Allows the creation of empty SSE events

[spericas]

Description

Allows an SSE event to be created using no data. Uses a static constant SseEvent.NO_DATA to avoid returning null in getters. Updates serialization code and tests. See issue #9165.

[tjquinno]

The changes look fine to me, although I do have one question.

I see the writer of the issue cited the SSE spec saying that "empty" events are legal.

I know very little about SSE, but is it in any way a potential problem that the SseEvent data for an empty event would contain this special constant value (as opposed to having no value at all)? For example, does this distort some representation of the event size or length or something like that that could otherwise be used to infer an "empty" event?

[spericas]

The changes look fine to me, although I do have one question.

I see the writer of the issue cited the SSE spec saying that "empty" events are legal.

I know very little about SSE, but is it in any way a potential problem that the SseEvent data for an empty event would contain this special constant value (as opposed to having no value at all)? For example, does this distort some representation of the event size or length or something like that that could otherwise be used to infer an "empty" event?

The code in SseSink currently checks for that constant and does not serialize anything into the stream. Would that answer your question?

[tjquinno]

Yes. Thanks.