is the one you experience with your script, and the other is Fatal Python error: PyEval_RestoreThread: the function must be called with the GIL held, but the GIL is released (the current Python thread state is NULL).

[anxuae]

I've tried to get more logs by exporting those variables:

PYTHONTRACEMALLOC=1
PYTHONTHREADDEBUG=1
PYTHONFAULTHANDLER=1

But I can't get more trace info. Can you tell me how can I get C backtraces ?

Note : I'm using pycups==2.0.1

[zdohnal]

When I got a similar bug about incorrect thread handling, python3 process itself got segfault and generated coredump, where I could see thread conflict causing the crash. Does python3 process running pibooth get SIGSEGV as well? If it does, you can get the backtrace from its coredump.