Skip to content

Conversation

dakrauth
Copy link

Fix up settings and skip certain tests that we probably don't care about.

Fix up settings and skip certain tests that we probably don't care about.
@dakrauth dakrauth requested review from defbyte, fdintino and obswork July 10, 2020 04:18
@@ -68,8 +70,10 @@ def test_post_process(self):
self.assertTrue(('screen.css', 'screen.css', True) in processed_files)
self.assertTrue(('scripts.js', 'scripts.js', True) in processed_files)

@unittest.skip("don't care")
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Doesn't this test a condition we'll soon have, with the s3 boto storage (which doesn't have a path method implemented). Was this disabled because it's failing, or for some other reason?

@@ -249,7 +249,7 @@ def test_cssmin(self):
self._test_compressor('pipeline.compressors.cssmin.CSSMinCompressor',
'css', 'pipeline/compressors/cssmin.css')

@skipUnless(settings.HAS_NODE, "requires node")
@skip("don't care")
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We have this compressor re-implemented and modified in our code base, so if it's failing, we ought to make sure that our rewrite isn't affected by the same issue.

Copy link
Author

@dakrauth dakrauth Jul 10, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

All of the failures that I've skipped fail in all Django versions tested (1.11-2.2). I'm not aware if -- or when -- we've ever run this test suite, and as much as I agree with your concerns, I think we should add a similar test as this and the others in Ollie, so that we can catch and fix any pertinent failures.

Copy link
Member

@fdintino fdintino Jul 10, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Or, if it's passing in travis but failing locally, I can probably help with the configuration. I actually wrote these tests; they were merged into upstream a while back (jazzband#531)

Copy link
Author

@dakrauth dakrauth Jul 10, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might just be a better-documentation thing that I can flesh out (meaning, the local tox failures), so if you have any suggestions beyond what I added to the README, I can give that a try.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also, as for the Java tests that I skipped, I was assuming that we didn't need any of those. If that is incorrect, I guess I can install the Java runtime, but I'd really hate to go down that rabbit hole locally.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe that you need to run npm install from the repo root for some of these tests to pass. Beyond that, you might need to rebase jazzband/django-pipeline@3a0e3cc and jazzband#644. If those don't do the trick, then I think it's not worth the trouble—in which case this PR LGTM

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just saw your comment now about the java tests: we don't need these. They could perhaps be wrapped in a skipUnless for settings.HAS_JAVA, but that's a very nitpicky concern and I think it's also fine as is.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I definitely did npm install. Unless I was too quick about it, when I stepped through this particular test failure, it seemed to me that the custom test storage class PipelineNoPathStorage was the cause of the error.

Copy link
Member

@fdintino fdintino left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On further reflection, I think the sourcemap changes on this branch would need to be rebased with the commits from jazzband#490, which is a mostly-abandoned upstream PR that might differ from the sourcemap implementation on atl/1.6.x in subtle ways. So I have no more objections to disabling the tests in this PR.

@dakrauth dakrauth merged commit d3e9101 into atl/1.6.x Jul 10, 2020
@dakrauth dakrauth deleted the atl/django2x branch July 10, 2020 07:02
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

Successfully merging this pull request may close these issues.

2 participants