Re: 0.0.9.pre6 --list-fingerprint(key generation) for server fails

[tor <tor@algae-world.com>]

Roger Dingledine wrote:

> On Wed, Nov 17, 2004 at 03:37:00PM -0800, tor wrote:
>  
> > >       I have a 1-6 mb down and 1-4 mb line up and would gladly 
> > > contribute at least a middleman server
> > > but even though I have followed instructions on the wiki and the 
> > > documentation and have defined the server
> > > parameters in torrc(DataDirectory etc,) when running tor -f torrc 
> > > --list-fingerprint I get 1 line denoting version
> > >      
> The --list-fingerprint option is only there to output a hash of your
> identity key and then exit. Since clients don't have fixed identity
> keys, they no longer output one.
>
> But if you want actually run tor, you shouldn't use --list-fingerprint
> as an option. And all the keys and so on will be written to your
> DataDirectory just like normal whether or not you use --list-fingerprint.
>
>  
> > hmm the keys finally showed up in the proper directory... guess I am 
> > just too impatient...
> >    
> Ah, yes, it takes a second or so to generate keys the first time it
> starts, on some processors.
>
> So it is working now? Please consider registering your key fingerprint,
> per the instructions on
> http://www.freehaven.net/tor/cvs/tor/doc/tor-doc.html#server
>
> Thanks,
> --Roger
>  
Hi Roger,
        its now core dumping when using --list-fingerprint BEFORE 
complete key generation is achieved(the secret keys get produced and it 
cores,(this is on OpenBSD), may be the current CVS, when its just 
started without --list-fingerprint then key generation appears as 
normal. just a minor issue for latter at this point. This is being 
SQA-ed:) in my lab, as I am developing
a script to write already keyed memory images suitable for CDROM, 
Compact-flash and USB-flash. These will be used for
internal distribution only to run further integration testing. Hmm 
couldnt find a --generate-key option and I could just fork tor as a 
child process, wait until all the files for server usage are 
generated(keys etc) and then kill via PID , then write all files produced
to my memory image. Should be a rather amusing shell script. After the 
images are installed at the user sites then we will
register the keys AND think about exit policies, got to coordinate that 
with the owners of the Internet connections.

just tried it again AND OpenBSD likes to core when list-fingerprint is 
used after key generation occurs leaves a  decimal 139
in the $? rc field of the shell, havent tried the other flavours yet.. 
got to get... more later.



         Thanx
          a Tor user