Vadim Gritsenko (JIRA) wrote: > [ > http://issues.apache.org/jira/browse/COCOON-1658?page=comments#action_12355965 > ] > > Vadim Gritsenko commented on COCOON-1658: > ----------------------------------------- > > (where is reply button in jira?) > > >>So, my question is why is the buffer right now set to "unlimited"? >>Is there any specific caveat for this? > > > If you are using caching pipeline then complete response will have to be > buffered > in any case. For non caching pipelines, complete response will have > to be buffered if serializer requires content length to be set. > For all other cases... The only reason for "unlimited" is error handling. If an error happens when some content has already been streamed to the client, we can't roll back the already streamed part. Without buffering you would get corrupted pages. So if you're sure that you don't have exceptions during streaming, you can turn off buffering. We decided a long time ago, to set the default to "unlimited" just for making error handling work in all cases.
If the caching pipeline is used and in the case the serializer needs to set the content length, then the pipeline takes care about buffering. > I think this outputBuffer is not really needed. > I think this is another example of running modes. During development you might want to have an unlimited buffer while in production (where never an error occurs :) ) you can disable it. Carsten -- Carsten Ziegeler - Open Source Group, S&N AG http://www.s-und-n.de http://www.osoco.org/weblogs/rael/
