[Webkit-unassigned] [Bug 153085] AX: Add a boundary value to AXTextStateChangeType

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=153085

--- Comment #33 from Doug Russell <d_russell at apple.com> ---
(In reply to comment #32)
> Comment on attachment 269085 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=269085&action=review
> 
> >>> Source/WebCore/editing/VisiblePosition.h:125
> >>> +    bool m_isBoundary = false;
> >> 
> >> I think it’s dangerous that now we have a type of VisiblePosition that won’t survive a round trop if broken into position and affinity and then made into a VIsiblePosition again. Do we have other options here rather than adding this state to VisiblePosition?
> > 
> > We could thread error state through the modifyMovingX logic to return why the VisiblePosition it's returning is NULL. It would mean touching a lot more code though.
> 
> One way to do it would be to add a new type that’s either a VisiblePosition
> or a boundary, rather than adding it to the original VisiblePosition type.
> You would need to touch more code, but just to change the type from
> VisiblePosition to the new type. Even though it makes the patch bigger, I
> think that would likely be better than introducing this concept into all
> other code using VisiblePosition for any other purpose. Could start as
> simply as a struct with a boolean and a VisiblePosition in it.

My latest patch threads a bool * through all the relevant logic. If this works for you I'm happy with it, but I could take the above approach if that's preferable.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160125/4b50e85d/attachment.html>