On Fri, Oct 17, 2014 at 1:31 PM, Manish Goregaokar
wrote:
> I'm pretty sure the must_root lint stops this from happening, except in a
> couple of cases which are pretty convoluted and not likely to come up.
>
Those are the worst cases, cause they'll still happen, you can just never
reproduce the
> Potential issues I see with the design:
>
> - there are Root, Handle and MutableHandle types that work on SM GC
> things, but there's no equivalent of the rooting analysis (that I could
> see) that prevents you from keeping unrooted GC thing pointers live over an
> operation that could GC.
>
>
Jon Coppeard provided some useful feedback on the SpiderMonkey upgrade
PR that probably shouldn't just languish in my inbox.
Forwarded Message
Subject:Re: Reviewing the SpiderMonkey upgrade for Servo
Date: Fri, 10 Oct 2014 04:45:34 -0700 (PDT)
From: Jon Coppeard
T
3 matches
Mail list logo
