<html>
<head>
<base href="https://bugs.webkit.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - AX: Add a boundary value to AXTextStateChangeType"
href="https://bugs.webkit.org/show_bug.cgi?id=153085#c20">Comment # 20</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - AX: Add a boundary value to AXTextStateChangeType"
href="https://bugs.webkit.org/show_bug.cgi?id=153085">bug 153085</a>
from <span class="vcard"><a class="email" href="mailto:d_russell@apple.com" title="Doug Russell <d_russell@apple.com>"> <span class="fn">Doug Russell</span></a>
</span></b>
<pre>(In reply to <a href="show_bug.cgi?id=153085#c19">comment #19</a>)
<span class="quote">> Comment on <span class=""><a href="attachment.cgi?id=269010&action=diff" name="attach_269010" title="Patch">attachment 269010</a> <a href="attachment.cgi?id=269010&action=edit" title="Patch">[details]</a></span>
> Patch
>
> View in context:
> <a href="https://bugs.webkit.org/attachment.cgi?id=269010&action=review">https://bugs.webkit.org/attachment.cgi?id=269010&action=review</a>
>
> > Source/WebCore/ChangeLog:9
> > + an editable elements boundaries.
>
> elements -> element's</span >
will do
<span class="quote">>
> > Source/WebCore/accessibility/mac/AXObjectCacheMac.mm:385
> > + [userInfo setObject:[NSNumber numberWithBool:intent.selection.focusChange] forKey:NSAccessibilityTextSelectionChangedFocus];
>
> @(intent.selectin.focusChange)</span >
will do
<span class="quote">>
> > Source/WebCore/editing/FrameSelection.cpp:1110
> > + // FIXME: BIDI
>
> do you have a WK bug to reference</span >
I'll file one
<span class="quote">>
> > Source/WebCore/editing/FrameSelection.cpp:1164
> > + return (AXTextSelection) { intentDirection, intentGranularity, false };
>
> is the cast necessary (it may very well be, but i'm curious)</span >
This syntax was suggested by Darin in past reviews, I'll see if the cast can be dropped
<span class="quote">>
> > Source/WebCore/editing/VisiblePosition.h:57
> > + , m_isBoundary(false)
>
> m_isBoundary should be false by default. you may not need to initialize</span >
That is how I had it, trying this way to see if that's what was tripping up the windows build (it's one of the bigger things I changed between patches when windows stopped building)
<span class="quote">>
> > Source/WebCore/editing/VisiblePosition.h:60
> > + VisiblePosition(bool boundary) : m_affinity(VP_DEFAULT_AFFINITY), m_isBoundary(boundary) { };
>
> should this be "bool boundary = false" so you get the default behavior
> without changing code to be no?
> or should we make a static boundaryVisiblePosition() class method that
> returns the right object</span >
If you want boundary to be false you can just call VisiblePosition(). I can do a static if you prefer but it feels odd to not use a constructor.
<span class="quote">>
> i know people don't generally like passing around bools where the value is
> not clear what its for</span >
And we'd solve that with the static method?
<span class="quote">>
> > LayoutTests/accessibility/mac/selection-boundary-userinfo.html:8
> > +<div role="textbox" tabindex=0 id="textbox" contenteditable=true>
>
> can you also add a test with a input type="text"</span >
Yup. As is this patch doesn't post a boundary notification when arrowing up in a input (since they're single line controls), but I can file an enhancement for that.
<span class="quote">>
> > LayoutTests/accessibility/mac/selection-boundary-userinfo.html:30
> > +         shouldBe("results[0]", "AXTextStateChangeTypeSelectionMove");
>
> looks like this file has some tabs in it</span >
I'll fix that.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the assignee for the bug.</li>
</ul>
</body>
</html>