[kde-linux] KDE 4 and monitor powering off.

Dale rdalek1967 at gmail.com
Fri Feb 26 00:15:12 UTC 2010


chrome://messenger/locale/messengercompose/composeMsgs.properties:
> Dale posted on Thu, 25 Feb 2010 16:40:13 -0600 as excerpted:
>
>    
>>> Hmm... I don't believe the standard konsole menu entry runs as root...
>>> and certainly, typing in konsole in krunner, and selecting the "run
>>> konsole" entry, should run it as the same user (the "terminal" entry
>>> simply invokes the menu entry, so it would run whatever the menu is
>>> setup to run).
>>>
>>>        
>> It doesn't.  I changed it so that it would run as root since I have to
>> be root to run portage.  I'm not a real security nut.
>>      
> =:^)
>
> I actually have three user logins I use, my normal user, an "admin" user,
> and (seldom) root itself.
>
> I have sudo configured so my normal user has very limited access (some
> with password, some without) to only a few specifically nailed down
> commands, with specific parameters.  One of those commands, however, with
> password, allows me to sudo to my admin user.
>
> The admin user has full passwordless access to do everything root could
> do, except that for most things, I have to sudo<command>  where root
> wouldn't have the sudo in front.  (Actually, I'm lazy and don't like to
> type the full sudo, so I have simply "s" aliased to sudo, and just type
> the s and a space.)
>
> For my frequently run admin commands, including portage related commands,
> I have a /l/sudobin/ directory (again, I'm lazy, the traditional /usr/
> local/ is simply a symlink to /l/, so that's the equivalent of /usr/local/
> sudobin/), which mostly consists of symlinks to a single executable
> script, /l/sudobin/exec-sudo, which is the following:
>
> #!/bin/bash
>
> exec sudo ${0##*/} $*
>
> So all it does is re-execute the command it was called as (thus the
> symlinks) with sudo in front.  IOW, if I run "eas", it executes "sudo
> eas" (where eas is another shortcut, short for emerge --ask
> (--update --deep --newuse) @system, e=emerge, a=ask, s=system, I have
> similar shortcuts for the whole set, epw for e=emerge p=pretend w=world,
> for instance).
>
> So as my admin user, esyn will sync (esyn is a script that syncs both
> portage and the layman managed overlays), eas will emerge ask the system
> update, and eaw will emerge ask the world update.  Those take no passwords
> and don't require the s/sudo in front, as I use them enough to have setup
> the symlinks in the sudobin dir, which is in my admin-user's path, to do
> so.
>
> FWIW, I have auto-complete setup to work for s as it would for sudo, and
> ea* and ep* just as it would for emerge, too, so I can do ep
> konq<tab><tab>  and have it supply the possible packages (konqueror, konq-
> plugins, konquest) I could emerge...
>
> The idea is to avoid logging in as root unless it's really necessary, tho
> I will if I'm doing something major like setting up, gdisking and mkfsing
> raid arrays or something.  And as the admin user, the common admin tasks
> require no "s " in front, so when I'm doing something like rm as root, I
> generally construct the command as a the admin user using tab completion,
> etc, run it, get the permission error, hit the up-arrow to recall the last
> command, home, and s<space>  to add the sudo invoke at the beginning, to
> actually run it as root.  By that point I know what I'm actually rming
> (the permissions error told me, so I can double-check before I do the
> actual rm as root), so when I actually add the s to the front of the
> command and hit enter, there's no accidentally rming the wrong thing!
>
> So for syncing and doing the normal updates, it's generally start konsole
> (which I have assigned a hotkey, so it's a two-stroke sequence to start a
> new konsole), ". admin" , to source the script that sudos me to the admin
> user, then esyn, then eps to see what the @system updates are going to be,
> epl<pkg>  if any of them look interesting, to get the changelog, edit
> package.use (se /et<tab>po<tab>pa<tab>u<tab>, se being short for sudoedit,
> of course, and using shell autocompletion) or my global USE flags if
> there's any new USE flags in the updates I need to attend to, eas to merge
> them.  Meanwhile, I run a second konsole session and epz (z for @selected,
> since s was already taken for @system) to get the world_sets updates, and
> again inspect them for anything interesting like USE flag changes and
> changelogs I might wish to check.  By the time I get thru checking them,
> the @system update is generally done, and I can run the @world update.
>
> When I'm thru, I eup (short for etc-update), ear (runs revdep-rebuild with
> --ask, I just kept the ea theme), ead (emerge --ask --depclean), and again
> ear if anything was cleaned.  Thus, when I'm fully done, the system is
> again fully config updated (etc-update), self-consistent (revdep-rebuild),
> and has no stale and no longer needed dependencies (depclean).  And most
> of it is simple 3-4 letter commands, with tab-completion on the
> parameters, if needed. =:^)
>
>    

OK.  I keep mine simple, in my eyes anyway.  I just open a Konsole as 
root, eix-s and hit tab for eix-sync.  That syncs everything.  Then I 
run emerge -uvDNa world and answer yes if I like it.  I only run 
--depclean once a month or so and run revdep-rebuild when I think it 
needed or after --depclean.

I still use world instead of the @<whatever> sets thingy.  I'm not big 
on the sets thing anyway.

That's my update.  It's simple even tho it makes me type a little more.

Dale

:-)  :-)



More information about the kde-linux mailing list