Next time, check to make sure gmail properly picked up to reply to the list instead of just to Jason...
On Fri, Sep 30, 2011 at 1:43 PM, Nick Hurley <[email protected]> wrote: > > > On Fri, Sep 30, 2011 at 1:25 PM, Jason Duell <[email protected]>wrote: > >> On 09/30/2011 12:47 PM, Patrick McManus wrote: >> >>> I thought this is why nick did necko net. >>> >> >> Yes, sorry, we'd need both the access pattern described above AND >> necko-net to get this to work. The key is to have some test running that >> has better hit rate with LRU on (and a small cache), and has network times >> greater than disk access. There's all sort of refinements that could be >> made, but even something stupid would be progress as far as regression >> coverage goes. >> > > Unless I screwed something up horribly, it should be entirely (and easily) > possible to take a manifest in the mode of Geoff's modified tp4m manifest, > run it on a machine running web-page-replay (to save the page loads), and > take that manifest + generated archive and shove them in a running neckonet, > pair those with a low-bandwidth, high-latency network configuration, and > we've got what we want. > _______________________________________________ dev-tech-network mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-network
