Drop events.

[Piotr Jasiun]

Part of #11437.

Editor should emit drag & drop events with data transfer facade.

[Piotrek Koszuliński]

We don't need editor#drop (as agreed in #11437) and we don't see yet need for editor#drag.

[Piotr Jasiun]

We decided that we need such facade.

[Piotr Jasiun]

Code, docs and tests done. Branch in based on #12168 because both tickets deeply modify tests structure so merge could be inconvenient. Changes in ​t/12169.

[Piotr Jasiun]

Rebased on newest #12168.

[Piotr Jasiun]

Rebased on newest #12168.

[Piotr Jasiun]

Rebased on major.

[Piotrek Koszuliński]

• ​https://github.com/cksource/ckeditor-dev/blob/2f98e75bbedd3f12beff015f044dc12212ecb3d9/plugins/clipboard/plugin.js#L1235 - why do you prevent the default there? The same is done inside the fireDragEvent.
• Run linter.
• ​https://github.com/cksource/ckeditor-dev/blob/2f98e75bbedd3f12beff015f044dc12212ecb3d9/plugins/clipboard/plugin.js#L1412-L1415 - mixed indentation.
• ​https://github.com/cksource/ckeditor-dev/blob/2f98e75bbedd3f12beff015f044dc12212ecb3d9/plugins/clipboard/plugin.js#L1401 - use dom.event#getTarget().
• "Facade for native editable.drop event." - don't mention editable here, becuase it's confusing.
• ​https://github.com/cksource/ckeditor-dev/blob/2f98e75bbedd3f12beff015f044dc12212ecb3d9/tests/plugins/clipboard/drop.js#L150 - don't do that - as I said many times, longer timeouts should be avoided. Use afterPaste plus, if required, 0ms timeout.

[Piotr Jasiun]

Fixed.

About:

• ​https://github.com/cksource/ckeditor-dev/blob/2f98e75bbedd3f12beff015f044dc12212ecb3d9/tests/plugins/clipboard/drop.js#L150 - don't do that - as I said many times, longer timeouts should be avoided. Use afterPaste plus, if required, 0ms timeout.

I cannot use 'afterPaste', because this is a case when there is no paste nor afterPaste events and I need to wait to check if no paste event were fired. But I decreased timeout from 300 to 0.

[Piotrek Koszuliński]

I cannot use 'afterPaste', because this is a case when there is no paste nor afterPaste events and I need to wait to check if no paste event were fired. But I decreased timeout from 300 to 0.

Hmm... Yeah, so this is more problematic. You can verify that 0ms is enough by breaking code and checking if test failed. If you used TDD you would be sure ;)

[Piotr Jasiun]

I set expectedPasteEventCount to 0 and all tests which fire paste event were red, so timeout 0 is not too small (paste was fired before event count was checked).

This should be checked again for #12173.

[Piotrek Koszuliński]

Replying to pjasiun:

I set expectedPasteEventCount to 0 and all tests which fire paste event were red, so timeout 0 is not too small (paste was fired before event count was checked).

I rather meant breaking plugin's code, not tests. You may not predict that in case in which event is cancelled paste event may be fired differently. That's why you should break the place where a decision is made whether to continue firing events process. That's why TDD is much better, because you first see the test failing.

[Piotr Jasiun]

I change if the change in the code causes red test and it does.

Majorized with hash: ​git:1bb7369.