- From: Noah Mendelsohn <nrm@arcanedomain.com>
- Date: Fri, 09 Aug 2013 12:04:21 -0400
- To: Jer Noble <jer.noble@apple.com>
- CC: Srikumar Karaikudi Subramanian <srikumarks@gmail.com>, "robert@ocallahan.org" <robert@ocallahan.org>, "K. Gadd" <kg@luminance.org>, Chris Wilson <cwilso@google.com>, Marcus Geelnard <mage@opera.com>, Alex Russell <slightlyoff@google.com>, Anne van Kesteren <annevk@annevk.nl>, Olivier Thereaux <Olivier.Thereaux@bbc.co.uk>, "public-audio@w3.org" <public-audio@w3.org>, "www-tag@w3.org List" <www-tag@w3.org>
On 8/7/2013 11:56 PM, Jer Noble wrote: > Solutions which require the former usually also require the latter two, Certainly not in all systems, but I'll assume it's indeed the case here. There are certainly systems in which one of the reasons for copying is so that input from an I/O buffer (which might be fixed and reused) is copied to a fixed application buffer that is also reused. In such cases: lots of copying, no malloc or GC churn. I'd say this is very typical of Unix I/O, for example (granting that Unix doesn't do GC anyway). > which may explain why they've all been lumped together under "copying". OK. Noah
Received on Friday, 9 August 2013 16:04:42 UTC