Skip to content
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

test failures in halotools due to test runner incompatibility #17

Closed
duncandc opened this issue Feb 2, 2015 · 5 comments
Closed

test failures in halotools due to test runner incompatibility #17

duncandc opened this issue Feb 2, 2015 · 5 comments

Comments

@duncandc
Copy link
Contributor

duncandc commented Feb 2, 2015

The package is failing the test suite when merging. This is a astropy problem with the versioning of pytest. @eteq is looking into this for us.

@eteq
Copy link
Member

eteq commented Feb 2, 2015

Oops, missed that this issue exists - d62c292 may have fixed this... It's still failing but I think that's a real test failure rather than the testing system itself?

@eteq
Copy link
Member

eteq commented Feb 2, 2015

Also, for background on the solution: the original problem stems from the fact that the test runner invoked by python setup.py test actually imports astropy to run the tests. Moreover, the latest astropy_helpers are not compatible with older versions of astropy. Travis installs astropy from conda for halotools, but versions later than 4.1 are only available with numpy 1.9, and halotools was running its tests with numpy 1.8 up until d62c292.

So that's super-convoluted why it failed, but the solution was basically just making sure that astropy is an up-to-date version when installed in travis.

To avoid this sort of thing in the future, astropy could do a few things: there could be more checks in the test runner to work around added feature, conda could be updated to include other numpy versions (@embray or @astrofrog - was this intentional, or just an oversight?), or the test runner should be fully integrated with the helpers (see astropy/astropy-helpers#73 for other discussion on this last point).

@aphearin
Copy link
Contributor

aphearin commented Feb 2, 2015

@eteq - I'm not sure about this. Both on my local version of master, and on the local copy @duncandc has, neither of us get failures when running the test suite.

Do keep in mind this is not urgent.

@astrofrog
Copy link
Member

@eteq - we don't have control over conda versions of astropy, these are managed by the Continuum developers. They only build it with the latest Numpy version at the time. However, I've recently been investigating how to build conda packages and I've created a channel (astrofrog) from which you can install e.g. 1.0rc1 and if needed I could make it build 1.0 final (and other versions) with older versions of Numpy:

https://binstar.org/astrofrog/astropy/files

You can install these with conda install -c astrofrog astropy=1.0rc1

@eteq
Copy link
Member

eteq commented Feb 3, 2015

@astrofrog - ahh, gotcha - I thought we had been submitting some builds to them.

Also, as @embray just mentioned in astropy/astropy-helpers#73, he anticipated exactly this issue in astropy/astropy#3019, but it accidentally didn't end up in master. So that should fix the problem directly, and then once thats in someone can update halotools' astropy_helpers to close this.

@eteq eteq changed the title test failures in halotools test failures in halotools due to test runner incompatibility Feb 3, 2015
eteq pushed a commit to eteq/halotools that referenced this issue Jul 15, 2015
Don't call out to git if we're sure it will fail
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants