Bill Sommerfeld wrote: > On Tue, 2007-06-12 at 14:14 -0700, Bart Smaalders wrote: >> I'm planning on maintaining Wireshark in the sfw consolidation by >> keeping it up-to-date with the community releases. I do not currently >> plan on removing snoop from ON. If those who currently maintain >> snoop wish to redirect their efforts to making sure that wireshark >> protocol support is a proper superset of that in snoop and then EOF >> snoop, that's fine w/ me. Apparently, snoop has been placed in >> sustaining mode already. If no resources are available to do even >> that much, leaving snoop alone and declared Obsolete certainly won't >> set any precedents; much of our userland code is defacto in the same >> state. > > PSARC and the ON cteam have operated for some time with a general > architectural principle that our observability tools (snoop, truss, > dtrace, pfiles, etc.,) ought to be able to observe what our > implementation does and thus projects which add new artifacts ought to > extend these tools accordingly. > > Do you believe this principle is in error? If not, how should we > instruct project teams to let us deliver this system property via > wireshark? > > > >
I would suggest the following: 1) project teams introducing new network protocols should be contributing these changes back to the wireshark project. 2) failing this, teams could enhance our copy of wireshark with a new plug-in. I would much prefer 1) above, as it's simpler and allows users using other OSes to observe traffic generated on Solaris more completely. - Bart -- Bart Smaalders Solaris Kernel Performance barts at cyber.eng.sun.com http://blogs.sun.com/barts
