Re: progress events and error

Anne van Kesteren wrote:
> On Sat, 21 Apr 2007 08:15:50 +1000, Jonas Sicking <jonas@sicking.cc> wrote:
>> Looking at the new progress event spec it looks great. I don't 
>> understand why it requires that 'loadstart' and 'progress' be fired 
>> before error though. For example if the requested uri can't even be 
>> parsed as a real uri it seems sensible to fire an error event right 
>> away and abort without doing any further processing.
> 
> Wouldn't that depend on the situation? With XMLHttpRequest for instance 
> if the URI can't be parsed you'd get a SYNTAX_ERR and never arrive here. 
> The error event is only thrown for network errors. (Similarly, the load 
> event is dispatched if everything went ok.)

Good point. This is probably going to be true in most specs actually. 
I'm still not convinced it's a good idea to mandate that these events 
are fired first unless there is a real usecase for it. See below.

> I think you do want at least 
> one loadstart and progress event before that event though.

Why? ('that event' means the error event here, right?)

/ Jonas

Received on Monday, 23 April 2007 12:45:07 UTC