Hello.
There exists much of subject to solve about this task ...
I will solve the individual subjects.
Firstly, I want to think about doing layer B stream for all cases or not.
In Bryan's comment
Or, we could take Kathey's suggestion and change that path so that it
also uses Layer B stream. That is, have writeScalarStream() *always*
do Layer B streaming, even if the length is known ahead of time.
I left processing of not doing layer B streaming expecting the benefit
that we can lower amount of buffer in proportion to the length.
However, amount of buffer is same size of DSS segment or less.
Then benefit is considerably limited.
And I realized different kind of benefit in doing layer B streaming
reading Bryan's comment next.
Instead, we use Layer B streaming.
Therefore, we no longer have this implicit contract to deliver the
entire blob,
and we are free to terminate our send at any point.
Now, I think we should do layer B streaming for all cases.
Best regards.
--
/*
Tomohito Nakayama
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Naka
http://www5.ocn.ne.jp/~tomohito/TopPage.html
*/
