Closed
Description
Let's start from creating a set of manual tests on which we'll be able to reliably and quickly test typical performance-oriented scenarios.
Some ideas for those scenarios:
- Loading large HTML data with a lot of attributes.
- Coming from an editor (so, only elements and structures that the editor could've produced).
- Coming from external sources (e.g. other website), so with structures, elements and attributes that the editor doesn't understand and many inline styles (so we can test the new style normalization).
- An editor with a long content (same cases as above) but focused on testing features inside the editor (so the content should be already preloaded).
- Inserting content (like pasting) into an editor.
- Also, 3-4 data sets.
- The editor content should be diverse as well so you can test various scenarios (pasting into a table, image caption, middle of something, etc.)
- Loading a couple of editors at once. We expect to see a linear time growth, so we could test just one, but let's have ~3-5 as at those points the load times start to be noticeable.
Other considerations:
- The action to be tested should be triggered by a button (instead of happening automatically).
- Use rich setups of editor plugins (so we cover diverse types of those).
Subtasks:
- Test with empty editor and
setData
button (Added editor setData manual performance test #6103) - Test with content initialization (after pressing a button) (Added editor initialization manual performance test #6124)
- Generate content (small, medium, large)
- Base text (content that will serve as a reference for other small, medium, large variations) (Performance manual tests: proposed the content for small, medium and long tests #6147)
- Semantic content
- Attributes/Inline (known and unknown to editor), combined CSS styles (Performance manual tests: inline css variant of fixtures #6148)
- Editor with many features enabled (content preloaded) (Performance manual tests: added an editor with most of our plugins enabled #6149)
- Loading multiple editors at once (Added a manual performance test for initializing multiple editors at once #6144)
- Review performance, check for any potential bottlenecks (Performance research #5880 (comment))