Add logging of callbacks in prof-lazy mode.

Add logging of callbacks in prof-lazy mode.

This is needed to show calls to DOM in CPU profiles. I can think
of a better approach like adding specific functions into V8 API
for explicitly providing callback names and modifying bindings codegen
appropriately. My plan is as follows:
 - submit this CL;
 - implement anything I need to process log data and display DOM 
   calls in profiles;
 - think again about adding specific functions and modifying bindings
   codegen.

BUG=http://code.google.com/p/chromium/issues/detail?id=27613

Committed: http://code.google.com/p/v8/source/detail?r=3340

[Søren Thygesen Gjesse, 2009-11-20 08:04:02]

LGTM, however we need to iterate on this to make it more general. As far as I
can see changes made through the API is not handled after profiling has started.

http://codereview.chromium.org/402100/diff/1001/2001
File src/log.cc (right):

http://codereview.chromium.org/402100/diff/1001/2001#newcode650
Line 650: // TODO(mnaganov): Don't use a bogus size.
Please file a bug and change this to TODO(bugnumber).

http://codereview.chromium.org/402100/diff/1001/2001#newcode1239
Line 1239: void VisitFTI(Object* o) {
Looks as if this has some assumptions to how the object model has been build. It
is OK for now, but it needs to be more general, perhaps with information from
the embedding application on how to name the callback functions.

http://codereview.chromium.org/402100/diff/1001/2001#newcode1257
Line 1257: // Properties are triples: [property name, entry point, attributes].
No mentioning of Chromium.

http://codereview.chromium.org/402100/diff/1001/2002
File src/log.h (right):

http://codereview.chromium.org/402100/diff/1001/2002#newcode117
Line 117: V(BOUND_TAG,                      "Bound",                  "bo")     
 \
I don't see why this is called bound object. A more natural name would be just
callback.

http://codereview.chromium.org/402100/diff/1001/2002#newcode204
Line 204: // Emits a code event for a callback of a bound object.
"callback of a bound object" -> "callback function"

http://codereview.chromium.org/402100/diff/1001/2002#newcode205
Line 205: static void BoundObjectCallbackEvent(const char* class_name,
Just CallbackEvent or CallbackFunctionEvent.

http://codereview.chromium.org/402100/diff/1001/2002#newcode275
Line 275: // Used for logging Chromium bindings callback entry points
Please don't mention Chromium in the V8 code. This is really not Chromium
specific.

http://codereview.chromium.org/402100/diff/1001/2002#newcode278
Line 278: static void LogChromiumBindingsCallbacks();
LogChromiumBindingsCallbacks -> LogCallbackFunctions

[mnaganov (inactive), 2009-11-20 08:46:27]

Thanks!

Your point about dynamic updates of callback entries is really good, I forgot
about it.

http://codereview.chromium.org/402100/diff/1001/2001
File src/log.cc (right):

http://codereview.chromium.org/402100/diff/1001/2001#newcode650
Line 650: // TODO(mnaganov): Don't use a bogus size.
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> Please file a bug and change this to TODO(bugnumber).

Just removing comment and settings this to zero.

http://codereview.chromium.org/402100/diff/1001/2001#newcode1239
Line 1239: void VisitFTI(Object* o) {
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> Looks as if this has some assumptions to how the object model has been build.
It
> is OK for now, but it needs to be more general, perhaps with information from
> the embedding application on how to name the callback functions.

Yes, this is what I'm talking about in CL description. Currently I just looked
at how Chromium bindings fill this information. I haven't yet grasped a full
understanding on how the whole interaction between an application and VM is
supposed to be, so this solution is temporary. After I'll implement needed
pieces to use callbacks info and to display their entries in profiles, I'll
return back to the entry points info discovering problem.

http://codereview.chromium.org/402100/diff/1001/2001#newcode1257
Line 1257: // Properties are triples: [property name, entry point, attributes].
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> No mentioning of Chromium.

Right. This is actually about how Temlate::Set works.

http://codereview.chromium.org/402100/diff/1001/2002
File src/log.h (right):

http://codereview.chromium.org/402100/diff/1001/2002#newcode117
Line 117: V(BOUND_TAG,                      "Bound",                  "bo")     
 \
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> I don't see why this is called bound object. A more natural name would be just
> callback.

Done.

http://codereview.chromium.org/402100/diff/1001/2002#newcode204
Line 204: // Emits a code event for a callback of a bound object.
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> "callback of a bound object" -> "callback function"

Done.

http://codereview.chromium.org/402100/diff/1001/2002#newcode205
Line 205: static void BoundObjectCallbackEvent(const char* class_name,
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> Just CallbackEvent or CallbackFunctionEvent.

Done. "CallbackEvent"

http://codereview.chromium.org/402100/diff/1001/2002#newcode275
Line 275: // Used for logging Chromium bindings callback entry points
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> Please don't mention Chromium in the V8 code. This is really not Chromium
> specific.

Done.

http://codereview.chromium.org/402100/diff/1001/2002#newcode278
Line 278: static void LogChromiumBindingsCallbacks();
On 2009/11/20 08:04:02, Søren Gjesse wrote:
> LogChromiumBindingsCallbacks -> LogCallbackFunctions

I think "LogCallbacks" will be enough and is consistent with "CallbackEvent".