Is FORMAT DEBUG still any useful?

Can someone has an explanation for this (R2020B)

clear
clc

x=1;
y=x;
x=x+1;

format debug
y
x

that produces

y =


Structure address = 23ebc936f80
m = 1
n = 1
pr = 23edbdd6660

     1


x =


Structure address = 23ebc936f80
m = 1
n = 1
pr = 23edbdd6660

     2

So x and y have the same structure address and data address, yet data are different (1 for y and 2 for x).

Finish the time where FORMAT DEBUG returns useful information for us.

[James Quinlan]

Hi Bruno, were you able to submit a bug report? I am running into similar issues. I am performing memory tests and tests are unreliable in light of your comment. I reproduce results from both you and James Tursa below giving same addresses for x and y even after x has been modified. I am running Version: 9.10.0.1602886 (R2021a).

My test is slightly different in the sense, I assign x a value, then pass it to a function and make assignment, y = x. From my understanding, if you pass x to a function and don't update it while there, it effectively is pass by reference.

[Bruno Luong]

Hi James,

I did not submitted a BUG report. IMO TMW has changed complemetly the shared memory paradism and they just are not willing to share the new mechanism.

Answer 1

[James Tursa]

Strange behavior and probably deserves a bug report. E.g.,

R2019a:

>> x = 1;
>> y = x;
>> x = x + 1;
>> format debug
>> y
y =

Structure address = fa4a7890
m = 1
n = 1
pr = 119f97b20

     1
>> x
x =

Structure address = fa4a76d0
m = 1
n = 1
pr = 17c869780

     2

R2019b:

>> x = 1;
>> y = x;
>> x = x + 1;
>> format debug
>> y
y =

Structure address = f7ef7ac0
m = 1
n = 1
pr = 145dbaec0

     1
>> x
x =

Structure address = f7ef7ac0
m = 1
n = 1
pr = 145dbaec0

     2

R2020a:

>> x = 1;
>> y = x;
>> x = x + 1;
>> format debug
>> y
y =

Structure address = 192d47e2d40
m = 1
n = 1
pr = 192d3b7a5c0

     1
>> x
x =

Structure address = 192d47e2d40
m = 1
n = 1
pr = 192d3b7a5c0

     2

So the problem goes back as far as R2019b.

[cr]

Is there a closure to this? I still see the same behaviour in 2023a.