[Someone to help test from page 5 ?] Apt-sources-update ? (Re Open)

[melodie]

Ok, same as before it stops just after having kept the 2 lines for the PASS repository. Once done it does not offer to rewrite the /etc/apt/sources.list, nor to start Synaptic.

I'll retry at once because appearantly the last directory "repo-test-speed" is of yesterday. Removing it and restarting apt-sources-update...

[melodie]

Ok, now repo-test-speed/ is in /tmp. But it still does not go beyond the point said above. I stops after having kept the 2 lines for PASS server.

Code: [Select]

[root@localhost ~]# cd /tmp/repo-speed-test/
[root@localhost repo-speed-test]# ls -l
total 16
-rwxr-xr-x 1 melodie melodie  595 août   5 15:33 backup-n-overwrite-sources.sh*
-rw-rw-r-- 1 melodie melodie 1617 août   5 15:32 repo_goodlist.txt
-rw-rw-r-- 1 melodie melodie 3050 juin  13 23:30 sources.list
-rw-rw-r-- 1 melodie melodie 3358 août   5 15:33 sources.list.new
[root@localhost repo-speed-test]#

[daniel]

then i've no idea any more, sorry...
Stopping my work for a while.

[melodie]

then i've no idea any more, sorry...
Stopping my work for a while.

Hi Leiche,

I might have THE solution.

I have taken your version apt-sources-update-0.3.1-4leiche2011.i586.rpm and reinstalled it. See ?

# apt-cache show apt-sources-update
Package: apt-sources-update
Section: System/Configuration/Packaging
Installed Size: 33528
Packager: leiche
Version: 0.3.1-4leiche2011
Depends: zenity, curl, wget
Provides: apt-sources-update = 0.3.1-4leiche2011, apt-sources-update(x86-32) = 0.3.1-4leiche2011
Architecture: i586
Size: 1
MD5Sum:
Filename:
Summary: Utility for optimizing PCLinuxOS repository sources.list
Description:
 apt-sources-update is a utility that will download the latest PCLinuxOS apt
 sources.list from the main ibiblio server, verify that they
 have been recently synced to the ibiblio server, and test
 all updated repositories for download speed.
 
 Once testing is completed the user will be given a speed
 rated list of repositories from which they can select a
 default repository to use for future software needs.

I have ktsuss installed in my system. What is ktsuss ?

Description :
ktsuss stands for "keep the su simple, stupid", and as the
name says, is a graphical version of su written in C and
GTK+ 2.The idea of the project is to remain simple and bug
free.

Simple and bug free...

I have modified your apt-sources-update of that one version (which was working well apart from the ldconfig message... ) and changed theses 2 lines:

line 133 :

Code: [Select]

   synaptic &
to

Code: [Select]

   ktsuss synaptic &
and line 455 :

Code: [Select]

synaptic -r
to

Code: [Select]

ktsuss synaptic -r
Then did a sync (is that useful ? I'm not sure... ) then started apt-sources-update from the menus, then it worked asked for the repo I wanted, I chose another one, not so fast, but to see one that is different... when it was time to overwrite the /etc/apt/sources.list file I was prompted for the root password, then Synaptic was opened, with the window for the list of repos, and then I installed

Commit Log for Sat Aug  6 12:38:20 2011


Les paquets suivants ont été installés :
emelfm2 (0.7.3-1pclos2010)
gkrellm (2.3.5-1pclos2010)
gkrellm-plugins (2.3.5-2pclos2011)

No more ldconfig message !

If other testers could confirm this find, I would suggest to make ktsuss king of the windowed su commands !!! \o/

[Vorteggs]

In the past there were many posts questioning the veracity of these tests, showing similar real life usage of varying mirrors differing greatly from the results of the tests. I think this is part of the reason Texstar chose to not update the script. If it is inaccurate, it's really pretty much useless.

So the one to blame is the script itself. Blame it, hehe.

[melodie]

In the past there were many posts questioning the veracity of these tests, showing similar real life usage of varying mirrors differing greatly from the results of the tests. I think this is part of the reason Texstar chose to not update the script. If it is inaccurate, it's really pretty much useless.

So the one to blame is the script itself. Blame it, hehe.

No, the script is good, it's Synaptic environment, or ldconfig's maybe that was changed and now needs to be full root.

gksu usually works well, but "gksu -l" which is the best option in the desktop file is not the right option in this script, as then the script does not work well anymore. So it is possible that (if I understood well) if gksu -l is the equivalent of sudo, it might be buggy. (I write lot's of "may" and "might" because I don't have the knowledge that allows me to be sure). But the ktsuss program, written in C, light and desktop free is not buggy : have used it since several years and never got a problem.

KTSUSS happens to make the job fine for me in the apt-sources-update script of the above version !
And I am also getting information for a newer version with a patche that replaces now a lib that does not exist anymore. In the meanwhile all who would be ok to test the couple "ktsuss" and the change to that version of apt-sources-update will be welcome, so to have more feedbacks than just mine.

PS: I got into the message error problems, for the speed tests Leiche might be able to make it more accurate...

