Closed kenn closed 6 years ago
Thanks for the report, it's updated!
Thanks @kenn! But why was this just a documentation change @Rykian? I'm disappointed that 7418d24 went out disguised as a patch release as it did cause breakage for Rails apps despite its intention not too. In fact the test in c055d0c clearly gives it away as it had to include active_support/time
just for them to pass. That should have been a red flag. I want to help contribute to correct this. Would you be open to:
maybe consider pulling down the 2.0.3 release?
The workaround is not egregious, and many people probably don't even notice the bug, having already loaded the library. Yanking is an earthquake. This is simply a bug like any other, calling for a new patch.
@pboling I agree, I think a patch is sufficient. Keep in mind people won't notice because its a silent failure which isn't good.
@Rykian would you be open to it? I can work on that and have it ready for you.
Yep, I’m open to a patch.
Thank you in advance ;)
@Rykian great thanks! I've set aside time tomorrow so I'll work on it then and have it ready for you. I'll make some suggestions too for those that would like to remove the dependency whether as a 3.0 or a 2.x release. I think it could be removed earlier if the library makes it an optional dependency.
Breaking change: clockwork stopped working after updating from 2.0.2 to 2.0.3 due to
3.minutes
getting NoMethodError, and it turned out, this change caused the issue.https://github.com/Rykian/clockwork/commit/7418d24026ab4bcad3cd3a19d46d45c61275ac68
I've fixed the issue by adding back 'active_support/time'.
As we have config examples with
3.minutes
or1.hour
in README, I believe we should also haverequire 'active_support/time'
in the examples as well.