On 2012/08/21 13:29:46, Toon Verwaest wrote:
Can someone clarify the context of this CL? Michael is currently on
holidays
and
I do not really know what I'm looking at.
The proposed changes seem to alter the semantics of what the script
originally
intended to do. Is this intentional?
http://codereview.chromium.org/10827379/diff/1/tools/gen-postmortem-metadata.py
File tools/gen-postmortem-metadata.py (right):
http://codereview.chromium.org/10827379/diff/1/tools/gen-postmortem-metadata.py#newcode86
tools/gen-postmortem-metadata.py:86: 'value': 'TransitionArray::kFirstIndex'
},
All these changes change the semantics of the fields. Is this on purpose?
Do
we
care about transitions rather than descriptors now?
http://codereview.chromium.org/10827379/diff/1/tools/gen-postmortem-metadata.py#newcode113
tools/gen-postmortem-metadata.py:113: 'Map, instance_descriptors,
uintptr_t,
kTransitionsOrBackPointerOffset',
This field does not contain the instance_descriptors but rather either the
transition array or the back pointer. If it contains a transition array,
that
array will store the descriptors.
I think it's best if I retract this patchset and let someone with a vested
interest in postmortem support (*hint* Joyent *hint*) address it.
The issue is that node.js won't build with V8 HEAD right now but I can work
around that by disabling postmortem support for now.
http://codereview.chromium.org/10827379/
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
