In documents that use CSS :before and :after, the text for the inserted frames is not exposed in accessible tree.

Louie or Jessie, do you see this problem or the problem in bug 242594 using at-poke?

Suggested fix: When walking accessible tree, make sure to create accessible objects for anonymous frames. if (childFrame && (childFrame->IsGeneratedContentFrame()) { frame = childFrame; childFrame = frame->GetFirstChild(nsnull); } } I will be able to do this after the patch for 242589 goes in.

The frame tree looks something like: Block(h1) Inline(h1) Text "Section: " Text "Fruits"

Aron, I have tested the patch on Linux (based on provided testcase). Below are some info about my testing. Before applying the patch (using the latest trunk code) 1. All texts wrapped by <h1> (with style "before:...") have no corresponding at-object, which means that <h1> (with style "before:...") nodes aren't accessible through at-poke. 2. All texts wrapped by <p> are fine. The content of "After:..." is correctly exposed. After applying the patch 1. <h1> (with style "before:...") still can't be exposed. 2. Through at-poke, there are 4 additional "at-object", which are all for the last <p> node -- "Life is short, eat this first".

The differences between how we expose text in ATK and MSAA are very different. I need to get my Linux build going so I can understand the difference better. This patch definitely improves things for MSAA. Can we #ifndef MOZ_ACCESSIBILITY_ATK on the frame walking parts for now, so we can get this in and make it work for ATK later?

Louie, Can you try this patch with and without the |#ifndef MOZ_ACCESSIBILITY_ATK| in nsAccessibleTreeWalker::UpdateFrame() ? I think an extra if () statement messed us up there. If it still doesn't work let's check it in with the |#ifndef| and we can leave the bug open for the ATK part of it.

Comment on attachment 150260 [details] [diff] [review] Fix for MSAA now. Will fix for ATK later. Seeking r=Louie Zhao

Comment on attachment 150260 [details] [diff] [review] Fix for MSAA now. Will fix for ATK later. Since it has no harm to Linux build, you can fix it for MSAA first.

Comment on attachment 150260 [details] [diff] [review] Fix for MSAA now. Will fix for ATK later. >Index: src/base/nsAccessible.cpp >=================================================================== >RCS file: /cvsroot/mozilla/accessible/src/base/nsAccessible.cpp,v >retrieving revision 1.101 >diff -u -r1.101 nsAccessible.cpp >--- src/base/nsAccessible.cpp 5 Jun 2004 17:56:59 -0000 1.101 >+++ src/base/nsAccessible.cpp 8 Jun 2004 12:11:59 -0000 >@@ -277,6 +277,15 @@ > // This node has been shut down > return NS_ERROR_FAILURE; > } >+ if (!mParent) { >+ nsCOMPtr<nsIAccessible> parent; >+ GetParent(getter_AddRefs(parent)); >+ if (parent) { >+ PRInt32 numChildren; >+ GetChildCount(&numChildren); // Make sure we cache all of the children >+ } >+ } >+ This part of code doesn't do any really work. Right?

(In reply to comment #11) > >+ if (!mParent) { > >+ nsCOMPtr<nsIAccessible> parent; > >+ GetParent(getter_AddRefs(parent)); > >+ if (parent) { > >+ PRInt32 numChildren; > >+ GetChildCount(&numChildren); // Make sure we cache all of the > This part of code doesn't do any really work. Right? You're right. It's supposed to say |parent->GetChildCount(&numChildren)| not just GetChildCount(&numChildren)| This code makes GetNextSibling() always use the cached mNextSibling. If there's no cached parent yet (e.g. if an accessible is created from an event in the middle of the document), and GetNextSibling() is called on that, then this code will make sure we have the cached mNextSibling. Calling parent->GetChildCount(&numChildren) makes sure that all of the parent's children are cached via mParent::mFirstChild and mNextSibling on each of the children.

Comment on attachment 150260 [details] [diff] [review] Fix for MSAA now. Will fix for ATK later. sr=Henry Pls make the related change (parent->GetChildCount) before checking in.

Checking in src/base/nsAccessNode.h; /cvsroot/mozilla/accessible/src/base/nsAccessNode.h,v <-- nsAccessNode.h new revision: 1.12; previous revision: 1.11 done Checking in src/base/nsAccessible.cpp; /cvsroot/mozilla/accessible/src/base/nsAccessible.cpp,v <-- nsAccessible.cpp new revision: 1.103; previous revision: 1.102 done Checking in src/base/nsAccessibleTreeWalker.cpp; /cvsroot/mozilla/accessible/src/base/nsAccessibleTreeWalker.cpp,v <-- nsAccessibleTreeWalker.cpp new revision: 1.5; previous revision: 1.4 done Checking in src/base/nsAccessibleTreeWalker.h; /cvsroot/mozilla/accessible/src/base/nsAccessibleTreeWalker.h,v <-- nsAccessibleTreeWalker.h new revision: 1.4; previous revision: 1.3 done Checking in src/html/nsHTMLFormControlAccessible.cpp; /cvsroot/mozilla/accessible/src/html/nsHTMLFormControlAccessible.cpp,v <-- nsHTMLFormControlAccessible.cpp new revision: 1.51; previous revision: 1.50 done Checking in src/html/nsHTMLTextAccessible.cpp; /cvsroot/mozilla/accessible/src/html/nsHTMLTextAccessible.cpp,v <-- nsHTMLTextAccessible.cpp new revision: 1.33; previous revision: 1.32 done Checking in src/xul/nsXULFormControlAccessible.cpp; /cvsroot/mozilla/accessible/src/xul/nsXULFormControlAccessible.cpp,v <-- nsXULFormControlAccessible.cpp new revision: 1.37; previous revision: 1.36 done

