End of buffer error after test failure

[AshtonKem]

Steps to reproduce

1. Find a testing namespace
2. Eval NS
3. Run all tests (success)
4. Run individual test (success)
5. Edit a test to fail.
6. Run all tests (failure correctly reported)
7. Run failing test individually (End of buffer error)
8. Run all tests (End of buffer error).

After repeating this process, the only way to return to successful tests is to restart Cider.

[bbatsov]

I'm guessing this happens on 0.9.0, right?

[bbatsov]

I can't reproduce this with a simple project created with lein new and the following tests in it:

(deftest a-test
  (testing "something"
    (is (= 1 1))))

(deftest b-test
  (testing "something else"
    (is (= 1 1))))

[AshtonKem]

Correct.

Ashton

Sent from my iPhone

On Jun 20, 2015, at 12:45 AM, Bozhidar Batsov notifications@github.com wrote:

I'm guessing this happens on 0.9.0, right?

—
Reply to this email directly or view it on GitHub.

[bbatsov]

Can you cook some minimal project in which the problem can be reproduced?

[weejulius]

steps to reproduce:

load from file and run tests, it works correctly.
and then load the test form again and run tests, you will see the error.

by the way , the souce code and test code is in the same file in my case.

[bbatsov]

@weejulius Yep, that did it. The problem is that when you're doing interactive evaluation the location metadata for the form is not set and this causes this error. There's nothing I can do about the location metadata and I'd advise everyone affected by this to ask @cemerick to merge the patch attached to http://dev.clojure.org/jira/browse/NREPL-59

I can, however, figure out some way to show a nicer message.

Here's some background info about the problem and why this was changed in CIDER 0.9 - #1088.