Random test failures on windows after editing a python file

[pytestbot]

Originally reported by: BitBucket: pombredanne, GitHub: pombredanne

---

I get random failures on windows only (not on mac and not on linux) with win7 and Python 2.7.8 and py.test 2.6.1.

This heisenbug occurs some times after editing a file and running py.test.
Somehow, I found that the previous version of the code was executed, not the one saved.
Clearing the pycache solved the issue each time.

I reckon that I could disable bytecode creation (as per #200 )
However why should I need to worry about a special python flag or env var just for py.test?

Note: the issue could be due to poor filesystem timestamps resolution on windows when files are updated.

Could it make sense to have a flag disable caching? or only enable caching with a command flag?

---

• Bitbucket: https://bitbucket.org/pytest-dev/pytest/issue/576

[pytestbot]

Original comment by BitBucket: pombredanne, GitHub: pombredanne:

---

By the way, I do not know when Windows updates the modification time stamps in NTFS, but this is rather troubling:

http://msdn.microsoft.com/en-us/library/windows/desktop/ms724290%28v=vs.85%29.aspx

"The NTFS file system delays updates to the last access time for a file by up to 1 hour after the last access."

Otherwise, NTFS resolution is 100 nanosec.

[pytestbot]

Original comment by Benjamin Peterson (BitBucket: gutworth, GitHub: gutworth):

---

Hopefully a433de9ca4f0 will help.

[pytestbot]

Original comment by BitBucket: pombredanne, GitHub: pombredanne:

---

@gutworth thx++ ... let me test that. It can sure help
though I think that pycache should not be used on windows at all.
I will test and report1

[RonnyPfannschmidt]

Closing as stalled