topic title: [antiX M12] several little bugs
Posts: 96
melodie
Joined: 15 Feb 2008
#1
Hi,

I am not new to Linux, but a bit new to antiX. I just posted, and it took me some time, but the forums have decided to invalidate my session, so I lost my post when I hit submit.

The machine used : laptop Omnibook 6000 PIII 800 Mhz - 192 MB ram.

Here is a pic:

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://img15.hostingpics.net/pics/455262ramandswap.jpg"
linktext was:"http://img15.hostingpics.net/pics/455262ramandswap.jpg"
====================================


Little bug #1 : There is a little button on the top left, I
would like to get it not to be started with the session. It's the thingy that is used to see all the windows on the desktops. I'll be back with more little bugs soon... before the forums invalidate my session again, I post ! // SOLVED The solution is to uncomment the line which starts icewmtray in the startup script of the ~/.icewm directory. It fixes it correctly in a sure way.

Bad bug #2 : Midori crashes with message"illegal instruction", as well in a session using the antiX i486 kernel as one using the liquorix i686 kernel // SOLVED by uninstalling flashplugin-nonfree, not compiled with the set of instructions fit for that CPU

Little bug #3 : in 3 machines tested (virtual machine, PIII and a T30 PIV) there is one entry menu after install, and after first reboot there are three : how come ?[/ s] // SOLVED

Little bug #4 : when installing a new kernel, (the liquorix i686) the new entry is not written to Grub menu.lst // SOLVED but could't grub-update be invoked by a post install script ?

Little bug #5 : at modprobe 2 lines of message state:

modprobe : module unix not found in modules.dep
modprobe : module ext4 not found in modules.dep // Bug fix acknowledged

Little bug #6 : the keyboard key F10 triggers the desktop contextual menu, even when the console containing mc is selected, so that after I have changed the console configuration to be able to use F10 to quit mc, the desktop contextual menu appear just besides, unexpectedly. I would like to reconfigure it, so to get the contextual menu of the desktop being displayed only when I right-click on the desktop : how can I do that ?


Little bug #7 : I can't remove Gparted, or it will pull out

Code: Select all

     Remove the following packages:
1)     antix-cc-full-antix         
2)     system-antix
I would like it not to pull out the two above items. I don't know what"system-antix" contains, but for antix-cc I must say I like it very much. At same time, Gparted is really not useful in that one machine.

I think it the list of depends for Gparted there are a few which could preferably be removed, which generally come along. (libparted0debian1, libgtkmm-2.4-1c2a)

Little bug #8 : In the menu"Applications > Preferences" Gparted and Synaptic have two menu entries.[/ s] // SOLVED I removed one of each in the directory /usr/share/applications/ : because they also exist in /usr/share/applications/antix and are therefore displayed twice.

Little bug #9 : The package uzbl from Debian testing, a lightweight web browser, does not have a menu entry, and does not have an icon. Mostover, it is absolutely outdate, as of 2010, whereas there is a git repos with recent files.[/ s] // Partly SOLVED Here is a uzbl.desktop file which I pulled out of a midori.desktop file, and a uzbl.png icon which I pulled from the git repository (I downloaded the sources):

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://meets.free.fr/antiX/desktop-files"
linktext was:"http://meets.free.fr/antiX/desktop-files"
====================================


========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://meets.free.fr/antiX/images"
linktext was:"http://meets.free.fr/antiX/images"
====================================


The git place if someone would like to look at it:

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://www.uzbl.org/get.php"
linktext was:"http://www.uzbl.org/get.php"
====================================


Else, I have broken the install a bit by trying to uninstall fluxbox. __{{emoticon}}__
I repaired it with packages I made with dpkg-repack which I installed in another box, and uploaded the 4 packages in the same space.

