[Click] Make SetUDPChecksum more verbose when something is wrong
Ian Rose
ianrose at eecs.harvard.edu
Thu Feb 25 20:49:47 EST 2010
Sounds reasonable to me!
Eddie Kohler wrote:
> Hm. The behavior I implemented was this:
>
> - If the second output exists, emit fragments & short packets to the
> second output with no message.
>
> - If the second output does not exist, then the user was obviously
> thinking there would be no fragments or short packets. Print a message
> the first time a surprise is encountered.
>
> No VERBOSE keyword.
>
> Does this make sense? SetUDPChecksum isn't a Check... elekment, so
> defaulting to true seems OK to me.
>
>
> Ian Rose wrote:
>> Very minor point here, but you might want that keyword to default to
>> false instead. Not only does this avoid behavior changes in all
>> legacy code, but its more consistent with similar, existing elements
>> with a VERBOSE keyword (such as CheckARPHeader, CheckIPHeader and
>> CheckTCPHeader).
>>
>> - Ian
>>
>>
>> Eddie Kohler wrote:
>>> Hi Bart,
>>>
>>> Totally reasonable. A patch to this effect is checked in.
>>> E
>>>
>>>
>>> Bart Braem wrote:
>>>> Hi,
>>>>
>>>> SetUDPChecksum now uses checked_output_push to output packets it
>>>> can not checksum. However, it would be nice if some warning would
>>>> be output, this silent behaviour is quite hard to debug. I suggest
>>>> giving a warning when packets are in this case, based on a verbose
>>>> keyword that defaults to on.
>>>>
>>>> Regards,
>>>> Bart Braem
>>> _______________________________________________
>>> click mailing list
>>> click at amsterdam.lcs.mit.edu
>>> https://amsterdam.lcs.mit.edu/mailman/listinfo/click
More information about the click
mailing list