[Webkit-unassigned] [Bug 68833] Discard event data not followed by an empty line before eof when parsing an event-stream

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=68833


Per-Erik Brodin <per-erik.brodin at ericsson.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #108722|0                           |1
        is obsolete|                            |
 Attachment #108722|review?                     |
               Flag|                            |
 Attachment #108858|                            |review?
               Flag|                            |




--- Comment #7 from Per-Erik Brodin <per-erik.brodin at ericsson.com>  2011-09-27 09:16:06 PST ---
Created an attachment (id=108858)
 --> (https://bugs.webkit.org/attachment.cgi?id=108858&action=review)
updated patch

(In reply to comment #6)
> (From update of attachment 108722 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=108722&action=review
> 
> I see, that explains why we need two member variables.
> 
> Would "m_currentlyParsedEventId" and "m_latestCompleteEventId" work? You could also null out m_currentlyParsedEventId after transferring its value to m_latestCompleteEventId.

I would really like to leave m_lastEventId as it is since that's a name used in the spec and the lastEventId property on MessageEvent and the Last-Event-ID request header are set from it, but I've renamed m_eventId to m_currentlyParsedEventId to better reflect how it's used.

> > Source/WebCore/page/EventSource.cpp:319
> > +            if (!m_eventId.isNull())
> > +                m_lastEventId = m_eventId;
> 
> What will break if this if() check is removed?

If the EventSource reconnects and data is discarded, and the first event after reconnecting doesn't contain an 'id' field, then m_lastEventId will be incorrectly overwritten without that if() check. I've slightly updated the eventsource-reconnect test so that this case is covered by the tests.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.