New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Recurrences (tests?) have an end-of-day edge case #1722
Comments
that could well be the problem. I do not think we set the time when we use Travel. I'll look into this before 2.3final, although this is a false positive I presume |
A quick look at the failing tests makes me think time travel is not the right answer here. Could you please do this as a PR before merging anything into 2.3? I'd like the chance to review it. |
Did you manage to reproduce the bug? I tried changing my system clock back to the error state, changing the time zone and changing the time_zone setting in the site.yml. I tried various permutations of these things without any success. Any other idea on how to reproduce it? |
I don't think time_zone in site.yml will have an effect, but an easy way to override the system time zone is to set the |
I tried with the following three commands:
Between each test I was changing my local timezone (as you can see by the date result). I really can't understand what was going on :( |
I have seen test failures when the edge-case was an end-of-month in the middle of the week or something like that. So these tests rule out timezone issues, but not other edge-cases |
I was able to reproduce this bug in a VM with a disconnected clock:
|
Once I got the conditions right I was also able to reproduce this on my Mac with libfaketime:
|
Also note that although the Travis build was on the master branch, I have confirmed the bug in |
Thanks @dnrce I can reproduce the bug now. BTW now I experience a fourth failure:
both from master and from 2.3 branch. Isn't that strange that you don't experience the same error? Am I doing something weird? |
@matjack1 That's a weird one. Are you using libfaketime? |
@matjack1 I'm also getting seemingly random failures of |
Yes @dnrce I'm using libfaketime as well. It looks like a library "strangeness", especialy if you look to the strange dates... |
The PR #1743 should address the test problems. Most of the problems come from the time zone handling that causes problems near the end of the day (I reckon). If you have any doubt we can continue the discussion directly in the code maybe. |
See https://travis-ci.org/TracksApp/tracks/jobs/36919173 which ran at 2014-10-02T22:39:54Z.
From the Travis log, the actual start and stop times of the
rake ci
task were Thu, 02 Oct 2014 22:44:55 GMT and Thu, 02 Oct 2014 23:02:56 GMT, respectively.The text was updated successfully, but these errors were encountered: