- From: Robin Berjon <robin@berjon.com>
- Date: Wed, 23 May 2012 16:19:28 +0200
- To: Marcos Caceres <w3c@marcosc.com>
- Cc: Anssi Kostiainen <anssi.kostiainen@nokia.com>, "public-device-apis@w3.org" <public-device-apis@w3.org>
On May 23, 2012, at 12:07 , Marcos Caceres wrote: > On Wednesday, 23 May 2012 at 08:34, Anssi Kostiainen wrote: >> (To make sure on-one gets confused, I'll have to mention the above draft is an old one despite ReSpec.js showing the current date.) > > I was going to bring this up because it's causing me a bit of confusion - it means I need to be checking the HG repo also for changes (which is ok), but it would be really nice if the spec linked to the repo. > > About the date problem, in other working groups we commonly have: > > Overview.src.html (unprocessed) and Overview.html (processed) online. Yeah but that only works if you produce a processed version with every commit, which is a hassle. The issue here is mostly that there were two documents when there should have been only one. > That avoid the problem of having the wrong date show up…. Actually, that's not where the problem of the wrong date showing up comes from. ReSpec is a good citizen, unless a specific date is given to it (e.g. to produce a dated snapshot) it uses document.lastModified to figure out when the document was last updated. The problem is that the hg server is misconfigured (or the Mercurial web bridge is broken) and always returns right now as the last modified date — which is a bug. > could also determine the last modified date from a HTTP head request, though I've tried doing that before and ran into cross browser issues. There are occasional issues with Opera in using document.lastModified, but that's really a bug in Opera. And a HEAD won't help you with the hg server alas. -- Robin Berjon - http://berjon.com/ - @robinberjon
Received on Wednesday, 23 May 2012 14:20:02 UTC