Steve McClure
Steve McClure
This sounds like a potentially serious issue (false positive), and I'm all for defensive-coding, but IMHO a lack of repro steps gives me pause to action anything and feel confident...
Does this pre-hashing match the existing unique hashes during the run? The existing hash mechanism modifies clashing test-names (white whitespace suffixes) to get unique IDs. I want to ensure these...
What's to become of the `QUnit.test` syntax for either "flat" (non-nested) modules, or if no modules are used at all? Not to scope creep, but is the proposed concept moving...
It's true that as-written, the repro steps would fail with enough tests above it, or with a very short window; this is accounted for (but was left out of the...
That all sounds great, thanks for digging deep on this. I made a one-liner CSS change in our extension and the masses of tests looked good again. There were a...
I took a pass at untangling the start/load logic: https://github.com/qunitjs/qunit/compare/master...smcclure15:untangle-load-logic Most of that endeavor was for my own understanding, getting my head around what the valid sequences were among these...
The following analysis starts with a trivial passing test: ```js QUnit.test( "example", assert => { assert.true( true ); } ); ``` and adds various combinations of calls to `QUnit.load`, `QUnit.start`,...