I think this caused bug 248378 -- it fits the regression window. Either we're doing something wrong or the VC98 compiler is.

Backing out. I believe this caused bug 248378 becuase of compiler differences. Checking in src/base/nsAccessNode.h; /cvsroot/mozilla/accessible/src/base/nsAccessNode.h,v <-- nsAccessNode.h new revision: 1.13; previous revision: 1.12 done Checking in src/base/nsAccessible.cpp; /cvsroot/mozilla/accessible/src/base/nsAccessible.cpp,v <-- nsAccessible.cpp new revision: 1.107; previous revision: 1.106 done Checking in src/base/nsAccessibleTreeWalker.cpp; /cvsroot/mozilla/accessible/src/base/nsAccessibleTreeWalker.cpp,v <-- nsAccessibleTreeWalker.cpp new revision: 1.6; previous revision: 1.5 done Checking in src/base/nsAccessibleTreeWalker.h; /cvsroot/mozilla/accessible/src/base/nsAccessibleTreeWalker.h,v <-- nsAccessibleTreeWalker.h new revision: 1.5; previous revision: 1.4 done Checking in src/html/nsHTMLFormControlAccessible.cpp; /cvsroot/mozilla/accessible/src/html/nsHTMLFormControlAccessible.cpp,v <-- nsHTMLFormControlAccessible.cpp new revision: 1.53; previous revision: 1.52 done Checking in src/html/nsHTMLTextAccessible.cpp; /cvsroot/mozilla/accessible/src/html/nsHTMLTextAccessible.cpp,v <-- nsHTMLTextAccessible.cpp new revision: 1.34; previous revision: 1.33 done Checking in src/xul/nsXULFormControlAccessible.cpp; /cvsroot/mozilla/accessible/src/xul/nsXULFormControlAccessible.cpp,v <-- nsXULFormControlAccessible.cpp new revision: 1.39; previous revision: 1.38 done

Backing this out fixes bug 248378 in Mozilla 1.8a2 Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a2) Gecko/20040625 We need to find out what part of this fix doesn't work for VC98

Rechecked in without bad line: mState.domNode->GetNextSibling(getter_AddRefs(mState.domNode)); getter_AddRefs() nulls out that arg. Checking in src/base/nsAccessNode.h; /cvsroot/mozilla/accessible/src/base/nsAccessNode.h,v <-- nsAccessNode.h new revision: 1.14; previous revision: 1.13 done Checking in src/base/nsAccessible.cpp; /cvsroot/mozilla/accessible/src/base/nsAccessible.cpp,v <-- nsAccessible.cpp new revision: 1.108; previous revision: 1.107 done Checking in src/base/nsAccessibleTreeWalker.cpp; /cvsroot/mozilla/accessible/src/base/nsAccessibleTreeWalker.cpp,v <-- nsAccessibleTreeWalker.cpp new revision: 1.7; previous revision: 1.6 done Checking in src/base/nsAccessibleTreeWalker.h; /cvsroot/mozilla/accessible/src/base/nsAccessibleTreeWalker.h,v <-- nsAccessibleTreeWalker.h new revision: 1.6; previous revision: 1.5 done Checking in src/html/nsHTMLFormControlAccessible.cpp; /cvsroot/mozilla/accessible/src/html/nsHTMLFormControlAccessible.cpp,v <-- nsHTMLFormControlAccessible.cpp new revision: 1.54; previous revision: 1.53 done Checking in src/html/nsHTMLTextAccessible.cpp; /cvsroot/mozilla/accessible/src/html/nsHTMLTextAccessible.cpp,v <-- nsHTMLTextAccessible.cpp new revision: 1.35; previous revision: 1.34 done Checking in src/xul/nsXULFormControlAccessible.cpp; /cvsroot/mozilla/accessible/src/xul/nsXULFormControlAccessible.cpp,v <-- nsXULFormControlAccessible.cpp new revision: 1.40; previous revision: 1.39 done

Fixed, without causing any regressions this time.

Comment on attachment 150260 [details] [diff] [review] Fix for MSAA now. Will fix for ATK later. >+ mState.domNode = mState.frame? do_QueryInterface(mState.frame->GetContent()): nsnull; This only works because of bug 253941 allowing the conversion to if (mState.frame) mState.domNode = do_QueryInterface(mState.frame->GetContent()); else mState.domNode = do_QueryInterface(nsnull);