[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [E-devel] entrance is now bloated.



On Tue, 26 Sep 2006 19:58:36 +0100 Essien Ita Essien <essien@wazobialinux.com>
babbled:

> Essien Ita Essien wrote:
> > Eugen Minciu wrote:
> >   
> >> On Tue, 26 Sep 2006 14:31:04 +0900
> >> Carsten Haitzler (The Rasterman) <raster@rasterman.com> wrote:
> >> My entrance is built on September 11th and I don't have this problem.
> >>
> >> Hope this helps you narrow it down.
> >>   
> >>     
> > Ok... found it after a few minutes of poking around web-cvs.
> >
> > Actually, the problem came in with the shell utilization patches...
> > http://e.kevb.net/cgi-bin/viewvc.cgi/e17/apps/entrance/src/client/entrance_session.c?r1=1.85&r2=1.86&sortby=date
> >
> > Now that I've identified it. It will be fixed soonest.
> >   
> all fixed now.
> 
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  
> COMMAND                                                          
>  1955 root      15   0  6952 1936 1528 S  0.0  0.4   0:00.04 
> entranced                                                        
> 16418 root      16   0  2008  712  576 S  0.0  0.1   0:00.80 
> entrance_login                                                   
> 16429 essiene   16   0 18532 8064  724 S  0.0  1.6   0:00.00 
> entrance                                                         
> 16430 essiene   15   0 23592  17m 4260 S  0.0  3.4   0:05.05 
> enlightenment                                                    
> 
> booyaa!!
> 
> I'll be more carefull in reviewing patches in future.

actually not all fixed. 1 entrance process floating around. there's no need for
it to hang about like a bad smell. :) i am pretty damn sure that when i fixed
it before entrance spawned off the user's login session then REPLACED itself
with entrance_login so it became small and unobtrusive. there were not 2
entrances and entrance became entrance_login :) i'm pretty damn sure that's how
it was... :)

> Cheers,
> Essien
> > Cheers,
> > Essien
> >   
> >>> OK - a quick check today:
> >>>
> >>>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND  
> >>>  4183 root      15   0 79712  28m 5216 S    0  2.9   0:35.94 Xorg      
> >>>  9385 raster    15   0  106m  21m  11m S    0  2.1   0:03.04 sylpheed  
> >>>  4189 root      16   0 65112  13m 3060 S    0  1.3   0:04.72 entrance  
> >>>  4600 raster    16   0 65112  11m  704 S    0  1.1   0:00.00 entrance  
> >>>  4601 raster    15   0 70988 9604 3984 S    0  0.9   1:50.63 enlightenm
> >>>  4168 root      16   0 41956 1248  860 S    0  0.1   0:00.00 entranced 
> >>>
> >>> ... now yes - i know how top and ps can lie - BUT, i only now JUST
> >>> noticed - someone has removed my work of using entrance_login. ie
> >>> entranced spawns (and manages) X and entrance (entrance is the gui front
> >>> end). when you authenticate
> >>> - entrance runs the login session. in order to make sure that when your
> >>> login ends everything is cleaned up login-wise, entrance_login sits and
> >>> waits for the user's login to end. entrance_login is small - it uses no
> >>> efl libs. it is intended to be so small as to use very little ram while
> >>> you are logged in. this seems to have been thrown out the window now with
> >>> entrance spawning 2 incarnations of itself and just siting around hogging
> >>> stale unused memory that will eventually need to be swapped out - while a
> >>> user is logged in.
> >>>
> >>> why was this removed? why suddenly consume about 20MB worth of resident
> >>> pages of ram when this is really not needed? just exec() entrance_login
> >>> so it replaces the entrance process with a brand new one that has no
> >>> allocated memory, fragmented pages etc. if any housekeeping needs to be
> >>> added - add it to entrance_login and pass it in via cmd-line, stdin or
> >>> something. as it stands entrance is hogging a huge amount of memory for
> >>> no useful purpose. entranced should stick around - and really - it's not
> >>> fat at all (400k of resident pages or so). look at the above -
> >>> enlightenment is leaner than entrance - by < 50% mem use. and entrance
> >>> should be doing NOTHING - it should be idle and consuming basically no
> >>> resources as it's only waiting for my login session to end.
> >>>
> >>> now - this brings me to my point. people need to think about what they are
> >>> patching and changing - not just in terms of features, but what they might
> >>> break, destroy and then bloat-up. this isn't gnome people. we CARE about
> >>> resource usage. if you are going to use ram or cpu - do it for a reason -
> >>> get something out of it (functionality, prettiness, etc.) but profile
> >>> your work - check its memory usage, check its process spawning - do the
> >>> processes go away? so you leave them hanging around consuming precious
> >>> ram and/or cpu.
> >>>
> >>> pay attention to the details. details make the difference between a $1
> >>> piece of plastic imitation junk churned out by a chinese factory and a
> >>> masterpiece of craftsmanship.
> >>>
> >>> i am sorry to have to say this - but i'm disappointed to see lax work like
> >>> this. :(
> >>>
> >>> -- 
> >>> ------------- Codito, ergo sum - "I code, therefore I am" --------------
> >>> The Rasterman (Carsten Haitzler)    raster@rasterman.com
> >>> 裸好多
> >>> Tokyo, Japan (東京 日本)
> >>>
> >>> -------------------------------------------------------------------------
> >>> Take Surveys. Earn Cash. Influence the Future of IT
> >>> Join SourceForge.net's Techsay panel and you'll get the chance to share
> >>> your opinions on IT & business topics through brief surveys -- and earn
> >>> cash
> >>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> >>> _______________________________________________ enlightenment-devel
> >>> mailing list enlightenment-devel@lists.sourceforge.net
> >>> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> >>>     
> >>>       
> >> -------------------------------------------------------------------------
> >> Take Surveys. Earn Cash. Influence the Future of IT
> >> Join SourceForge.net's Techsay panel and you'll get the chance to share
> >> your opinions on IT & business topics through brief surveys -- and earn
> >> cash
> >> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> >> _______________________________________________ enlightenment-devel
> >> mailing list enlightenment-devel@lists.sourceforge.net
> >> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> >>   
> >>     
> >
> >
> > -------------------------------------------------------------------------
> > Take Surveys. Earn Cash. Influence the Future of IT
> > Join SourceForge.net's Techsay panel and you'll get the chance to share your
> > opinions on IT & business topics through brief surveys -- and earn cash
> > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> > _______________________________________________
> > enlightenment-devel mailing list
> > enlightenment-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> >   
> 


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster@rasterman.com
裸好多
Tokyo, Japan (東京 日本)