If you throw a load of handles at it, it'll probably push all the older handles out, but you'll still be left with a load of handles knocking around.
I'd appreciate if this collectgarbage thing was official too (I also dislike having to edit the v8 Source). This was mentioned in that other thread: > If you want to trigger a GC, you could look at > the GC extension. For example, when running with --expose-gc, a gc() > method will be installed on the global object. Which I expect would have the exact same effect as CollectAllGarbage.. (it just feels wrong to be calling a js function to do this). garry On Oct 2, 2:03 am, Pete Gontier <[EMAIL PROTECTED]> wrote: > More specifically: > > http://groups.google.com/group/v8-users/msg/47d3cc9c9a3e9d0a > > On Sep 27, 2008, at 12:30 PM, Garry wrote: > > > Thanks - for future reference for anyone looking for the same thing.. > > I ``fixed'' this by exposing v8::internal::Heap::CollectAllGarbage() > > by adding it to the v8:V8 interface in api.cc > > > void V8::CollectAllGarbage() > > { > > i::Heap::CollectAllGarbage(); > > } > > > Calling this at choice places (right before exit, after disposing a > > context etc) solves the problem completely for me. > > This is just what the doctor ordered. > > I'm not terribly enamored of the idea of altering V8 sources because > it becomes a maintenance hassle to stay in synch, and anyway I > ultimately want to link to the inevitable official Ubuntu binary > shared library which won't have my changes, but one does what one must. > > Another technique discussed in the thread was to allocate (and leak, > from a C++ perspective) a bunch of handles to nudge V8 into deciding > for itself to collect garbage. I suppose this could be done without > altering V8 sources, but then the question becomes: how many handles > is enough? The thread says it's around 4000... for now... > > Pete Gontier <http://pete.gontier.org/> > > On Oct 1, 2008, at 8:10 AM, deanm wrote: > > >http://groups.google.com/group/v8-users/browse_thread/thread/9effc949... > > > On Oct 1, 5:07 pm, Pete Gontier <[EMAIL PROTECTED]> wrote: > >> I would like to be able to test a WeakReferenceCallback. > > >> It would seem that in order to do that I need to induce garbage > >> collection at a predictable time. > > >> So far, I don't see a way to do that. Ideas? > > >> Pete Gontier <http://pete.gontier.org/> --~--~---------~--~----~------------~-------~--~----~ v8-users mailing list [email protected] http://groups.google.com/group/v8-users -~----------~----~----~----~------~----~------~--~---
