py.test not running any tests when using -n (xdist)

[pytestbot]

Originally reported by: BitBucket: vmalloc, GitHub: vmalloc

---

When I run py.test without any arguments all my tests are running as expected:

$ py.test tests
========================================================= test session starts =========================================================
platform darwin -- Python 2.7.9 -- py-1.4.26 -- pytest-2.6.4
plugins: capturelog, xdist
collected 438 items
....... (omitted)

=============================================== 435 passed, 3 skipped in 33.90 seconds ================================================

However, when I pass -n 4, nothing gets actually run:

$ py.test -n 4 
========================================================= test session starts =========================================================
platform darwin -- Python 2.7.9 -- py-1.4.26 -- pytest-2.6.4
plugins: capturelog, xdist
gw0 [439] / gw1 [439] / gw2 [439] / gw3 [439]
scheduling tests via LoadScheduling

==========================================================  in 3.03 seconds ===========================================================

---

• Bitbucket: https://bitbucket.org/pytest-dev/pytest/issue/669

[pytestbot]

Original comment by BitBucket: vmalloc, GitHub: vmalloc:

---

Managed to pin the problem on a specific use case: it happens when parameters are different between workers in pytest.mark.parametrize. For instance, the following simple test will reproduce it:

#!python

import time

import pytest

@pytest.mark.parametrize('param', [time.time(), 1, 2, 3])
def test_something(param):
    pass

It would be helpful to at least get a clear error message explaining why things aren't running...

[pytestbot]

Original comment by Rok Garbas (BitBucket: garbas, GitHub: garbas):

---

i also stumble upon this problem. (pytest-2.6.2, pytest-xdist-1.11)
somebody has any idea how to go about solving this? any pointers would be welcome.

[pytestbot]

Original comment by Floris Bruynooghe (BitBucket: flub, GitHub: flub):

---

AFAIK when using xdist the tests collected on all hosts must have exactly the same test IDs. This means you can not have different parameters depending on the host.

You might be able to work around this in the above example by explicitly giving parameter IDs (which are used to create the test IDs): @pytest.mark.parametrize('param', params=[time.time(), 1, 2, 3], ids=['time', '1', '2', '3']).

[pytestbot]

Original comment by Floris Bruynooghe (BitBucket: flub, GitHub: flub):

---

Actually, I believe this is a duplicate of issue #594.

[pytestbot]

Original comment by Rok Garbas (BitBucket: garbas, GitHub: garbas):

---

same error is also present in pytest-2.7.0

[The-Compiler]

When I run this with an up-to-date pytest and pytest-xdist, I get an error message instead:

================================== ERRORS ===================================
___________________________ ERROR collecting gw0 ____________________________
Different tests were collected between gw2 and gw0. The difference is:
--- gw2

+++ gw0

@@ -1,4 +1,4 @@

-foo.py::test_something[1436343943.2776034]
+foo.py::test_something[1436343943.234645]
 foo.py::test_something[1]
 foo.py::test_something[2]
 foo.py::test_something[3]
___________________________ ERROR collecting gw1 ____________________________
Different tests were collected between gw2 and gw1. The difference is:
--- gw2

+++ gw1

@@ -1,4 +1,4 @@

-foo.py::test_something[1436343943.2776034]
+foo.py::test_something[1436343943.2553692]
 foo.py::test_something[1]
 foo.py::test_something[2]
 foo.py::test_something[3]
___________________________ ERROR collecting gw3 ____________________________
Different tests were collected between gw2 and gw3. The difference is:
--- gw2

+++ gw3

@@ -1,4 +1,4 @@

-foo.py::test_something[1436343943.2776034]
+foo.py::test_something[1436343943.3050516]
 foo.py::test_something[1]
 foo.py::test_something[2]
 foo.py::test_something[3]

So can this be closed, @garbas?

[RonnyPfannschmidt]

Closing as duplicate of #920