This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
An InputMethodContext can be obtained from any HTMLElement, but in some cases it is unrealistic that all elements in a page have different contexts. This point is ambiguous in the spec. For example, consider the following HTML: <html> <body> <input type=text> <textarea> ... </textarea> <div contenteditable=true> <span id=span1> ... </span> <span id=span2> ... </span> </div> </body> </text> It makes sense <input> and <textarea> can have different context, but if span1 and span2 can have different contexts, - once cursor enters any of the span, switch to the active context? - if selection encompasses 2 <span>s, what context should be activated? This may complicate the implementation too much. Probably a guideline under one contentable element, all DOM elements under the element usually share the same context makes sense for most implementations.
Yes, it’s our thinking that if the developer queries for input method context on child elements inside the same contentEditable element, he’ll get the same context as on the parent contentEditable element.
I'll clear this in the next working draft spec.