The Inspector should support console.trace for Firebug parity.
<rdar://problem/5950861>
I read CallStackSidebarPane.js but it looks like you can't get the CallFrame unless the debugger is enabled and paused. >JavaScriptCallFrame* JavaScriptDebugServer::currentCallFrame() >{ > if (!m_paused) > return 0; > return m_currentCallFrame.get(); >} My initial thought was I could do it using arguments.callee.caller but I think ArgList doesn't have the callee info.
Created attachment 23840 [details] Screenshot
Created attachment 23841 [details] patch I'm not sure if it's ok but I called Machine::retrieveCaller recursively to get a stack trace. Seems to be working.
Comment on attachment 23841 [details] patch Why does trace() need to be custom? [Custom] void trace(); because it passes along the exec state?
Created attachment 23843 [details] patch Just replaced a tab.
That was my thinking when I did that. It doesn't compile when I don't do that. (In reply to comment #5) > (From update of attachment 23841 [details] [edit]) > Why does trace() need to be custom? [Custom] void trace(); because it passes > along the exec state? >
Comment on attachment 23843 [details] patch JSConsole::trace doesn't use the "arguments" argument. Either remove the name or the argument altogether. In Console::trace use this->page() instead of checking the m_frame, we made a helper to do this easier. Also please file a bug that says we should include and name (program) and (eval) nodes. Also please edit 19229 to also give better names than (anonymous function) when possible.
Created attachment 23862 [details] fixed patch I couldn't remove the argument so I just removed the name.
I file a bug https://bugs.webkit.org/show_bug.cgi?id=21157 (In reply to comment #8) > Also please file a bug that says we should include and name (program) and > (eval) nodes.
Landed in r37012.