Thanks for the paste, present and future improvements.
Last edited by melodie on 13 Aug 2012, 15:23, edited 7 times in total.
Posts: 325
male
Joined: 04 Nov 2011
#2
melodie wrote:Little bug #4 : when installing a new kernel, (the liquorix i686) the new entry is not written to Grub menu.lst

Code: Select all

# update-grub
and reboot. __{{emoticon}}__
Posts: 279
afab4
Joined: 17 Oct 2009
#3
melodie wrote: Here is a pic:

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://img15.hostingpics.net/pics/455262ramandswap.jpg"
linktext was:"http://img15.hostingpics.net/pics/455262ramandswap.jpg"
====================================


Little bug #1 : There is a little button on the top left, I
would like to get it not to be started with the session. It's the thingy that is used to see all the windows on the desktops.
I will address the"bug" you have described.
This little window is actually caused by the volume icon not being fully loaded. I have installed antiX-12 on several laptops and the fix requires a bit of fiddling with the home/~/.icewm/startup script. You can easily modify this by using the antixcc and Edit IceWM settings.
At the bottom of the script are these 3 lines. You may not have the 3rd line in your script.

Code: Select all

(sleep 3 && volumeicon) &
(sleep 5 && toram-eject) &
(sleep 5 && sudo persist-config --startup) &
What you need to do is change the delay on the volumeicon line to

Code: Select all

(sleep 5 && volumeicon) &
(sleep 5 && toram-eject) &
(sleep 5 && sudo persist-config --startup) &
You may have to experiment a bit with different settings. I believe I had to set the volumeicon delay to 7 on one of the machines I installed on.
Finally, if I recall correctly, I could always log out/log in and the little window would disappear, even in a live session.
Posts: 96
melodie
Joined: 15 Feb 2008
#4
Thank you male, and afa4 ! For grub, I have already edited the menu.lst, but I would remember for next time, that a simple"update-grub" as root will do the trick. For the fiddling with the icewm config, I'll do that tomorrow, and hopefully it will solve it. I hope the other little annoyances will also get sooner or later a solution.

(And within a few days I will have to come talk about how it goes in another machine... and perhaps I'll play the install again once before).
anticapitalista
Posts: 5,956
Site Admin
Joined: 11 Sep 2007
#5
Bad bug2 is not an antiX bug as far as I can tell (we ship with iceweasel), unless you mean that midori crashes only in antiX and not Debian testing (on which it is based).

Little bug 3 - is a feature! __{{emoticon}}__ The first entry is set via the installer, the other two, after an update-grub is triggered on first installed boot so that grub is configured correctly.

Little bug 5 is a Debian bug that is 'fixed' in Sid/unstable. It is completely harmless.
Posts: 96
melodie
Joined: 15 Feb 2008
#6
Hello anticapitalista,

First thing, I want to tell you I found your recent versions (RC2 and final which I tested and used just these days) mega brilliant ! It is a fantastic tool ! And even if you already know it, I want you to know that I find it great ! With this post a list of little bugs, I mainly want to bring a modest contribution, while eventually solving one or two that are bothering me a bit.

About Midori, I will ask people using Debian testing, because I am mainly an Archlinux user, and at Archlinux (my box is a 64bits) I have no problem using it whatsoever.

This makes me think : what version of Midori is available in antiX ? Aha ! after a"apt-cache show":
version is 0.4.3-1.

In Archlinux, a"pacman -Si midori" tells me:
Version : 0.4.6-2

Could we fancy that the newer version (without illegal instructions) could be compiled and packaged for antiX ? (Debian does not have newer than 0.4.3, even in unstable). (I don't offer to do it, because the first and only package I did took me not less than one week : packaged"
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://code.google.com/p/mimarchlinux/wiki/OpenBoxMenu"
linktext was:"openbox-menu"
====================================
" for Debian, then for PPA Ubuntu).

About feature"little bug # 3" : I would like to ask what"nouveau.modeset=0" does in the system ? This would allow me to know if I can reduce the number of entries. ie: if this is a tweak dedicated to nvidia gpus, this machine having an ATI Rage3 it will not be necessary. I didn't find any other difference with entry number 2, as the first one is also for the same smp i486 kernel.

I have found another little bug, which I would like to solve, I'll add it as #6 is the list of the first post. It is related to keyboard shortcuts, F10 triggers the desktop contextual menu, even when the console containing mc is selected, so that after I have changed the console configuration to be able to use F10 to quit mc, the desktop contextual menu appear just besides, unexpectedly.
anticapitalista
Posts: 5,956
Site Admin
Joined: 11 Sep 2007
#7
melodie - thanks for the comments and yes I do see your bug list as a positive contribution to improving antiX.

