Search options (Page 1 of 3)

Thanks for trying it out. I intended it to give an easy way to undo the most recent changes, basically just like "undo" in a word processor (for the "I just installed 4 apps to test and now I want to get rid of them, but I forgot the package names of the first two" or the "I just uninstalled a list of things i didn't think I needed but something broke and I don't remember the names of any of them" or the "I screwed something up, I just want it back the way it was 1hr ago" situations). Maybe I don't understand what you are suggesting (feel free to correct me) but what would the advantage be of what you are suggesting over say apt-get auto-remove? Do you mean just so you can find the the point to get back to more easily? BTW, how long was the list it gave you?

Some windows on some desktops:
Same zoomed in on Ooo:
Like 1 with an extra window or two:
Zoomed in on Opera:

And yes slashdot and random code are my two cures for productivity. ; P

Do you mean to ask, how big would a checksum/signature have to be to be truly unique to the file, to ensure that there is no possible collision?
I'm not well-versed in information theory, but it would seem to have to be as long as the original file to guarantee that there is no possible collision.

Dosn't this violate some major rule of information theory?
I didn't think checksums were unique to a file of a particular size. I just thought they were statistically unlikely, so that the odds of a couple errors in a transfer compensating for one another (so the checksum comes out right, even though the file is corrupt) were negligably low.
If it isn't mathematically undetermined, then wouldn't it just be easier to write an algorithim that will work backwards, rather than guess and check?
Maybe I just misunderstood the process.

EDIT: This page has links to two Postscript files of the same size with the same with the same MD5, and also happen not the be gibberish: http://th.informatik.uni-mannheim.de/Pe … ollisions/

I'm in the process of writing a network manager that will detech when wifi starts to crap out and reload the module. I posted a (now very out of date and primitive) version earlier on the forum (if you are interested let me know an I can post the current, much improved, version). It isn't really a great fix, because things have to get bad before it tries to reload it, and then you lose networking for a couple secs while it actually reloads it and reconnects, but it is better than having to stop what you are doing and manually bring it down.

I know it's not helpful, but I was laughing really hard reading over these posts (in the i've-been-there-before way).
Best of luck.

+1 on the sidux idea. I've been using it and I'm really impressed.
I think starting from sidux, running the smxi script to get the non-free stuff, then installing openbox, conkey, lxpanel, etc, copying your #! config files over, and adding any other highend aps you need (e.g., claws, gimp, etc) from the sidux/sid repos would probably give a pretty close experience.
You could even make a script to do it in case someone else want to imitate it.

for network printers

If you know what you are doing or are willing to do the research, in order to reduce your memory use after boot (and reduce boot time) you should look through /etc/rcX.d directories (especially rc2.d) and see if things are getting started that you don't need. Just make sure you actually know what you are doing before playing around or you may end up taking out something important.
To keep something from running rename it with a K instead of an S (the README in the directory explains). If you don't use bluetooth often, you can pull that out. If you rarely print, pull out cupsys. Rarely have to use removable media, pull out ivman. All these services can be started manually if you need them.
Same idea for openbox's autostart.sh (although nothing in there will really screw up the system if you mess it up, that I can think of at least).
I don't know how much you will be able cut out (or whether it will be worth the inconvenience of having to start the stuff manually) but if you are hell-bent on decreasing the footprint this is a place to consider.

Snowpine makes a good point, make sure you are checking the mem used by applications, not the total. htop show both in different colors. You shouldn't worry about linux keeping a big cache, but it is perfectly legitimate to try to reduce non-system mem use (the less memory being used by apps the more linux can use to cache).
And, just fyi, windows from XP does the same thing but actually caches things ahead of time with prefetch (XP) or superfetch (Vista, 7?, which predicts when you will need what in memory). This is why Vista starts out after boot with such a high reported memory use, which hardly increases as you start opening apps: The app was already loaded into memory.

if you mean the options in the menu, mainmenu > preferences > openbox config > gui menu editor || edit menu.xml
at least these were the options in 8.4.1 they may be in a different place or have different labels in the later versions.
If it isn't there you can start it with the command obmenu (from a terminal or from the alt+f2 run dialog)