| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
function serialization in Safari.
* Add azure pipelines CI.
* Publish test results.
* Enable coveralls
* Use Safari for OSX browser tests.
* Adding missing lib files to karma (only Safari complained).
* Rename with dot so it gets better placed in order.
* Use ci instead of install, so we stick with package-lock.json
* Use npm test instead of mocha directly. It just needed a double -- escape.
* Move DISPLAY to Linux only, with value set just once.
* Use variables for DISPLAY, browser tests and coveralls.
* Simplify steps with a browser name variable.
* Run junit reporter on CI only.
* DISPLAY should be already an environment variable.
* Use vmImage for consistency with other OS.
* Don't use verbose npm@1 task.
* No need to setTimeout, let's `done` on the callback.
* Don't resume twice.
Check queue length on callback for safety, setTimeout may be delayed.
* Avoid time dependency, act on events instead.
* Avoid dependency on timing to final assertion.
* Rewrite test to avoid setTimeout
* Fix formatting.
* Make diff more significant (125+50 ~= 200)
* Make diff more significant to avoid race conditions.
* Wait a bit more so default 5 retries did happen.
* Make delays shorter to avoid "Timeout of 250ms exceeded"
* Shorten delay to avoid 200ms test timeout in browser tests.
* Shorten timeout to mitigate 200ms test timeout.
* Add more diff on delays to enforce test behavior.
* Make delays more significant to avoid race conditions.
* Stop using setTimeout and rely on events to push new items, so order is enforced.
* Fix async function serialization in Safari.
* Add more delay so 4 never completes before 3.
* Ensure 3 never completes before 4.
* Use drain instead of task count.
Trying to ensure done is called from Windows browser tests that fail frequently.
* Ensure 2 starts before 3 arrives.Avoid expected 'process 2 3' to equal 'process 2'
* Add retries to browser tests.
* Don't fail all on Mac & Windows browser tests.
* Ensure 3 happens before 2, it fails sparely on OSX.
AssertionError: expected [ 1, 2, 3 ] to deeply equal [ 1, 3, 2 ]
* Avoid unnecessary multilines.
* Use Edge for Windows browser tests.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* cancelable foreach
* cancelable waterfall
* cancellable auto
* fix lint
* fix tests
* cancelable whilst/until/during/forever
* fix waterfall test. It WILL get there
* docs
* use rest params instead of slice
* clean up internals
* remove property func
* clarify uses of createTester
* happy path async funtions in asyncify
* stop using arguments
* DLL to class
* moar arrows
* fix merge issues
* remove forOwn
* moar arrows
* fix merge mistake
* even more arrows, what can stop him
* mo more fn.apply(null,...)
* remove more spurious uses of apply
* update lint config
* just when you thought there couldn't possibly be more arrows
* use eslint:recommended
* even less uses or aguments
* get rid of prototype cuteness
* fix concat tests
* fix more tests
|
