- From: Anne van Kesteren <annevk@opera.com>
- Date: Tue, 19 Sep 2006 10:55:19 -0000
- To: Óscar Toledo G. <uno@biyubi.com>
- Cc: "Web APIs WG (public)" <public-webapi@w3.org>
On Mon, 17 Jul 2006 17:14:40 -0000, Óscar Toledo G. <uno@biyubi.com> wrote: > Also generating exceptions on reading status and statusText, > is a little awkward, I think that zero and empty string > are more sensible selections. There's some amount of legacy content and implementations we have to deal with... > Other suggestions, the send method can include a second > argument to indicate the desired timeout in milliseconds, > to generate an event, would be better than setTimeout, as > the event would not be generated if the answer comes first > or the onerror handler exists and it is invoked first. Not for this version. We are considering something like that for XMLHttpRequest 2. > The onerror event is necessary to detect a failed request > (by example: server not responding HTTP connection), so is > worth to have it into the draft, along with onload. Other > alternative would be a special status 5 (Request error) > for readyState. Error handling for XMLHttpRequest 1 comes down to going to readyState 4 and having fields set to their initial values which is what most UAs do today. XMLHttpRequest 2 will deal with this. > In the responseXML description, the draft could say > something like: > "An UA with a HTML implementation, may support > additionally text/html, the Document node may have > properties related to DOM HTML.". XMLHttpRequest 2. -- Anne van Kesteren <http://annevankesteren.nl/> <http://www.opera.com/>
Received on Tuesday, 19 September 2006 10:56:05 UTC