Nicolas FROMENT a écrit :
As you seem convinced that my decoder is wrong can i suggest that Mina
can failed too ? :)
As a rule of thumb, for any library that has been widely used for more
than one year, I would suggest the following steps :
1) first, suspect your program. This is basically the correct approach,
say, 99% of the time.
2) if you don't see anything that could be wrong, just add some log to
the point you can exhibit the duplicated message, or you just can't
because it has slowed down so much the program that any threading pb has
been hidden.
2-1) Supposing you still have the problem with logs, well, that's a good
point. Now, the bad news is that we will have to analyse the logs...
Goto (3), but as we have logs, this is really easier
2-2) No way to get the pb with logs ... Goto to (3)
3) Well, now, this is the hard part : try to define a scenario were the
problem can be reproduced. I mean, even if it's one time each 100 000
messages, this should be the case every time you launch the program.
Now, there is _no_ guarantee we can reproduce the problem... You will
have to give more information : machine, cpu frequence, OS, version, JDK
version and origin...
What I mean is that MINA can be buggy. But finding the reason of a MINA
failure can be difficult without a scenario which exhibit such a problem.
