Python federated execution sometimes segfaults

[housengw]

minimal program to reproduce the problem:

target Python {
    timeout: 10 msec
}

reactor a {
    timer t(0, 1 msec)
    output o;
    reaction(t) -> o {=
        o.set(1);
        print("a: sent a message to b")
    =}
}

reactor b {
    input i;
    reaction(i) {=
        print("b: received a message from a")
    =}
    
}

federated reactor {
    a = new a();
    b = new b();
    a.o -> b.i;
}

program output

oysteryio-2:scratch wonghouseng$ bash /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld
Federate HelloWorld in Federation ID 'd773bb9a0ec81be80fb99144eb7d010a10cf5ce4ac8259ad'
#### Launching the runtime infrastructure (RTI).
RTI: Federation ID: d773bb9a0ec81be80fb99144eb7d010a10cf5ce4ac8259ad
RTI: Number of federates: 2
RTI: Clock sync: init
RTI: Clock sync exchanges per interval: 10
Starting RTI for 2 federates in federation ID d773bb9a0ec81be80fb99144eb7d010a10cf5ce4ac8259ad
RTI using TCP port 15045 for federation d773bb9a0ec81be80fb99144eb7d010a10cf5ce4ac8259ad.
RTI: Listening for federates.
#### Launching the federate a.
#### Launching the federate b.
#### Bringing the RTI back to foreground so it can receive Control-C.
RTI -i ${FEDERATION_ID} -n 2 -c initial exchanges-per-interval 10
Federation ID for executable /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/src-gen/HelloWorld/b/HelloWorld_b.py: d773bb9a0ec81be80fb99144eb7d010a10cf5ce4ac8259ad
Python(1330,0x10e12de00) malloc: Incorrect checksum for freed object 0x7ff04b70cb60: probably modified after being freed.
Corrupt value: 0x1094176b0
Python(1330,0x10e12de00) malloc: *** set a breakpoint in malloc_error_break to debug
Federate 1: Connected to RTI at localhost:15045.
Federate 1: ---- Start execution at time Sat Oct 30 22:14:29 2021
---- plus 453717000 nanoseconds.

The program seems to loop infinitely there. If I press CTRL-C, I get:

^CFederate 1: ERROR: Read 0 bytes, but expected 9.
Federate 1: FATAL ERROR: Failed to read MSG_TYPE_TIMESTAMP message from RTI.
Federate 1: ---- Elapsed logical time (in nsec): 0
Federate 1: ---- Elapsed physical time (in nsec): 74,308,920,000
/Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld: line 68:  1330 Abort trap: 6           python3 /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/src-gen/HelloWorld/a/HelloWorld_a.py -i $FEDERATION_ID
#### Received ERR signal on line 68. Invoking cleanup().
Killing federate 1330.
Killing federate 1331.

It happens once every 4 ~ 5 times I run the program. Also, the error message is sometimes different. Here's another error message I got from running the same program

oysteryio-2:scratch wonghouseng$ bash /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld
Federate HelloWorld in Federation ID '0b536f4d41d915e7ad3a7f8c9ef7d18930048adae8b7db68'
#### Launching the runtime infrastructure (RTI).
RTI: Federation ID: 0b536f4d41d915e7ad3a7f8c9ef7d18930048adae8b7db68
RTI: Number of federates: 2
RTI: Clock sync: init
RTI: Clock sync exchanges per interval: 10
Starting RTI for 2 federates in federation ID 0b536f4d41d915e7ad3a7f8c9ef7d18930048adae8b7db68
RTI using TCP port 15045 for federation 0b536f4d41d915e7ad3a7f8c9ef7d18930048adae8b7db68.
RTI: Listening for federates.
#### Launching the federate a.
#### Launching the federate b.
#### Bringing the RTI back to foreground so it can receive Control-C.
RTI -i ${FEDERATION_ID} -n 2 -c initial exchanges-per-interval 10
Federation ID for executable /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/src-gen/HelloWorld/b/HelloWorld_b.py: 0b536f4d41d915e7ad3a7f8c9ef7d18930048adae8b7db68
Federation ID for executable /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/src-gen/HelloWorld/a/HelloWorld_a.py: 0b536f4d41d915e7ad3a7f8c9ef7d18930048adae8b7db68
Python(1351,0x111f81e00) malloc: Incorrect checksum for freed object 0x7fcc86d10ab0: probably modified after being freed.
Corrupt value: 0x10febd6b0
Python(1350,0x11240be00) malloc: Incorrect checksum for freed object 0x7fd7e35263a0: probably modified after being freed.
Corrupt value: 0x10e3fd6b0
Python(1350,0x11240be00) malloc: *** set a breakpoint in malloc_error_break to debug
Python(1351,0x111f81e00) malloc: *** set a breakpoint in malloc_error_break to debug

It loops infinitely there. If I press CTRL-C, I get:

^C/Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld: line 68:  1350 Abort trap: 6           python3 /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/src-gen/HelloWorld/a/HelloWorld_a.py -i $FEDERATION_ID
/Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld: line 68:  1351 Abort trap: 6           python3 /Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/src-gen/HelloWorld/b/HelloWorld_b.py -i $FEDERATION_ID
#### Received ERR signal on line 68. Invoking cleanup().
Killing federate 1350.
Killing federate 1351.
/Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld: line 15: kill: (1350) - No such process
/Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld: line 15: kill: (1351) - No such process
#### Killing RTI 1348.
/Users/wonghouseng/lingua-franca-master/runtime-EclipseXtext/scratch/bin/HelloWorld: line 17: kill: (1348) - No such process

[Soroosh129]

Wow, nice catch! I will look into it.