-
Notifications
You must be signed in to change notification settings - Fork 308
Conversation
Assuming this lands in |
Will it fail here? :] |
http://faker.readthedocs.io/en/master/providers/faker.providers.date_time.html |
Seemingly(?) fixed under #4130 (comment) by 3dda2a1. Not what I would have expected since "now" is supposedly the default. If the presence of the bug varies with the time of day, then rerunning the build on #4157 that doesn't have the fix #4130 has—we might expect that to also pass. |
Build restarted. |
Build passes! I don't think we're out of the woods yet. |
Harumph. Passing locally. Can I blame it on caching at Travis? |
/me restarts build ... |
Man, I just figured we had the right version of faker. Huh. |
We upgrade dependencies manually, to avoid the risk that newer versions break things we were depending on. We could certainly do a better job of upgrading dependencies on a schedule as opposed to haphazardly when stuff like this happens—the generalized case of gratipay/inside.gratipay.com#776. |
Apparently not. |
@@ -33,7 +33,10 @@ | |||
./vendor/contextlib2-0.5.4.tar.gz | |||
./vendor/raven-5.32.0.tar.gz | |||
|
|||
./vendor/fake-factory-0.3.2.tar.gz | |||
./vendor/six-1.8.0.tar.gz |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is already in the next section?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, but it needs to be installed before Faker since Faker depends on it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Same with the next two lines, actually.
joke2k/faker#332 is this issue, but we're installing from .tar.gz. |
Heh ... COS fixed it in CenterForOpenScience/osf.io@381c851 by pinning to version 0.5.3. Does that version include the bugfix we need for our own presenting issue? |
fe08989
to
82e5eb9
Compare
|
Local working
|
Here's where the breakdown occurs: Working Locally
Broken at Travis
|
With dffcac0: Working Locally
Broken at Travis
|
The broken version at Travis includes |
Cache cleared and build restarted ... 🙈 |
Not a cache issue. Hmm ... 8faba0e to confirm joke2k/faker#332 (comment) ... |
Confirmed.
|
Local:
|
I note that there are also |
Different red. |
Green! Does caching predict the result? |
|
I'm going to try removing all caches. |
My hypothesis is that when we don't specify the cache directory for the branch, Travis is falling back to a spoiled cache from master or something like that. |
Well, that should be green, but it won't give us info about the hypothesis that a fall-through is causing failures to persist. For that I think we would need to remove all caches after all. |
|
Here we go! |
💃 |
3dff9ab
to
45cb68d
Compare
Debugging commits rebased off. Was b2a33f9. |
Build restarted for 45cb68d ... |
Green! Ready to declare victory here, @mattbk? |
Nice work! |
From #4157 (comment) and #4130 (comment) ...