That's great to hear. Thanks for working on this!
On Wed, Feb 12, 2014 at 3:25 PM, Jim Blandy <[email protected]> wrote: > As hinted at previously, the DevTools group would like to introduce a new > representation of saved (not live) JS stacks to SpiderMonkey. Our immediate > need is to record the stacks at which JS objects are allocated, for our > memory profiling tools, but we'd like to introduce something general enough > that Error, and other sorts of captured stacks, could switch to using it. > > The representation that Nick Fitzgerald and I have prototyped: > > - shares common old-end tails of captured stacks; > > - retains principal information for each frame, so that stack traces can > omit privileged frames in unprivileged contexts, but show full information > in privileged contexts (unlike Error's present behavior, which entirely > forgets frames that the capturing context shouldn't see); > > - will support the optimizations described before on this list: > https://groups.google.com/forum/#!msg/mozilla.dev.tech. > js-engine.internals/crDZ1vU7hkY/uYg3xxD552oJ <https://groups.google.com/ > forum/#%21msg/mozilla.dev.tech.js-engine.internals/ > crDZ1vU7hkY/uYg3xxD552oJ> > > We're discussing our prototype design in bug 972045, but since we want > this to be part of SpiderMonkey, we can only proceed with a design that the > team would look forward to. > > The patch soon to be attached there will simply add the stack > representation, along with a TestingFunctions.cpp primitive that captures > stacks, and unit tests, for the time being. > > _______________________________________________ > dev-tech-js-engine-internals mailing list > [email protected] > https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals > _______________________________________________ dev-tech-js-engine-internals mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
