[Xmonad] xmobar-0.4 - the xmonad status bar - released!

Robert Manea rob.manea at googlemail.com
Thu Jul 5 12:57:27 EDT 2007


* Andrea Rossato (mailing_list at istitutocolli.org) wrote:
[...]
> > 12954 robert    15   0 37932  2512  1508 S  0.0  0.7  0:00.09 ./dist/build/xmobar/xmobar
> 
> well, this does not tell me much, actually... xmobar is linked to
> everything needed to run a haskell binary in your system, and all this
> stuff must be loaded.

True for the virtual memory size. Still, the question remains why does
xmonad which is obviuosly also linked to "everything needed to run a
haskell binary" _not_ use equally much memory? 

The answer might be: because xmobar uses some really huge libraries that
are not being used by/shared with any other haskell prog i use. So if I
go one step further i could conclude that I'm spending a lot of precious
memory for only a single application.


> have a look here:
> http://virtualthreads.blogspot.com/2006/02/understanding-memory-usage-on-linux.html
> 
> > Just for comparision purposes:
> > 
> >   2982 robert    15   0  5640  2436  1528 S  0.0  0.7  0:01.95 /home/robert/bin/xmonad
> >   2981 robert    15   0  4044   984   796 S  0.0  0.3  0:00.14 dzen2
> 
> This is just a misleading statement: dzen is using a lot of library
> that are already loaded into memory. This is not a useful comparison.
> You should know that.

Well, in my opinion the comparision _is_ useful because it shows 2
things:

* A haskell programm with a fair bit higher complexity than xmobar needs
  way less memory or said in other words does not need that many huge
  supporting librarieѕ as xmobar.

* A program (dzen) with a somewhat comparable feature set needs way less
  memory. 

Sure you could argue that ps shows bogus for multiple instances of a
process etc. but used libraries need to be mapped and if they are huge
and only used by a single program they simply waiste memory.

Please do not take this as any kind of offense, i just wanted to point
out the obvious :):

> Ciao
> Andrea 

Bye, Rob.


More information about the Xmonad mailing list