Opened 12 years ago
Closed 11 years ago
#10837 closed Task (worksforme)
Image2: EventsBuffer integration
| Reported by: | Olek Nowodziński | Owned by: | Olek Nowodziński | 
|---|---|---|---|
| Priority: | Normal | Milestone: | |
| Component: | General | Version: | 4.3 Beta | 
| Keywords: | Cc: | 
Description
The plugin should integrate CKEDITOR.tools.eventsBuffer to improve the performance of mousemove callback (Firefox, IE).
Attachments (3)
Change History (8)
comment:1 Changed 12 years ago by
| Status: | new → confirmed | 
|---|
Changed 12 years ago by
| Attachment: | img2ebuff_chrome.png added | 
|---|
Changed 12 years ago by
| Attachment: | img2ebuff_ie9.png added | 
|---|
Changed 12 years ago by
| Attachment: | img2ebuff_ie8.png added | 
|---|
comment:2 Changed 12 years ago by
| Owner: | set to Olek Nowodziński | 
|---|---|
| Status: | confirmed → review | 
comment:3 Changed 12 years ago by
| Milestone: | CKEditor 4.3 | 
|---|
I don't see a significant improvement too and, moreover, resizing works now smoothly even on IE9 on VM. Actually... it works there better than on Firefox running on my main OS.
Anyway, AFAIR it didn't work so well when I reported this issue. The current performance is good, so I'm removing a milestone, but we can keep this ticket if users will report us that something is wrong.
BTW. Avoid this kind of cache:
+ var $ = evt.data.$; + + screenX = $.screenX; + screenY = $.screenY;
You use the variable twice, so you may lost more time on defining it than on accessing.
comment:4 Changed 11 years ago by
| Status: | review → assigned | 
|---|
comment:5 Changed 11 years ago by
| Resolution: | → worksforme | 
|---|---|
| Status: | assigned → closed | 

Created branch t/10837 with
eventsBufferwhile resizing.The basic problem is that to keep UX at the high level, the maximum buffer interval for
mousemovecan be ~30ms. Otherwise the experience becomes choppy.Such interval, however, makes
eventsBufferintegration pointless because benefits (expressed by CPU load) are questionable and the code gets complicated. For example:In Chrome a slight improvement can be observed:
In IE9 there's no improvement:
In IE8 there's a measurable drop of the load:
I'm leaving this ticket as something to be discussed.