Re: [NSE] proposal of size_tohuman() in stdnse library

[Djalal Harouni <tixxdz () gmail com>]

On 2010-06-08 11:16:53 -0600, David Fifield wrote:
> On Tue, Jun 08, 2010 at 02:09:03AM +0100, Djalal Harouni wrote:
> > I've coded a function which converts the size in bytes to
> > human readable format, I'm using this function in my NFS code to show
> > files size, but perhaps this function would be more useful if it is in
> > the stdnse library, it's a generic function.
> >
> > I've make it simple, the %.1f format will do some round up/down.
> >
> >
> > --- Converts the size in bytes to a human readable format
> > --
> > -- An optional second argument is the size of the block
> > --
> > -- @usage
> > -- size_tohuman(1024) --> 1024.0B
> > -- size_tohuman(926548776) --> 883.6M
> > -- size_tohuman(246548, 1024) --> 240.8K
> > -- size_tohuman(246548, 1000) --> 246.5K
> > --
> > -- @param size in bytes
> > -- @param blocksize represents the number of bytes per block
> > --        Possible values are: 1024 or 1000
> > --        Default value is: 1024
> > -- @return String that represent the size in the human readable
> > --         format
> > function size_tohuman(size, blocksize)
> >     local idx, bs = 1, 1024
> >     local unit = { "B", "K", "M", "G" }
> >     if blocksize and blocksize == 1000 then
> >             bs = blocksize
> >     end
> >         for i=1, #unit do
> >             if (size > bs) then
> >                             size = size / bs
> >                             idx = idx + 1
> >                     end
> >         end
> >         return string.format("%.1f%s", size, unit[idx])
> > end
>
> This looks fine to me. In tcpip.cc there is a function called
> ll2shortascii that is used for the same purpose in Nmap. (ll2shortascii
> is a bad name; it's going to be changed to format_bytecount when Luis
> merges his nmap-dedup branches.) How about the name format_bytecount?
> That goes with the names format_difftime and format_output that are
> already in stdnse.
Well, it's fine for this name (format_bytecount) or perhaps format_sizeof (found it on the net).

the ll2shortascii() functions seems to use only a size block of 1000
(which is used by some hdd manufacturers) but as I'm planning to emulate
the tool "ls" in my NFS code which uses 1024 as a size block, so adding
a new parameter to the function to choose the block size will be the
best bet.

For the moment I'm using this code in my nfs-ls but I'll change it and
export this function to NSE in the stdnse library, so I'll wait for
Luis's merge.

> David Fifield

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