For midori I would prefer not to package the latest, which means users will just have to wait for it to enter Testing. The reason being that I want antiX to be totally compatible with Debian Testing and not add/remove any necessary libraries for a later version of an application. It *might* break at some point, not just the app itself, but maybe other things as well.

nouveau.modeset=0 - disables KMS

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://nouveau.freedesktop.org/wiki/TroubleShooting"
linktext was:"http://nouveau.freedesktop.org/wiki/TroubleShooting"
====================================


========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"http://nouveau.freedesktop.org/wiki/KernelModeSetting"
linktext was:"http://nouveau.freedesktop.org/wiki/KernelModeSetting"
====================================
Posts: 96
melodie
Joined: 15 Feb 2008
#8
Someone told me it could be a matter of flags present on the machine of the packager and not present in the machine which I am using now. He also added that it could come from webkit rather than from midori.

So I went to wikipedia to seek for other browsers using webkit, and available at Debian. I found uzbl, which crashed. the Wikipedia page also mentioned Chromium. I installed it, and I am now using it : does it use webkit ? It does not seem to crash.

Here are the specs of the cpu Coppermine in this machine:

Code: Select all

$ cat /proc/cpuinfo 
processor    : 0
vendor_id    : GenuineIntel
cpu family    : 6
model        : 8
model name    : Pentium III (Coppermine)
stepping    : 10
microcode    : 0x4
cpu MHz        : 796.531
cache size    : 256 KB
fdiv_bug    : no
hlt_bug        : no
f00f_bug    : no
coma_bug    : no
fpu        : yes
fpu_exception    : yes
cpuid level    : 2
wp        : yes
flags        : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pse36 mmx fxsr sse up
bogomips    : 1593.06
clflush size    : 32
cache_alignment    : 32
address sizes    : 36 bits physical, 32 bits virtual
power management:
$
I have asked some Debian users, one told me it works. Another one tested in vbox and told me it works (but the test was too short, here if I don't do anything in the browser it can stay still a few seconds). And one tested in Crunchbang in vbox and said it works.

I tried to find out what compilation flags were used in the webkit version which is available at antiX but could not find out.

I still hope these few elements can help debug (ie: the proc here does not have sse2 in the flags).

Thank you for the nouveau.modeset=0 information.
Posts: 96
melodie
Joined: 15 Feb 2008
#9
Hi,

I am posting from within Midori. The problem was coming from flashplugin-nonfree. I'll be back from within the other machine from where I was logged in"ssh -X", to show the gdb output.

/Edit : here it is:

Code: Select all

$ gdb midori 
GNU gdb (GDB) 7.4.1-debian
Copyright (C) 2012 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type"show copying"
and"show warranty" for details.
This GDB was configured as"i486-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/bin/midori...(no debugging symbols found)...done.
(gdb) run
Starting program: /usr/bin/midori 
[Thread debugging using libthread_db enabled]
Using host libthread_db library"/lib/i386-linux-gnu/libthread_db.so.1".
Xlib:  extension"RANDR" missing on display"localhost:10.0".
[New Thread 0xb36c1b70 (LWP 5812)]
[New Thread 0xb2da1b70 (LWP 5813)]
[New Thread 0xb251ab70 (LWP 5821)]
[New Thread 0xb1d1ab70 (LWP 5823)]
[New Thread 0xac1f3b70 (LWP 5861)]
[New Thread 0xaad15b70 (LWP 5869)]
[New Thread 0xaa515b70 (LWP 5870)]
[Thread 0xaad15b70 (LWP 5869) exited]
[Thread 0xaa515b70 (LWP 5870) exited]
[New Thread 0xaa515b70 (LWP 5896)]
[New Thread 0xaad15b70 (LWP 5898)]
[New Thread 0xa99ffb70 (LWP 5899)]
[New Thread 0xa91ffb70 (LWP 5900)]
[New Thread 0xa89ffb70 (LWP 5901)]
[Thread 0xa89ffb70 (LWP 5901) exited]
[Thread 0xaad15b70 (LWP 5898) exited]
[Thread 0xa91ffb70 (LWP 5900) exited]
[Thread 0xaa515b70 (LWP 5896) exited]
[New Thread 0xaa515b70 (LWP 5913)]
[New Thread 0xa91ffb70 (LWP 5914)]
[Thread 0xa99ffb70 (LWP 5899) exited]
[Thread 0xa91ffb70 (LWP 5914) exited]

Program received signal SIGILL, Illegal instruction.
0xa5bf2aa0 in ?? () from /usr/lib/flashplugin-nonfree/libflashplayer.so
(gdb) quit
A debugging session is active.

