Hi Vincent, Which version of V8 are you using? Did you build it yourself or you use prebuilt binary?
We had similar issue with WrapFrameId being optimized away with value range propagation optimization in GCC 4.4. We fixed it in http://code.google.com/p/v8/source/detail?r=5278 -- Vyacheslav Egorov On Wed, Nov 10, 2010 at 9:30 AM, Vincent Han <[email protected]> wrote: > "scope" and some other protocol occur forced termination in Android. > > My analysis is like below. > > First of all, a major cause is "JavaScriptFrameIterator > frame_it(id);" > in Runtime_GetScopeCount function. > The reason of above is "id" after doing below. > - StackFrame::Id id = UnwrapFrameId(wrapped_id); > A parameter of Runtime_GetScopeCount function is "args". > → args[0]: this.break_id_, args[1]: this.frameId() (in Mirror- > debug.js) > When args[1] is zero(frameId is zero), wrapped_id is zero. > - CONVERT_CHECKED(Smi, wrapped_id, args[1]); > Then id become zero after UnwrapFrameId(wrapped_id). > I think above routine got some problem. > > Now, look at Runtime_GetFrameDetails function. > There is id got some values. > Then, make frame_id using id through WrapFrameId function. > - frame_id value is zero. > And set frame_id to details[kFrameDetailsFrameIdIndex]. > I thiink args[1] of Runtime_GetScopeCount using that frame_id(zero). > > Is it a right processing? > Or some protocols have restrictions before requesting. > > Please help me. > > -- > v8-users mailing list > [email protected] > http://groups.google.com/group/v8-users > -- v8-users mailing list [email protected] http://groups.google.com/group/v8-users