[Old-Polack]

melodie:

You seem to miss the main point. Whether the script goes through all the motions correctly doesn't matter if the speeds reported for the individual mirrors, in relation to each other, are inaccurate.

You are obsessing about all sorts of details, while the main purpose for the existence of the script is not being properly fulfilled.

Think of it as a car you keep washed and waxed, the interior is vacuumed regularly, the windows are spotless, and the air in the tires is at the optimum pressure. The brakes and steering are mechanically perfect. The problem is there is no engine or transmission with which to perform the main function of an automobile, getting from point A to point B. If the main function cannot be properly done, all the rest means nothing.

[melodie]

melodie:

You seem to miss the main point. Whether the script goes through all the motions correctly doesn't matter if the speeds reported for the individual mirrors, in relation to each other, are inaccurate.

Hi O-P,

Are they really inaccurate ? Maybe there is the need for different methods of scanning the response of the mirrors, according to the kind and quality connection one has, but I don't think it is inaccurate.

I have sought for the fastest, and most often the one shown as fastest was really much faster than the one configured preceedingly ; I have recently sought for the slowest and less up to date, and thanks to this found rapidly,  without the need for visiting mirrors and mirrors through the web browser, so I say found a repos where packages a bit older could help me get the result I wanted (mozilla-sunbird so that I still have the menus in French, before pinoc's patched package reached the repositories, and addlocale for 2 members for whom the latest one was not working).

So I think it is a valuable program which is worth maintaining and continuing.

Regards,
Mélodie

[Old-Polack]

melodie:

You seem to miss the main point. Whether the script goes through all the motions correctly doesn't matter if the speeds reported for the individual mirrors, in relation to each other, are inaccurate.

Hi O-P,

Are they really inaccurate ? Maybe there is the need for different methods of scanning the response of the mirrors, according to the kind and quality connection one has, but I don't think it is inaccurate.

I have sought for the fastest, and most often the one shown as fastest was really much faster than the one configured preceedingly ; I have recently sought for the slowest and less up to date, and thanks to this found rapidly,  without the need for visiting mirrors and mirrors through the web browser, so I say found a repos where packages a bit older could help me get the result I wanted (mozilla-sunbird so that I still have the menus in French, before pinoc's patched package reached the repositories, and addlocale for 2 members for whom the latest one was not working).

So I think it is a valuable program which is worth maintaining and continuing.

Regards,
Mélodie

The only time I tried it, the mirrors I hand chose, based solely on the reputation of the mirrors, were in fact faster than the choices the script picked for me. My results were similar to those of as and the others that posted about the results being inaccurate. That is why I never use the script in normal operations.

I don't care if the script is maintained or not. If I can't depend on the results, I still won't use it.

[melodie]

O-P, after I read the 3 posts before this one, and this one:

The problem is that the time depends also from available internet connection speed:

On a low speed connection (less than 500 Kbit/s) I would try 250 Kb filesize,
On a medium speed  (5 Mbit/s) =>  2.5 Mb filesize,
On higher speed ( > 10 Mbit/s) => 5 Mb filesize or more.

Not a simple task ....    Sad

Maybe a list, where you can select your connection...will it added soon 
Thanks for suggestions 

I think there is still hope for this program. Perhaps do you know other distributions do have a similar feature in the package manager, and that can be useful.

[Vorteggs]

I agree that the speeds estimated by the script are not always the highest.

But in general if you do that test many times (say 10 -20) for about 2 or 3 months the script shows you the top 5 for you in speed and reliability. So I personally find it useful.
O-P, have in mind that the server situation here in Europe may differ from USA. Actually the closest to me servers never got the top 10. But Swiss, Netherland and Finland do and they reside in my sources list in Synaptic with only one of them enabled and changed some time if necessary.

[Old-Polack]

O-P, after I read the 3 posts before this one, and this one:

The problem is that the time depends also from available internet connection speed:

On a low speed connection (less than 500 Kbit/s) I would try 250 Kb filesize,
On a medium speed  (5 Mbit/s) =>  2.5 Mb filesize,
On higher speed ( > 10 Mbit/s) => 5 Mb filesize or more.

Not a simple task ....    Sad

Maybe a list, where you can select your connection...will it added soon 
Thanks for suggestions 

I think there is still hope for this program. Perhaps do you know other distributions do have a similar feature in the package manager, and that can be useful.

I'm only aware of that in yum, which Texstar is already working on to adapt to PCLinuxOS use.

[daniel]

last change for you melodie... 
http://dl.dropbox.com/u/2077694/apt-sources-update-0.3.1-10leiche2011.i586.rpm

What changes?
- add ktsuss for root privilegs, note ktsuss must be installed
- add selection box to use internet connection

Selection box will only add low or medium internet connection...

good luuuuck... 

[melodie]

last change for you melodie...  
http://dl.dropbox.com/u/2077694/apt-sources-update-0.3.1-10leiche2011.i586.rpm

What changes?
- add ktsuss for root privilegs, note ktsuss must be installed
- add selection box to use internet connection

Selection box will only add low or medium internet connection...

good luuuuck...  

Leiche, your dropbox does not want to drop it !

PS: Do want !!!

[daniel]

try again...