[Webkit-unassigned] [Bug 26279] REGRESSION: typing in gmail pegs CPU

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=26279





------- Comment #12 from ojan at chromium.org  2009-06-12 12:48 PDT -------
OK. Here's what I know. 

The problem is that in InsertTextCommand::input we have a typingStyle that
actually has a property in it. The property looks to be a font-size:
(gdb) p typingStyle->m_properties[0]
$15 = (const WebCore::CSSProperty &) @0x28097810: {
  m_id = 1046, 
  m_shorthandID = 0, 
  m_important = false, 
  m_implicit = false, 
  m_value = {
    m_ptr = 0x1c1b7870
  }
}

My guess is that the reason that hitting enter or whatever fixes the perf
problem is that it clears this phantom typingStyle. FWIW, the
lastEditingCommand is a WebCore::TypingCommand::DeleteKey. So, it might be that
it got this weird typingStyle when I hit delete. A number of the comments in
the Chromium bug mention doing a delete right before hitting the perf bug.

That said, that the lastEditingCommand is a DeleteKey is not itself a bug. That
will happen anytime you hit delete and then start typing normal characters. The
difference is that the first time you type a character, it should apply the
typingStyle, so that next time you type a character, there is a typingStyle,
but it's length() is 0.

So, I see two problems:
1. The typingStyle is never getting applied.
2. I never actually changed font-size in gmail, so I have no idea where this
typing style came from. I guess gmail could be doing it, but that wouldn't
explain why it's never getting applied.


-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.