Bug 206015

Summary: Text manipulation controller should not observe changes in new replacement elements
Product: WebKit Reporter: Wenson Hsieh <wenson_hsieh>
Component: WebCore Misc.Assignee: Wenson Hsieh <wenson_hsieh>
Status: RESOLVED FIXED    
Severity: Normal CC: bdakin, commit-queue, ews-watchlist, megan_gardner, mifenton, rniwa, thorton, webkit-bug-importer
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch none

Description Wenson Hsieh 2020-01-09 10:25:11 PST 
<rdar://problem/58353667>
Comment 1 Wenson Hsieh 2020-01-09 13:35:59 PST 
Created attachment 387267 [details]
Patch
Comment 2 WebKit Commit Bot 2020-01-09 16:40:21 PST 
Comment on attachment 387267 [details]
Patch

Clearing flags on attachment: 387267

Committed r254314: <https://trac.webkit.org/changeset/254314>
Comment 3 WebKit Commit Bot 2020-01-09 16:40:23 PST 
All reviewed patches have been landed.  Closing bug.
Comment 4 Ryosuke Niwa 2020-01-13 19:06:32 PST 
Comment on attachment 387267 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=387267&action=review

> Source/WebCore/editing/TextManipulationController.cpp:407
> +    m_document->eventLoop().queueTask(TaskSource::InternalAsyncTask, [weakThis = makeWeakPtr(*this)] {

Hm... this is probably not sufficient delay. We need to schedule this when the first call to didCreateRendererForElement happens.
Anyway, I'm gonna add a generic mechanism to track already translated content so it's sort of a moot point.
Comment 5 Ryosuke Niwa 2020-01-13 19:06:52 PST 
(In reply to Ryosuke Niwa from comment #4)
> Comment on attachment 387267 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=387267&action=review
> 
> > Source/WebCore/editing/TextManipulationController.cpp:407
> > +    m_document->eventLoop().queueTask(TaskSource::InternalAsyncTask, [weakThis = makeWeakPtr(*this)] {
> 
> Hm... this is probably not sufficient delay. We need to schedule this when
> the first call to didCreateRendererForElement happens.
> Anyway, I'm gonna add a generic mechanism to track already translated
> content so it's sort of a moot point.

This is because there is no guarantee that the style recalc happens by the next task in the event loop is executed.