This seems like a false dichotomy. I'd rather ask: do we want the scheduling library code to deliberately avoid resource exhaustion or do we want the Servo code to deliberately avoid resource exhaustion?
Kevin On Tue, Dec 24, 2013 at 11:12 AM, Carter Schonwald <carter.schonw...@gmail.com> wrote: > Indeed! And the producer can always have their own private backlog. There is > nothing preventing each producer having it's one internal backlog queue > that's not bounded. But such behavior should be deliberate rather than > accidental. > > > On Tuesday, December 24, 2013, Daniel Micay wrote: >> >> On Tue, Dec 24, 2013 at 1:56 PM, Patrick Walton <pwal...@mozilla.com> >> wrote: >> > I think we need them to handle the Web semantics though. We can't just >> > drop >> > input events. >> >> It would be better to use a bounded queue and report an error when it >> gets to megabytes or gigabytes in size than going out-of-memory >> though. A user will prefer a clean exit of the tab or browser compared >> to it becoming locked because the events aren't being promptly >> handled or it causing a significant system performance degradation >> ending with the browser being killed. >> _______________________________________________ >> Rust-dev mailing list >> Rust-dev@mozilla.org >> https://mail.mozilla.org/listinfo/rust-dev > > > _______________________________________________ > Rust-dev mailing list > Rust-dev@mozilla.org > https://mail.mozilla.org/listinfo/rust-dev > _______________________________________________ Rust-dev mailing list Rust-dev@mozilla.org https://mail.mozilla.org/listinfo/rust-dev
