Summary of potentially-relevant matrix discussion:

• The frame tree for the example below has a one block frame root with a unique inline frame parent for each text leaf (the code element becomes several frames)
• The accessible we generate for the code element has only one frame (the primary frame, whatever that is)
• We use TransformRect to compute the offset between subtree-spanning continuations -- transforming the continuation to the hypertext's frame. The hypertext's frame (in this instance) is not necessarily an ancestor of the continuation's frame, since the primary frame for the code element is only the parent to one (the first?) text leaf.

Here's a test case:

/**
 * Test character bounds in an intervening inline element with non-br line breaks
 */
addAccessibleTask(
  `
  <style>
    @font-face {
      font-family: Ahem;
      src: url(${CURRENT_CONTENT_DIR}e10s/fonts/Ahem.sjs);
    }
    pre {
      font: 20px/20px Ahem;
    }
  </style>
  <pre><code id="t" role="group">XX
XXX
XX
X</pre>`,
  async function(browser, docAcc) {
    await testChar(docAcc, browser, "t", 0);
    await testChar(docAcc, browser, "t", 3);
    await testChar(docAcc, browser, "t", 7);
    await testChar(docAcc, browser, "t", 10);
  },
  {
    chrome: true,
    topLevel: !isWinNoCache,
    iframe: !isWinNoCache,
  }
);

I'm wondering whether we can fix this by calling TransformRect with the continuation frame and the primary text frame rather than the ancestor frame. That is:

// ... No need for nearestAccAncestorFrame
if (prevTextFrame->GetParent() != currTextFrame->GetParent()) {
  contRect = frameRect;
  nsLayoutUtils::TransformRect(currTextFrame, frame, contRect);
}

As with bug 1809695, this is broken in the parent process (LocalAccessible) as well. I think we may have a JS test harness bug here rather than a C++ bounds bug.

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:morgan, could you have a look please?
If you still have some work to do, you can add an action "Plan Changes" in Phabricator.
For more information, please visit auto_nag documentation.

https://hg.mozilla.org/mozilla-central/rev/8ab29a76ed86