M2: Odd problems/interactions between xhr tests and JsonItemStore tests

[guest]

After the xhr tests were enabled in base, the JsonItemStore? tests started getting random failures. Often it would be the first test of JsonItemStore? would fail (fetch_all). Sometimes it would be a bunch of the tests. The only commonality seems to be it's all tests where JsonItemStore? is using xhrGet() to load the JSON item data.

So, it looks like there is some sort of state/timing/something issue in the xhrGet code that occurs when you run a lot of the tests together and mix in sync versus non-sync xhr. The problem so far manifests itself on Firefox 1.5 (I haven't had a chance to try other browsers yet. Here is what has been tried and the results:

Extract revision of 0.9: 8035:

1.) JsonItemStore? tests run with all other tests in-place.

JsonItemStore? seems to get random failures in tests. Sometimes its Deferred not firing. Others I get nulls when I shouldn't (implying sync mode xhr wasn't working right).

2.) JsonItemStore? tests run without xhr tests in the mix

Tests run fine, no errors, no failures, in the JsonItemStore? tests.

3.) JsonItemStore? tests run before the xhr tests.

Tests run fine, no errors, no failures, in the JsonItemStore? tests.

[alex]

(In [8080]) adds a delay between test execution. Refs #2721. May fixt it, but it'll need more testing and Jared's verification.

[guest]

The delay seems to fix it. I havent' seen any odd xhr test/JsonItemStore test failures anymore.

[(none)]

Milestone 0.9M2 deleted