Re: [patch] More NSE bugfixes

[Daniel Miller <bonsaiviking () gmail com>]

Ron,

In general, throwing an exception should be avoided in favor of 
returning nil from the action, or returning a description of the error 
if nmap.debugging() > 0. I would guess that the majority of them are due 
to bugs rather than an expected failure condition.

Stack traces are particularly interesting right now because of the 
change from Lua 5.1 to 5.2, so any bugs related to that can be worked out.

Dan

On 06/05/2012 08:53 AM, Ron wrote:
> A question related to this - are all stacktraces in -d output bad? Or are some expected?
>
> I've scanned 50,000+ Internet hosts in the last little while. There are a lot of stacktraces in my output. I can go through and 
> post them if they shouldn't be happening. Otherwise - if some are okay and expected - I'll go through at some point and 
> triage the ones that are probably legit bugs.
>
> Ron
>
> On Sun, 3 Jun 2012 21:36:12 -0500 Daniel Miller<bonsaiviking () gmail com>  wrote:
> > List,
> >
> > Decided to hunt bugs again, by running a debugging scan on my network
> > (sudo ./nmap -A --script 'default or (safe and discovery)' -d5 -sU -sS
> > -sY --log-errors 192.168.1.0/24 -T4 -oA ../test). Here are the bugs I
> > came up with (patches attached):
> >
> > Fix bug in traceroute-geolocation: ipops
> >
> > ./scripts/traceroute-geolocation.nse:100: variable 'ipops' is not
> > declared
> > stack traceback:
> >      [C]: in function 'error'
> >      ./nselib/strict.lua:80: in function '__index'
> >      ./scripts/traceroute-geolocation.nse:100: in function
> >      <./scripts/traceroute-geolocation.nse:82>
> >      (...tail calls...)
> >
> > ----
> > Fix bug in dns.encodeFQDN
> >
> > In lexmark-config.nse, dns.query is called with "" as the first
> > argument, which breaks dns.lua, resulting in this error:
> >
> > ./nselib/dns.lua:768: attempt to concatenate a nil value
> > stack traceback:
> >      ./nselib/dns.lua:768: in function 'encodeQuestions'
> >      ./nselib/dns.lua:848: in function 'encode'
> >      ./nselib/dns.lua:318: in function 'query'
> >      ./scripts/lexmark-config.nse:66: in function
> > <./scripts/lexmark-config.nse:56>
> >      (...tail calls...)
> >
> > Fixed by returning \0 byte from encodeFQDN if the name passed in is
> > the empty string. Not sure how this will work, but it shouldn't throw
> > exceptions, at least.
> >
> > ----
> > A change had left an undeclared variable, func. Fix was to strip out
> > the test for existence of func, since that logic was based on
> > previously loading the function from a file, instead of the current
> > method of indexing into an object to get the function. Error was:
> >
> > ./scripts/wsdd-discover.nse:53: variable 'func' is not declared
> > stack traceback:
> >      [C]: in function 'error'
> >      ./nselib/strict.lua:80: in function '__index'
> >      ./scripts/wsdd-discover.nse:53: in function 'main'
> >      ./nse_main.lua:803: in function<./nse_main.lua:803>
> >
> > ---
> >
> > Lastly, there's a bug with cups-queue-info (or possibly with the
> > ipp.lua library) that I can't hunt down at the moment. Perhaps with
> > access to a printer tomorrow I'll be able to hunt it down, but here's
> > the exception traceback:
> >
> > NSE: 'cups-queue-info' (thread: 0xb04eae0) against 192.168.1.2:631
> > threw an error!
> > ./scripts/cups-queue-info.nse:40: attempt to index upvalue 'ipp' (a
> > boolean value)
> > stack traceback:
> >          ./scripts/cups-queue-info.nse:40: in function
> > <./scripts/cups-queue-info.nse:39>
> >          (...tail calls...)
> >
> > Dan
> _______________________________________________
> Sent through the nmap-dev mailing list
> http://cgi.insecure.org/mailman/listinfo/nmap-dev
> Archived at http://seclists.org/nmap-dev/
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/