mirror of
https://github.com/php/pftt2.git
synced 2026-03-24 01:02:11 +01:00
Tests crash/skip for php-7.4 with opcache config #59
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @lavturo on GitHub (Dec 11, 2019).
I was testing php-7.4.1RC1 and saw that a lot more tests were classified as CRASH or SKIP than usual.
When pftt is not configured with opcache, there are:
~11,000 tests passed, ~600 tests skipped, ~2 tests crashed.
When pftt is configured with opcache, there are:
~15 tests passed, ~4,600 tests skipped, ~7,400 tests crashed.
Seems like this issue might have been around for quite for a while. Previously, I have been running an older version of pftt2 on VMs and this issue did not occur. I recently updated it and now here it is. Will have to look through PRs to see which one might have caused this.
@lavturo commented on GitHub (Dec 12, 2019):
Seems to be for any build, not just php-7.4.
@cmb69 commented on GitHub (Dec 12, 2019):
These are actually good news (otherwise we might have had a bad build).
Are you sure the crashes are caused by PFTT changes? Could as well be something different. Maybe try to run some of these crashing test cases with run-test.php, or even with plain PHP.
@lavturo commented on GitHub (Dec 12, 2019):
I am almost certain the crashes are caused by pftt changes. I still have a backup copy of pftt2 on the VMs and when I run the tests on them, they get tested normally (not all as crash/skip). However, I am a bit unsure which version of pftt (how many commits it was behind) the ones on the VM are.
Just in case, using run-test.php, I did try to check some tests that resulted as a crash and they pass.
@lavturo commented on GitHub (Dec 12, 2019):
Strange. I was able to replicate the issue locally yesterday, even after a clean build. However, I could not replicate the bug this morning. I exported this to the VMs and they work perfectly fine now.
However, I was able to replicate this bug again after cleaning and building pftt2 after switching my project back to what I was working on (issue 5). The tests would crash/skip again. Here, I realized if I restart my PC, it would work again.
@cmb69 commented on GitHub (Dec 13, 2019):
Are there perhaps hanging processes?
@lavturo commented on GitHub (Dec 13, 2019):
I made sure there weren't any processes left before being able to confirm that I could replicate the issue.
However, I will close this issue for now, since everything is being tested correctly now.