Thanks Filip. Buffering is what I do right now and that is what I'm trying to avoid. With a large number of concurrent users, the buffering is proving to be a performance bottleneck and is causing frequent GC sweeps.
I think what I'm trying to do with the Trailer header is valid http usage (correct me if that's wrong). And I was hoping tomcat supports this usage. --Keyur Filip Hanik - Dev Lists wrote: > > On 08/05/2009 02:14 PM, Keyur Shah wrote: >> Tomcat by default sends chunked responses => Transfer-Encoding:chunked >> >> I'm calculating the ETag for the response dynamically as the content is >> being >> written to the response stream. Since I'll know the final value of the >> ETag >> only >> when the stream is closed, I indicate that using the Trailer header => >> Trailer: >> ETag >> >> I'm ensuring that just before the stream is closed, I set the ETag header >> => >> ETag: SomeETagValue >> >> However, when I look at the response on the client, the ETag header was >> never >> written. >> >> Is this a tomcat specific bug or am I doing something wrong? >> > When you write, after you have filled up the buffer, tomcat will commit > the response. > You will need to not write any of your response until you have the tag, > so buffer your response, > calculate your tag, set the header, write the response > > Filip >> Thanks, >> Keyur >> > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > For additional commands, e-mail: users-h...@tomcat.apache.org > > > -- View this message in context: http://www.nabble.com/Trailer-header-and-tomcat-tp24835409p24836927.html Sent from the Tomcat - User mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
