- From: Andy Davies <dajdavies@gmail.com>
- Date: Thu, 7 Feb 2013 13:05:35 +0000
- To: "public-web-perf@w3.org" <public-web-perf@w3.org>
- Message-ID: <CABbusA+wC9a0iB_EmDA7u1vRjSKLcEjrdbkwo=fgFVBWCXkrcg@mail.gmail.com>
Is there any active work being carried out on the HAR spec? I'm aware the latest editors draft is here: https://dvcs.w3.org/hg/webperf/raw-file/tip/specs/HAR/Overview.html but that's not been updated since August I've a couple of observations and a question based on the implementation of a HAR generator I'm currently working on. 1. Timings assume a linear path I'm sure this has been brought up before but the timings section assumes the various events immediately follow each other. With DNS prefetch and pre-emptive TCP connections this is no longer true for some browsers. 2. dataURIs The spec doesn't mention how dataURIs should be handled i.e. should they be listed in the HAR on not. HARs exported from Chrome include them but if the entries are supposed to represent network activity this seems incorrect as their download is included in other entries. As Resource Timing excludes dataURIs so I'm guess HAR should exclude them too. 3. Calculating send, wait and receive In the implementation I'm working on I only have access to requestStart, responseStart & responseEnd, so I can either calculate send or wait but not both. What would people recommend, I'm currently using wait = responseStart - requestStart? Thanks Andy
Received on Thursday, 7 February 2013 13:06:06 UTC