- From: Harald Alvestrand <harald@alvestrand.no>
- Date: Mon, 09 Jun 2014 20:53:40 +0200
- To: Iñaki Baz Castillo <ibc@aliax.net>, "public-webrtc@w3.org" <public-webrtc@w3.org>
On 06/09/2014 08:34 PM, Iñaki Baz Castillo wrote: > Hi, > > I've realized, by measuring DataChannels performance in a JS > application, that in certain circumstances (network conditions?) > DataChannel.send() blocks for 1ms, 2ms or even 10ms. Tested in Chrome > 35. > > This is obviously catastrophic for a JS application since it blocks > the JS loop, and it is even worse given that existing periodic timers > (window.setInterval) get their next tick delayed the number of > milliseconds it took to DataChannel.send() to return. Please file a Chrome bug, preferably with a script that reproduces the behaviour. > > I assume this cannot be valid and thus it is an implementation bug. Am > I right? or must I assume that send() may, not just fail, but also > block the JS execution loop? > > Thanks a lot. > -- Surveillance is pervasive. Go Dark.
Received on Monday, 9 June 2014 18:54:15 UTC