tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sebastiaan van Erk <>
Subject Re: Comet: problem with request.getParameter() in Comet POST requests
Date Wed, 02 May 2007 13:11:03 GMT
Rémy Maucherat wrote:
> On 5/2/07, Sebastiaan van Erk <> wrote:
>> What you seem to want is more in line with the "asynchronous servlet"
>> (request, wait, response), which Filip and Remy pointed out is not the
>> quite the same as Tomcat's Comet. Ideally, both models would be possible
>> through  single unified API, but this is currently not yet the case.
> You're misrepresenting certain things. As you can see, it is already
> possible to do request, wait, response,

Sorry about that, did not mean to say it is not possible to do a 
request, wait, response with Comet. All I'm trying to say is that Comet 
was not designed (at least, Filip stated this) for the async servlet 
model, and I tried to point out some of the issues when using Comet in 
such a scenario, such as getParameter() not working, synchronization 
issues, etc.. There is a difference between "possible", and 
"easy/practical". What I meant was that the current API is not designed 
for both models, and it is not yet so practical for the async servlet model.

> but you'll have to wait for an IO event to free up the memory used by 
> the request processor.

You need the IO event to free up more than just the memory. The file 
descriptor also needs to be freed; this was causing problems for me earlier.

> It would seem easy to me to cause the client to disconnect after the 
> request, as keepalive is probably not extremely useful in this 
> scenario, producing the event.

With a combination of "Content-Length" or chunked encoding plus a 
"Connection: close" header this should be possible, I agree. I would 
still like to know the "official" way to synchronize this is, though.

I've seen it happen that my application was still in the "write()" 
method of the reponse output stream when I already received the END 
event in my CometProcessor. At the moment I'm synchronizing 
event.close() and all access to the response output stream with the same 
lock, but I do not know for sure if this is the correct approach. I 
would very much like to know what the "official" approach is.

> Rémy


To start a new topic, e-mail:
To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message