16999 2012-05-08 19:31:00 +0000 Clarify JS type for empty close reason 2012-05-10 07:22:43 +0000 1 1 1 Unclassified WebAppsWG WebSocket API (editor: Ian Hickson) unspecified All All RESOLVED INVALID P2 minor --- 1 jduell.mcbugs ian jonas mike Ms2ger public-webapps public-webapps-bugzilla oldest_to_newest 67494 0 jduell.mcbugs 2012-05-08 19:31:00 +0000 I don't believe the Web socket spec is clear on this: If there is no close 'reason' passed from the server in the protocol (which on a network frame level is different than sending an empty Unicode string), should we set the close event's 'reason' field to 'undefined', or to an empty string? Right now Firefox is setting it to an empty string. Logically, I feel it ought to be 'undefined'. But I don't see much harm in it being an empty string. Either way the spec should be clear about it. 67503 1 jonas 2012-05-08 21:04:09 +0000 Rather than undefined I think null would be more appropriate. But I'd be ok either way. 67522 2 Ms2ger 2012-05-09 17:07:31 +0000 > The reason attribute must return the value it was initialized to. > When the object is created, this attribute must be initialized to > empty string. What's unclear? (Except for the missing 'the') 67524 3 jduell.mcbugs 2012-05-09 20:59:12 +0000 > attribute must be initialized to empty string. Ah, didn't see that. So the spec is clear then--'reason' will never be null/undefined. 67542 4 zcorpan 2012-05-10 07:22:43 +0000 I think this is the relevant part "whose reason attribute is initialized to the WebSocket connection close reason decoded as UTF-8, with error handling, and dispatch the event at the WebSocket object. [WSP]" and then WSP says "If there is no such data in the Close control frame, _The WebSocket Connection Close Reason_ is the empty string." But the result is the same. :-)