Fix 'step in' after live edit stack manipulation

Fix 'step in' after live edit stack manipulation

Committed: http://code.google.com/p/v8/source/detail?r=5150

[Peter Rybin, 2010-07-27 17:49:32]

Hi Mads

It's stack manipulations again. This time I think it should be more modest
change.

Peter

[Peter Rybin, 2010-07-28 12:57:29]

Mads,

I guess Soren may have returned from his vacation.
Do you think it's better to pass this CL to him?

p.s. Hi Soren! :)

Peter

[Søren Thygesen Gjesse, 2010-07-28 15:29:09]

LGTM

http://codereview.chromium.org/3029033/diff/8001/9003
File src/debug.cc (right):

http://codereview.chromium.org/3029033/diff/8001/9003#newcode1304
src/debug.cc:1304: // to function it is goint to restart.
function it -> the function which
goint to -> going to be

http://codereview.chromium.org/3029033/diff/8001/9003#newcode1311
src/debug.cc:1311: // If it's CallFunction stub ensure target function is
compiled and flood
This comment should be inside the else if body.

http://codereview.chromium.org/3029033/diff/8001/9003#newcode1788
src/debug.cc:1788: thread_local_.restarter_frame_function_pointer_ =
restarter_frame_function_pointer;
Long line.

http://codereview.chromium.org/3029033/diff/8001/9004
File src/debug.h (right):

http://codereview.chromium.org/3029033/diff/8001/9004#newcode428
src/debug.h:428: //  b. being compatible with regular stack structure for
various stack iterators.
Long line.

http://codereview.chromium.org/3029033/diff/8001/9004#newcode429
src/debug.h:429: // Returns address of stack allocated pointer to restarted
function,
Stack allocated here means stored in the JavaScript stack?

http://codereview.chromium.org/3029033/diff/8001/9004#newcode430
src/debug.h:430: // the value that is called 'restarter_frame_function_pointer'.
Maybe mention where this return value is used.

http://codereview.chromium.org/3029033/diff/8001/9004#newcode979
src/debug.h:979: return
reinterpret_cast<Address>(Debug::restarter_frame_function_pointer_address());
Long line.

http://codereview.chromium.org/3029033/diff/8001/9005
File src/ia32/debug-ia32.cc (right):

http://codereview.chromium.org/3029033/diff/8001/9005#newcode299
src/ia32/debug-ia32.cc:299: Address fp = bottom_js_frame->fp();
Would it be possible to use some of the existing constants or (frames-ia32.h)
here instead of 1, 2, 3, 4 and 5? Or possibly add more constants?

http://codereview.chromium.org/3029033/diff/8001/9006
File src/liveedit.cc (right):

http://codereview.chromium.org/3029033/diff/8001/9006#newcode1242
src/liveedit.cc:1242: *restarter_frame_function_pointer =
Debug::SetUpFrameDropperFrame(bottom_js_frame, code);
Long line.

[Peter Rybin, 2010-07-28 17:03:15]

I'm sorry for those long lines, I got used to check them on web page, but the
limit was reset to 100 characters.

http://codereview.chromium.org/3029033/diff/8001/9003
File src/debug.cc (right):

http://codereview.chromium.org/3029033/diff/8001/9003#newcode1304
src/debug.cc:1304: // to function it is goint to restart.
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> function it -> the function which
> goint to -> going to be

Done.

http://codereview.chromium.org/3029033/diff/8001/9003#newcode1311
src/debug.cc:1311: // If it's CallFunction stub ensure target function is
compiled and flood
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> This comment should be inside the else if body.

Done.

http://codereview.chromium.org/3029033/diff/8001/9003#newcode1788
src/debug.cc:1788: thread_local_.restarter_frame_function_pointer_ =
restarter_frame_function_pointer;
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> Long line.

Done.

http://codereview.chromium.org/3029033/diff/8001/9004
File src/debug.h (right):

http://codereview.chromium.org/3029033/diff/8001/9004#newcode428
src/debug.h:428: //  b. being compatible with regular stack structure for
various stack iterators.
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> Long line.

Done.

http://codereview.chromium.org/3029033/diff/8001/9004#newcode429
src/debug.h:429: // Returns address of stack allocated pointer to restarted
function,
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> Stack allocated here means stored in the JavaScript stack?

Yes.

http://codereview.chromium.org/3029033/diff/8001/9004#newcode430
src/debug.h:430: // the value that is called 'restarter_frame_function_pointer'.
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> Maybe mention where this return value is used.

Done.

http://codereview.chromium.org/3029033/diff/8001/9004#newcode979
src/debug.h:979: return
reinterpret_cast<Address>(Debug::restarter_frame_function_pointer_address());
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> Long line.

Done.

http://codereview.chromium.org/3029033/diff/8001/9005
File src/ia32/debug-ia32.cc (right):

http://codereview.chromium.org/3029033/diff/8001/9005#newcode299
src/ia32/debug-ia32.cc:299: Address fp = bottom_js_frame->fp();
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> Would it be possible to use some of the existing constants or (frames-ia32.h)
> here instead of 1, 2, 3, 4 and 5? Or possibly add more constants?

I would like to reimplement this function and probably make it
platform-independent.

http://codereview.chromium.org/3029033/diff/8001/9006
File src/liveedit.cc (right):

http://codereview.chromium.org/3029033/diff/8001/9006#newcode1242
src/liveedit.cc:1242: *restarter_frame_function_pointer =
Debug::SetUpFrameDropperFrame(bottom_js_frame, code);
On 2010/07/28 15:29:10, Søren Gjesse wrote:
> Long line.

Done.