Wireshark-bugs: [Wireshark-bugs] [Bug 11917] Custom types dissection in RTPS

Date: Wed, 23 Dec 2015 14:39:32 +0000

Comment # 9 on bug 11917 from
(In reply to Michael Mann from comment #8)
> While Wireshark can't control the direction of protocols, the two thoughts I
> had were:
> 1. Schrodinger's cat
> (https://en.wikipedia.org/wiki/Schr%C3%B6dinger%27s_cat), in that you are
> involving Wireshark in something its intended to observe (really hope I have
> that analogy right).
> 2. This seems very similar to using capture comments to further dissection
> (https://code.wireshark.org/review/9686/), which also seems like a bad idea.
> 
> This is more a critique of the protocol than the patch that provides the
> functionality.  Don't have a reason to reject the functionality, I just
> really don't want to encourage it in other protocols that may be as
> flexible.  I think my biggest objection is that the string used is coming
> from Wireshark, NOT the protocol.

In my opinion, as long as we are consistent, we have the power to provide users
an standardized way to dissect their traffic. If we let users know that they
can dissect their user data types in RTPS by creating a dissector and
registering it using the Type Name, they will follow that process.

As always, processes need to be defined. And I think we are the ones that have
to define this process. That is why I provided the suggestion that I think
makes more sense: using the Type Name.

I have discussed this with RTPS experts and they all agree with me. If anyone
doesn't agree, I am more than willing to discuss and implement it in a
different way. I just want this to be implemented.


You are receiving this mail because:
  • You are watching all bug changes.