    Inferior 1 [process 5798] will be killed.

Quit anyway? (y or n) y
$
after that, I did a"aptitude purge flashplugin-nonfree" and I was able to use Midori normally to come post here. Great find ! Thanks again to captnfab who gave me the idea to use gdb, I had not think about it.
Posts: 96
melodie
Joined: 15 Feb 2008
#10
Hi,

I marked solved several in the list and added 2. I didn't solve the button bug yet, although I have tried several sleep times for volumeicon : is it sure that it is volumeicon which is concerned ? The button is related to displaying all windows, so it seems a bit strange that it would be triggered by a volumeicon icon being slow to start ?

I will comment out volumeicon in the startup script to see what happens then.
Posts: 609
dark-D
Joined: 02 Jun 2008
#11
hi,
bug 1: if volume icon doesn't make the window, sometimes i suspect that wicd makes this ghost window.
bug 4: you could use smxi to install kernels. it will install the kernel, headers, update grub and i think it installs gcc also.
bug 6: edit your keys setting from control center for icewm or fluxbox depending on what you are using.
bug 7: gparted is a dependency of antix-cc so i see the reason it wants to remove both of them. if you want free space deleting gparted won't give you lots of it. here is what i use to get more space on my installs: post18498.html#p18498
Posts: 96
melodie
Joined: 15 Feb 2008
#12

Code: Select all

bug 6: edit your keys setting from control center for icewm or fluxbox depending on what you are using.
bug 7: gparted is a dependency of antix-cc so i see the reason it wants to remove both of them. if you want free space deleting gparted won't give you lots of it. here is what i use to get more space on my installs: post18498.html#p18498
I will try your fix for #6, and for #7 it's not for space, there is enough, but for the program itself. I try to reduce the number of programs which are not needed in that box. The menus are very full of programs not needed (I would not have dared start with a basis version because I don't know the distro enough yet, to ensure getting a result at least as good).

By reporting it here I thought it would be a good idea to suggest a different tree dependency for Gparted.

I have commented volumeicon and the ghost window was not there anymore. I uncommented it, put 7 seconds, then it's not there but the screen flickers several times before staying still. I will try to look into the .xsession-errors file next.

n# 5 : I don't know the tool you are saying about. I will not change the kernel anymore, and for future uses on other machines I know the trick now. __{{emoticon}}__

Thanks for your answers.
Posts: 96
melodie
Joined: 15 Feb 2008
#13
Hi,

Here is the content of the .xsession-errors after just logged in :
The only warning I know I can fix it the shortcut for Midori. Do you have suggests to fix the other warnings/error messages ?

Thanks.
Posts: 96
melodie
Joined: 15 Feb 2008
#14
Hi,

Bug number 1 is solved ! and not by adding seconds to the volumeicon, but something else.

First, thanks to someone at the irc.freenode.net #icewm chan who has known at once where the trouble was coming from. Here an extract of our discussion (with his kind permission):
<mythos> melodie_, sorry, which window appears?
<melodie_> hi mythos in the pic, the very little thingy on the top left of the screen
<melodie_> just at the left of the htop window
<mythos> melodie_, is that a trayicon? do you have startet icewmtray (or was it icetray)?
<mythos> *have you started
<melodie_> it's not icewmtray
<melodie_> it's a full mystery
<mythos> no, i mean, icewmtray handles the trayicon-area for icewm. if that prozess does not run, many programs draw a seperate window for only this icon.
So what I did is re-open the file ~/.icewm/startup and I noticed the icewmtray is commented. I commented out and go !

Here is the startup script of mine now:

I have commented out wicd because it is not and will not be used, but this is a detail which is not related to the problem.

The volumeicon sleep time to wait is 3 as before, as you can see. I have shutdown and even unplugged from power source twice before booting agin since, to be sure, and the ghost window didn't re appear. __{{emoticon}}__
anticapitalista
Posts: 5,956
Site Admin
Joined: 11 Sep 2007
#15
Good catch there melodie!

You know, I must have commented out icewmtray for a reason, but for the life of me cannot remember why.

I'll add this to the bug fix.

thanks