Comment # 12
on bug 8674
from [email protected]
Hi Evan,
I've cleaned up the issues that you brought up but have some questions on how
to proceed because of a problem with the current implementation.
We are currently associating requests with responses using a 16-bit transaction
id. This is a problem because of rollover. The benefit of doing this is that it
allows us to correctly decode the responses and is very useful when
troubleshooting.
Because of this, would it make sense to make this dissector a plugin? This
would allow us to submit a dissector which would not be able to fully decode
the responses but still allow vendors to swap it out with one that does full
decoding (with the understanding that its use is limited to shorter captures
without rollover). In this case is it fine to comment/ifdef out the
request/response tracking code so others can enable it when building their own
dissector?
The other option is to keep it built in but have some kind of switch (presence
of a file?) that would enable the response decoding.
Is there a better way to handle this?
Thanks,
Warren
You are receiving this mail because:
- You are watching all bug changes.
