On 8/31/06, Trustin Lee <[EMAIL PROTECTED]> wrote:
Hi Vinod,
On 8/30/06, Vinod Panicker <[EMAIL PROTECTED]> wrote:
>
> Hi,
>
> Currently if the application calls a write() on MINA, it will get
> notified of any exceptions via the exceptionCaught() method. The
> problem is that the application does not have any information as to
> what data could not be written on the session. To figure that out, it
> either has to wait on the WriteFuture or implement some state
> mechanism within the IoSession.
Hmm, you are right. We need to provide additional information.
Any better ways to handle this? Would it make sense for
> exceptionCaught() to supply the buffer that could not be written?
What do you think about creating a new exception 'WriteException' which
extends IOException? If this sounds reasonable, please create a JIRA issue
for us. :)
I think this should do the trick. We can have more exceptions in the
future if needed.
I'll open a JIRA issue.
Regards,
Vinod.
