Support stepping out for recursive functions

Support stepping out for recursive functions.Simply flooding JS function from the calling stack frame with one shot breakpoints is not enough to support step out action in all cases since the function on top of the stack may be turn recursive and we may end up flooding itself. To overcome this a pointer to the stack frame where the debugger should be invoked after stepping out is strored in the debugger.Chromium bug:http://code.google.com/p/chromium/issues/detail?id=17967

Committed: http://code.google.com/p/v8/source/detail?r=2850

[Søren Thygesen Gjesse, 2009-09-08 14:06:02]

LGTM

Nice debugger change!

http://codereview.chromium.org/200041/diff/8/1008
File src/debug.cc (right):

http://codereview.chromium.org/200041/diff/8/1008#newcode869
Line 869: // out unless real breakpoint is hit.
out -> out to is reached,

http://codereview.chromium.org/200041/diff/8/1008#newcode1115
Line 1115: thread_local_.step_count_ = step_count;
Please set the thread local step count to 0 here and add a comment about how
step count is interpreted when performing step out.

http://codereview.chromium.org/200041/diff/8/1008#newcode1199
Line 1199: } else if (it.IsExit()) {
Please change this to plain else and assert it.IsExit().

http://codereview.chromium.org/200041/diff/8/1008#newcode1216
Line 1216: thread_local_.step_count_ = 0;
Remove this when setting thread local step count above.

[yurys, 2009-09-08 14:16:30]

http://codereview.chromium.org/200041/diff/8/1008
File src/debug.cc (right):

http://codereview.chromium.org/200041/diff/8/1008#newcode869
Line 869: // out unless real breakpoint is hit.
On 2009/09/08 14:06:02, Søren Gjesse wrote:
> out -> out to is reached,

Done.

http://codereview.chromium.org/200041/diff/8/1008#newcode1115
Line 1115: thread_local_.step_count_ = step_count;
On 2009/09/08 14:06:02, Søren Gjesse wrote:
> Please set the thread local step count to 0 here and add a comment about how
> step count is interpreted when performing step out.

Done.

http://codereview.chromium.org/200041/diff/8/1008#newcode1199
Line 1199: } else if (it.IsExit()) {
On 2009/09/08 14:06:02, Søren Gjesse wrote:
> Please change this to plain else and assert it.IsExit().

Done.

http://codereview.chromium.org/200041/diff/8/1008#newcode1216
Line 1216: thread_local_.step_count_ = 0;
On 2009/09/08 14:06:02, Søren Gjesse wrote:
> Remove this when setting thread local step count above.

Done.

[yurys, 2009-09-08 14:58:18]

Please look at the change one more time. I've added Debug::ClearStepping(); call
in Runtime_PrepareStep function so that there is no current step action when
Debug::PrepareStep is called. Otherwise stepping state can get messed up by
several consecutive calls to Runtime_PrepareStep like in
mjsunit/debug-continue.js test.

I could move Debug::ClearStepping at the very top of Debug::PrepareStep but this
would make Debug::Break code less clear.

[Søren Thygesen Gjesse, 2009-09-08 19:33:58]

LGTM

Maybe add some asserts to ActivateStepIn/ActivateStepOut that no other stepping
activity is active to help to keep a consistent state.

http://codereview.chromium.org/200041/diff/10/16
File src/runtime.cc (right):

http://codereview.chromium.org/200041/diff/10/16#newcode6936
Line 6936: // args[2]: number of times to perform the step
Update comment to indicate that for step out it is the number of frames to step
down.

[yurys, 2009-09-09 08:34:41]

http://codereview.chromium.org/200041/diff/10/16
File src/runtime.cc (right):

http://codereview.chromium.org/200041/diff/10/16#newcode6936
Line 6936: // args[2]: number of times to perform the step
On 2009/09/08 19:33:58, Søren Gjesse wrote:
> Update comment to indicate that for step out it is the number of frames to
step
> down.

Done.