- From: Ian Hickson <ian@hixie.ch>
- Date: Sun, 26 Apr 2009 21:01:53 +0000 (UTC)
- To: John Fallows <john.fallows@kaazing.com>, Anne van Kesteren <annevk@opera.com>
- Cc: "public-html-comments@w3.org" <public-html-comments@w3.org>
On Tue, 17 Mar 2009, John Fallows wrote: > > As discussed on IRC: > > <john_fallows> Hixie: the latest EventSource API for SSE looks great - > question: is it now illegal to send "open" and "error" event types in > an SSE stream because eventSource.addEventListener("open", ...) and > eventSource.addEventListener("error", ...) would be ambiguous if this > is permitted? > <Hixie> john_fallows: it's actually been impossible to set the event > name for some time now > <john_fallows> Hixie: okay, missed that removal from the spec, so we > are locked down as message always? > <Hixie> john_fallows: yeah, i believe so > <john_fallows> ok, great - that allieviates any concern, thanks. :) > <john_fallows> Hixie: am i missing something? > http://dev.w3.org/html5/spec/Overview.html#processField seems to still > have the ability to override the event name if the event field is > present. > <Hixie> john_fallows: huh, go figure. > <Hixie> john_fallows: i guess you _can_ send arbitrary events then > <Hixie> john_fallows: i wonder why we still support that > <john_fallows> Hixie: IMHO this is probably no longer necessary, and > as pointed out previously it is also ambiguous for "open" and "error" > events. > <Hixie> john_fallows: yeah > <Hixie> john_fallows: can you send mail to the list saying that i > should remove it? > <john_fallows> Hixie: would you like me to send email to remind you to > remove it? > <john_fallows> :-) > <Hixie> yes please :-) On Wed, 18 Mar 2009, Anne van Kesteren wrote: > > I think I requested to keep that in and I still think it is useful. It > allows for handling the incoming messages in separate ways without > having to parse the data field first. So you can have different handlers > for different types of incoming messages. It should be fairly trivial to > support, too. "error" and "open" would implement different interfaces so > they can be distinguished in that way. I suppose you could make them > fail to dispatch if you want to be careful though... Based on Anne's comments I haven't changed the spec. The feature seems mostly harmless; even faking 'error' events doesn't seem particularily threatening. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Sunday, 26 April 2009 21:02:30 UTC