Wrong error location in traceback inside an assert

[solanav]

Bug report

Bug description:

The location in the traceback for this failed assertion:

test = 3
assert test == 1 and test == 2, "Bug found?"

Seems to be wrong. It should show test == 1 as the source of the error:

Traceback (most recent call last):
  File "/home/.../test.py", line 3, in <module>
    assert test == 1 and test == 2, "Bug found?"
                         ^^^^^^^^^
AssertionError: Bug found?

CPython versions tested on:

3.11

Operating systems tested on:

Linux

Linked PRs

• [3.11] gh-116034: fix location info on the error of a failed assertion #116052
• [3.12] gh-116034: fix location info on the error of a failed assertion #116054

[Eclips4]

Hello!
Thanks for the report!
Traceback for current 3.11 branch:

Traceback (most recent call last):
  File "/Users/admin/Projects/cpython/example.py", line 2, in <module>
    assert test == 1 and test == 2, "Bug found?"
                         ^^^^^^^^^
AssertionError: Bug found?

Traceback for current 3.12 branch:

Traceback (most recent call last):
  File "/Users/admin/Projects/cpython/example.py", line 2, in <module>
    assert test == 1 and test == 2, "Bug found?"
AssertionError: Bug found?

Traceback for current main branch:

Traceback (most recent call last):
  File "/Users/admin/Projects/cpython/example.py", line 2, in <module>
    assert test == 1 and test == 2, "Bug found?"
           ^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: Bug found?

cc @pablogsal

[pablogsal]

Error locations are set by the compiler so this is likely a compiler bug. CC @iritkatriel

[iritkatriel]

in 3.13 it does this:

>>> assert test == 1 and test == 2, "Bug found?"
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
    assert test == 1 and test == 2, "Bug found?"
           ^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: Bug found?

[pablogsal]

Shouldn't this be just one or the other?

[iritkatriel]

Shouldn't this be just one or the other?

Since the line numbers are calculated statically, that would require emitting a different RAISE instruction for each clause. Currently we don't.

[pablogsal]

Then it seems there isn't much we can do for 3.13, no? For 3.12 it seems that we don't emit line info

[iritkatriel]

For 3.13 we could emit more instructions per assert, if we decide the tradeoff is worth it.

I'm looking at 3.11 / 3.12 and will see if there is an easy fix.

[iritkatriel]

I fixed 3.11 and 3.12 to behave like 3.13. I will close this issue as the problem it reports it resolved.

I'm not sure narrowing the ranges is worth more code. If it ever comes up as an actual problem, we can revisit.

Thank you @solanav .