A possible problem on update caused by the update of 0.3.6

[winnnnnny]

Sorry for disturbing you!
I observed an unexpected update when I use version 0.3.6
Here is an example.

a = sint(0)
b = sint(1)
c = sint(2)
# a is updated with the value of b, but it creates an unexpected shallow copy here.
a.update(b)
# a and b share the same value 1 here
print_ln('b: %s', b.reveal())
print_ln('a: %s', a.reveal())
# a is updated with the value of c, but b is also updated, which is not expected
a.update(c)
# a and b share the same value 2 here but b is expected to be still 1.
print_ln('b: %s', b.reveal())
print_ln('a: %s', a.reveal())

The result is shown below.

./Scripts/mascot.sh test
Running /home/mp-spdz/Scripts/../mascot-party.x 0 test -pn 10909 -h localhost -N 2
Running /home/mp-spdz/Scripts/../mascot-party.x 1 test -pn 10909 -h localhost -N 2
Using security parameter 40
b: 1
a: 1
b: 2
a: 2
The following benchmarks are including preprocessing (offline phase).
Time = 0.00150634 seconds 
Data sent = 0.041596 MB in ~49 rounds (party 0 only; use '-v' for more details)
Global data sent = 0.083192 MB (all parties)

By the way, the result from 0.3.5 is also shown below.

./Scripts/mascot.sh test
Running /home/mp-spdz-0.3.5/Scripts/../mascot-party.x 0 test -pn 19733 -h localhost -N 2
Running /home/mp-spdz-0.3.5/Scripts/../mascot-party.x 1 test -pn 19733 -h localhost -N 2
Using security parameter 40
b: 1
a: 1
b: 1
a: 2
The following benchmarks are including preprocessing (offline phase).
Time = 0.00252492 seconds 
Data sent = 0.041596 MB in ~537 rounds (party 0; use '-v' for more details)
Global data sent = 0.083192 MB (all parties)

The behavior is different from versions 0.3.5 and 0.3.6. Is this a new property introduced in 0.3.6 or a bug?

[mkskeller]

Thank you for raising this. You should find that 8e735da fixes it.

[winnnnnny]

Thank you for repairing this!