I believe we do this in the build-bot testing, by doing:
tshark -G values
Since that operation iterates across all the value_string arrays, a
non-terminated array will result in an error .... or at least it
should.
Is that enough testing?
--gilbert
On 9/20/06, Ulf Lamping <ulf.lamping@xxxxxx> wrote:
Hi List!
It seems to be a common mistake to forget the terminating zero entry in a value_string, I've done this myself before and it's hard to track it down if you don't have a clue what's going wrong.
Even worse, this mistake might not make any problems for a long time as usually the values rushing in will be in the value_string (I've found my mistake by fuzz-testing).
We could automatically check the termination of the value_string (at least for the registered values) while registering the hf_ values. This will indicate the problem a lot earlier and at the place where it's caused (and not much, much later).
Of cause this has a drawback: Iterating trough all value_strings at registration will take a little bit of time at startup (didn't tried so don't know how long this will take).
What do others think: Is the added time overhead worth adding some extra robustness or are there other mechanisms possible to detect these kind of errors (e.g. static code analysis)?
Regards, ULFL
_____________________________________________________________________
Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!
http://smartsurfer.web.de/?mc=100071&distributionid=000000000066
_______________________________________________
Wireshark-dev mailing list
Wireshark-dev@xxxxxxxxxxxxx
http://www.wireshark.org/mailman/listinfo/wireshark-dev