20768 2013-01-25 03:50:49 +0000 TCP/IP disconnect detection issue AFAIK, only Chrome(WebKit?) uses TCP keepalive with 45 sec. interval; anyway, it is not easy to use native EventSource in other browser: there are no tcp keepali [...] 2013-04-30 17:52:23 +0000 1 1 1 Unclassified WHATWG HTML unspecified Other other RESOLVED WONTFIX http://www.whatwg.org/specs/web-apps/current-work/#top P3 normal Unsorted 1 contributor ian ian mike vic99999 contributor oldest_to_newest 82128 0 contributor 2013-01-25 03:50:49 +0000 Specification: http://www.w3.org/TR/2012/WD-eventsource-20120426/ Multipage: http://www.whatwg.org/C#top Complete: http://www.whatwg.org/c#top Comment: TCP/IP disconnect detection issue AFAIK, only Chrome(WebKit?) uses TCP keepalive with 45 sec. interval; anyway, it is not easy to use native EventSource in other browser: there are no tcp keepalive to detect broken connections; there are no way to pass "lastEventId" to a new EventSource object; Posted from: 83.146.108.107 User agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.17 (KHTML, like Gecko) Chrome/24.0.1312.52 Safari/537.17 82129 1 vic99999 2013-01-25 03:53:22 +0000 see https://bugzilla.mozilla.org/show_bug.cgi?id=444328#c1 82410 2 ian 2013-01-31 17:40:05 +0000 You can pass lastEventId manually to the other side by encoding it in the URL, if you really need to. I would recommend contacting the browser vendors, though. This seems to be an issue at a different layer than the EventSource spec itself. 82495 3 vic99999 2013-02-02 03:40:55 +0000 https://www.w3.org/Bugs/Public/show_bug.cgi?id=13104 i see, that for WebSockets the note about "user agent may send ping..." was added. in case of EventSource: "User agent should use TCP-keepalives to detect disconnects" although, "heartbeat timeout" option will be better, because the default timeout may be too big. 84952 4 ian 2013-03-25 22:30:00 +0000 It doesn't seem like the Web page would be in a better position to determine what the right heartbeat value should be than the browser. 87025 5 ian 2013-04-30 17:52:23 +0000 WONTFIX per comment 4. Please re-open if you can elaborate on why it makes more sense for pages to control this than browsers; I don't really understand how an author could know the right values to use.