[prev in list] [next in list] [prev in thread] [next in thread] 

List:       metasploit-framework
Subject:    Re: [framework] why metasploit is so slow?
From:       ArcSighter Elite <arcsighter () gmail ! com>
Date:       2009-03-06 15:13:18
Message-ID: 49B13D8E.5020109 () gmail ! com
[Download RAW message or body]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

gaurav chaturvedi wrote:
> Hmm seems like i cant create a bug for this issue, anyways i guess i
> will have to provide all the info right here.
> I am using ubuntu 8.10, ruby 1.8.7, and msf from trunk
> msfconsole loads up rather slow.
> 
> msfconsole takes the following ammount of time while loading.
> 
> real    0m26.359s
> user    0m19.873s
> sys     0m2.128s
> 
> here is the output for ruby -v masfconsole
> 
> http://pastebin.com/m3214b35b
> 
> Hope that helps to debug the problem.
> 
> --Tazz
> _______________________________________________
> http://spool.metasploit.com/mailman/listinfo/framework
> 
I'm not sure I like this thread, particularly the tone being used.
So, I'll tell you what.
First, accurately profiling an application's performance it's a hard
task because there are so many environmental factors that arise which
make accurate results hard to achieve.
Of course, msf it's a bit slow, we all know, but we all love it anyway.
Do you guys have take a look at Canvas or Core IMPACT load times are at
least similar to msf's. So, being a rather long ruby project -
interpreted, rather than compiled - it's actually natural to have this
delays at startup essentially.
I'm not saying improvement couldn't be done, but I think the condition
is not for killing ourselves.
IMHO, I think first it's determining the source of this delay at msf's
code, if any, when solved, we'll get the interpreter itself ones, but
that's another story.
Manually inspecting the code, or running a ruby profiler along with a
coverage tool against most application paths, will provide us with the
ability to spot where are this delays, so that's the help nicholas was
talking about I think.
Once this done, then they'll determine solutions, caching, avoiding
module startup loading, etc. Giving msf's componentized architecture,
that wouldn't be hard to achieve.
Without aiming to hurt any-body's feelings, sincerely.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAkmxPXgACgkQH+KgkfcIQ8c2eQCeOMnho9NQzoOc0A3AwWD8QcvR
CUUAoKoWJ+oOnXueO/j5ovC6EIF7RJCp
=QnPk
-----END PGP SIGNATURE-----
_______________________________________________
http://spool.metasploit.com/mailman/listinfo/framework
[prev in list] [next in list] [prev in thread] [next in thread]