stdout during event_callback cannot be used

I have tried to run a background scan with:

arg='--send-ip -PE -PS21,22,23,25,53,80,110,111,139,443,445,5357 -PA80 -PP'
nmapscanner = NmapProcess(targets='10.1.0.80-100', options=arg, event_callback=lambda nmapscanner: callback(nmapscanner))
nmapscanner.run_background()

my callback function just try to parse the nmapscanner.stdout but it's impossible, the output is not xml well formed.
I have tried to use ElementTree and your NmapParser and they raise the following errors:
Wrong XML structure: cannot parse data (NmapParser)
no element found: line 4, column 0 (ElementTree)

So there is no way to access to datas discovered during the scan (for example if I want to know which host have been found alive before the end of the scan), the only way to access datas: stdout, isn't usable (or only with advanced regex).

I'm not sure this is an issue, but python-nmap permits to access to datas during the scan, so I guess there should be a method to do that here too.

Thanks for your work, it's a great lib

[savon-noir]

yep, I didn't implemented a partial xml block parser. This could be achieved with a pull-dom type parser but I haven't. The basic idea is already there but I only did it partially to collect the percentage done but no other events (see def __process_event in process.py).

If you fee like implementing this, do not hesitate, fork and issue a pull request, I would be happy to merge such feature but I will probably not have the time to do this myself for the moment.

Hi, in fact, the workaround that I use is hard adding at the
end of the partial xml returned, this work for me. I'm not sure this is
the solution that you expect ^^.
Thanks for your work

| David SORIA | Security Engineer
| ITrust France | IT Security Services & SaaS Editor
| Email: d.soria@itrust.fr | Fixe 05.67.34.67.87
| Bâtiment ACTYS/1 55 l'Occitane 31670 LABEGE

On 16/03/2014 12:32, Ronald wrote:

yep, I didn't implemented a partial xml block parser. This could be
achieved with a pull-dom type parser but I haven't. The basic idea is
already there but I only did it partially to collect the percentage
done but no other events (see def __process_event in process.py).

If you fee like implementing this, do not hesitate, fork and issue a
pull request, I would be happy to merge such feature but I will
probably not have the time to do this myself for the moment.

—
Reply to this email directly or view it on GitHub
#13 (comment).

[savon-noir]

Hey,

I'll add an attribute to parse() method "incomplete" which will basically enable you to parse interrupted or not yet finished scan.

I've tested you raw style method of appending a at the end of the stream and it seems to work (although the best would be to add another xml stream based parser engine, but I know, i'll never do this :p)

If, from your experience, you have improved the proposed method, do not hesitate to issue a pull request or provide me with some comments :)

feedback is always good.

Cheers!

Ronald

[savon-noir]

added in version 0.4.5 in parse.py, see commit: 12e9ad3

Hi,

I have not improved this method, it was sufficient for all my test cases.
But in fact, I have to switch to python-nmap because of CPU consumption.

You are using python thread, but this is not as performant as
muti-processing:
Sometimes, it takes 90% of one processor.
With python-nmap: less than 1%

So you should probably consider to use multi-processing.

Thanks for your reactivity

Good job

| David SORIA | Security Engineer
| ITrust France | IT Security Services & SaaS Editor
| Email: d.soria@itrust.fr | Fixe 05.67.34.67.87
| Bâtiment ACTYS/1 55 l'Occitane 31670 LABEGE

On 06/04/2014 13:23, Ronald wrote:

Hey,

I'll add an attribute to parse() method "incomplete" which will
basically enable you to parse interrupted or not yet finished scan.

I've tested you raw style method of appending a at the end of the
stream and it seems to work (although the best would be to add another
xml stream based parser engine, but I know, i'll never do this :p)

If, from your experience, you have improved the proposed method, do
not hesitate to issue a pull request or provide me with some comments :)

feedback is always good.

—
Reply to this email directly or view it on GitHub
#13 (comment).