Re: [tor-dev] Proposal: Controller events to better understand connection/circuit usage

[Karsten Loesing <karsten@xxxxxxxxxxxxxx>]

On 2/11/13 5:37 PM, Damian Johnson wrote:
>> Sure.  Please find the revised proposal in branch proposal218 of my
>> public torspec repository:
>>
>> https://gitweb.torproject.org/user/karsten/torspec.git/shortlog/refs/heads/proposal218
> 
> Hi Karsten. Looks good! Pushed some revisions to my proposal218 branch...
> 
> https://gitweb.torproject.org/user/atagar/torspec.git/shortlog/refs/heads/proposal218

Looks good.  I'm fine with making ConnType an enumeration type,
especially when controllers must tolerate unrecognized types (which I
didn't know was common in control-spec.txt).  But I wonder why we would
handle CellType differently and not turn that into an enumeration, too.
 There are 13 connection types and 17 cell types in or.h, so not totally
unreasonable to enumerate.  Or even better, cell types are specified in
Section 3 of tor-spec.txt which we could reference.  Also, should we
change CellType values to upper-case?

> For the 'Per-circuit cell stats' section it would be nice if there was
> an explanation about what the 'N' and 'P' prefixes mean (it's
> described as "app-ward" and "exit-ward" but I haven't a clue what
> those mean).

The P and N come from the Tor code, and I guess they stand for next and
previous.  We can also call them AppWard and ExitWard, like in
AddedAppWar or TimeExitWard, if that makes it easier to understand.

Thanks,
Karsten

_______________________________________________
tor-dev mailing list
tor-dev@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-dev