[Spread-users] daemon crash on illegal private name

Jonathan Stanton jonathan at cnds.jhu.edu
Fri Oct 24 16:14:29 EDT 2003


What a coincidence. Just a few days ago I received a contributed patch to 
fix the parser for the spread.conf file to reject machine names that are 
too long. They discovered the same crash problem that you did. 

I'm applying that patch to cvs this afternoon.

In general, I agree that the parser should not accept illegal syntax -- 
and we attempt to do just that. On many invalid configurations the daemon 
refuses to startup and prints the invalid configuration line to help the 
user fix the problem. There must be some cases that get through.

Do you have the spread.conf that failed? 

I'm guessing from your message that you don't. I'll still look and see if
it's obvious how an invalid config is not caught, but it might be tricky
to see as we handle a bunch of the obvious cases already.

Jonathan


On Fri, Oct 24, 2003 at 03:57:27PM -0400, Jeremy Hylton wrote:
> > In short, I believe that an integrity check that should be done on
> > startup on the daemons' names isn't being done correctly.  Or this could
> > be a symptom of a completely different bug, we'll try and find out.
> 
> Ok.  I understand.
> 
> On the subject of integrity checks, we've had trouble with syntax errors in
> config files.  I don't recall the details, because we were in a hurry to
> help a customer get spread running.  I believe the curly braces were
> unbalanced.  The daemon started but was effectively unresponsive.  It
> wouldn't communicate with other daemons.  It would be helpful if the daemon
> could complain about invalid config files rather than running in an unknown
> state.
> 
> Jeremy
> 
> 
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users

-- 
-------------------------------------------------------
Jonathan R. Stanton         jonathan at cs.jhu.edu
Dept. of Computer Science   
Johns Hopkins University    
-------------------------------------------------------




More information about the Spread-users mailing list