On Fri, Jun 15, 2012 at 1:37 AM, <[email protected]> wrote: > On 2012/05/22 10:39:53, Erik Corry wrote: >> >> In order to avoid problems with the incremental marker you could ensure >> that a >> GC is completed immediately before doing the replacement. See the code in >> HeapSnapshotsCollection::FindHeapObjectById for an example. > > >> LGTM > > > Thank you for advice. > > Actually I'm fine with patching dead objects amongst others. Is it the only > thing I should be caring about? In this case probably it's not required to > do > extra GC.
It's not just about patching dead objects, it's also because you have to notify the GC with a write barrier when you write pointers on the heap, so that the GC can keep track of where there are pointers in case some of the objects move in a compacting GC. > > http://codereview.chromium.org/10332101/ -- Erik Corry, Software Engineer Google Denmark ApS - Frederiksborggade 20B, 1 sal, 1360 København K - Denmark - CVR nr. 28 86 69 84 -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
