Re: [NSE][PATCH] string_ext library

["Patrick Donnelly" <batrick.donnelly () gmail com>]

On Wed, Oct 1, 2008 at 4:43 AM, Sven Klemm <sven () c3d2 de> wrote:
> Hmm extending string would really make more beautyful code IMO.
>
> Do you have any numbers regarding the performance difference? Does
> that imply that using ("abc"):foo() is also bad for performance since
> this probably also does a global string lookup instead of using a
> local copy of the string table.

I did a quick benchmark using os.clock (code attached):

For string.upper, in a tight loop with 5e6 iterations, it ends up
being about a 200ms linear difference between (worst to best):
string.upper (global lookup + upper lookup), string.upper (local
reference to string + upper lookup), upper (local reference to
string.upper). The exact numbers in order are (in seconds):

a =     1.85
b =     1.68
c =     1.4

The metamethod lookup ends up being about the same (as expected) as
the global lookup + upper lookup (a):

d =     1.78

After extending the string module and looking up tohex (which is just
string.upper for consistency):

a =     1.98
b =     1.78
c =     1.4
d =     1.91

As you can see, it's not a huge difference. I still would advocate the
use of local references to functions as in a register based virtual
machine it will always be optimal.

> Currently there is already a tohex() function in stdnse.lua which only
> works for numbers. It's probably best to rename my function to tohex()
> since this is more in line with the lua function names.
>
> Do you think we should put this function into its own module or do you
> think it should be added to stdnse?

I would actually recommend extending the tohex() function to handle
strings or numbers. Perhaps the other functions could be extended in
the same way as well.

Cheers,

-- 
-Patrick Donnelly

"One of the lessons of history is that nothing is often a good thing
to do and always a clever thing to say."

-Will Durant

Attachment: local_test.lua
Description:

_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://SecLists.Org