Add IE11 BrowserStack custom launcher.

[sculpt0r]

Add IE11 BrowserStack custom launcher

Closes #9

Closes #75

[sculpt0r]

Blocked by: #75

[github-actions]

It's been a while since we last heard from you. We are marking this pull request as stale due to inactivity. Please provide the requested feedback or the pull request will be closed after next 7 days.

[sculpt0r]

Seems that CI is green. However, locally few tests are down. I talked with @jacekbogdanski and he points me to ckeditor/ckeditor4-angular#119. However running single Chrome / Firefox stills fail on

SUMMARY:
✔ 43 tests completed
✖ 3 tests failed

FAILED TESTS:
 CKEditor Component
   when "component.value" changes to "foo"
     ✖ should call "instance.setData"
       Chrome 87.0.4280.66 (Linux x86_64)
     AssertError: expected setData to be called with arguments 
         at Object.fail (tests/component.js:30493:25)
         at failAssertion (tests/component.js:30450:20)
         at Object.assert.<computed> [as calledWith] (tests/component.js:30476:17)
         at Context.<anonymous> (tests/component.js:51840:63)

   when "component.readOnly" changes to "true"
     ✖ should call "instance.setReadOnly"
       Chrome 87.0.4280.66 (Linux x86_64)
     AssertError: expected setReadOnly to be called with arguments 
         at Object.fail (tests/component.js:30493:25)
         at failAssertion (tests/component.js:30450:20)
         at Object.assert.<computed> [as calledWith] (tests/component.js:30476:17)
         at Context.<anonymous> (tests/component.js:51840:63)

   when "component.readOnly" changes to "false"
     ✖ should call "instance.setReadOnly"
       Chrome 87.0.4280.66 (Linux x86_64)
     AssertError: expected setReadOnly to be called with arguments 
         at Object.fail (tests/component.js:30493:25)
         at failAssertion (tests/component.js:30450:20)
         at Object.assert.<computed> [as calledWith] (tests/component.js:30476:17)
         at Context.<anonymous> (tests/component.js:51840:63)

[sculpt0r]

Rebase onto newest master.

[sculpt0r]

After rebase, all local tests are green :)

[sculpt0r]

Summary:

• add BrowserStack IE11 launcher
• failing syntax on IE11 (also failing tests) - fixed with Incorrect files published ckeditor4-integrations-common#15. Now common integration distributes properly transpiled code
• still failing tests on IE11 fixed with 2f31df4:
  • Object.entries was in usage, but IE doesn't support it
  • Cannot use CKEDITOR.tools for Object.entries or reduce, because we are testing CKEditor namespace loading. There is cached Namespace before we delete window.CKEditor, but inner code of tools stills relay on global CKEditor variable, so it's hard to usage
  • Simplified object iteration: actually we are passing only simple object constructs ({ prop:value}) to fill element props. So there is no need to validate all specific cases like our tools are handling it.

Waiting for CI forever...
Finally, it's ready for review.

[github-actions]

It's been a while since we last heard from you. We are marking this pull request as stale due to inactivity. Please provide the requested feedback or the pull request will be closed after next 7 days.

[Dumluregn]

Looks good 👍🏻 Only one small thing left - in test code coverage I see that for one browser (I guess it's IE11 but I'm not sure) there is one line uncovered: 108. Could you examine if that's expected or the tests can be improved to check also this condition?

[sculpt0r]

Indeed - It's IE11...
And it's because watch for value is not fired. That's intentional, but I will check what were the reasons for such ignore flag.

ckeditor4-vue/tests/component.js

Lines 353 to 378 in 5df73b4

	[ {
	property: 'value',
	value: 'foo',
	spyOn: [ 'setData', true ],
	ignore: !!CKEditorNamespace.env.ie // (#4)
	}, {
	property: 'value',
	value: '',
	spyOn: [ 'setData', false ],
	ignore: !!CKEditorNamespace.env.ie // (#4)
	}, {
	property: 'readOnly',
	value: true,
	spyOn: [ 'setReadOnly', true ]
	}, {
	property: 'readOnly',
	value: false,
	spyOn: [ 'setReadOnly', true ]
	}, {
	property: 'readOnly',
	value: null,
	spyOn: [ 'setReadOnly', false ]
	} ].forEach( ( { property, value, spyOn: [ method, spyCalled ], ignore = false } ) => {
	if ( ignore ) {
	return;
	}

Here is the reason: #4 

There is also #4 (comment) that we should drop this flag. But now we are also testing with IE11.

IE11 also throws this error (MS admit it's bug).

So, could we leave it as it is?

[Dumluregn]

Thanks for checking the coverage thing. I've checked and the flag is still necessary in IE11, so as you wrote we have to leave it. Other than that - 👍🏻

Now let's wait for CI after rebase (I should've rebased before review but well).

[Dumluregn]

CI is green, so I'm merging the PR 👍🏻