1386342 - Add a special case to Selection::Collapse() to detect the no-op case where a selection is being collapsed onto itself and do nothing in that case

Status: RESOLVED INVALID

(Core :: DOM: Core & HTML, enhancement)

Comment 1

[(no longer active)]

Attachment: Add a special case to Selection::Collapse() to detect the no-op case where a selection is being collapsed onto itself and do nothing in that case

Comment 2

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 8892545 [details] [diff] [review]
Add a special case to Selection::Collapse() to detect the no-op case where a selection is being collapsed onto itself and do nothing in that case

r=me

Comment 3

[Olli Pettay [:smaug][[e-mail]]]

Why is the patch right when there are several ranges in selection?
And isn't collapse suppose to replace the range object from JS point of view.
See "// we should reuse it for reducing the recreation cost."

Comment 4

[(no longer active)]

(In reply to Olli Pettay [:smaug] from comment #3)
> Why is the patch right when there are several ranges in selection?

Oh yeah, that's a bug.  We don't have a test to catch it unfortunately.  :-/  

> And isn't collapse suppose to replace the range object from JS point of view.
> See "// we should reuse it for reducing the recreation cost."

Does this matter for nsFrameSelection objects we create for nsTextEditorState?  The selection objects inside those aren't exposed to content JS.

Comment 5

[Olli Pettay [:smaug][[e-mail]]]

It probably doesn't matter for internal frame selection objects. But then, I'd expect refcnt to be 1 usually in that case.

Comment 6

[(no longer active)]

I don't think the patch here is much useful after all, I'm going to close this bug.