All Projects

The new multiple screens betatesting is ready to be used, this task will be to better track all that is needed to betatest:

How to use it: Everything should be automated, so just upgrade your system with the command “apug” and restart to a new desktop configuration (from elive menus), which is important too. It should detect that you have multiple screens and ask you to configure them, or use an existing configuration, etc in contrary case you still have the configurator available (only one now, which is an elive tool)

Things to betatest: This is the list:

1. all the options and behaviours (what if select no, yes, save, not save, remove, etc), this is more or less already betatested by me
2. try to disconnect the extra screen, reboot computer, your desktop should be equally functional (and I assume not asking/saying anything about)
3. poweroff again the computer, connect again the extra screen, it should remember that configuration for your second screen and going back to these settings (without asking anything, i assume)
4. there’s some games that changes your screen resolution (in order to play in fullscreen), if the game exits with an error your desktop is messed up (desktop running in a strange resolution), the normal behaviour is to return to the correct resolution when the game exits correctly, but i dont know if this happens or even if goes to the correct multiscreen mode. Also: a reboot should perfectly return to the normal behaviour
5. maybe already said before, but switching between “using and not using” the external screen should be not a painful experience (and/or with reboots), it should be “work on the fly”
6. ecomorph (special effects) should work correctly, except for the “exposé effect” which doesn’t seems to like it so much (confused numbers of desktops, etc, not sure if this one works better after the full implementation)
7. detail: technically speaking, there’s two launching of the “load screen settings”, one happens before the desktop is loaded and the other when the desktop has started to load... in other words: if there’s a widget/question that asks for anything it should always happen when you already have access to the window controls (including window borders for example)
8. since i found a strange behavior with the “automated fonts sizing calculation” feature, i include this betatest point: test if the font sizes are not reconfigured “smaller than normal”
9. we have hotkeys to switch between virtual desktops, do we have a setup for between screens? i just tried to configure a new keyboard hotkey (ctrl + alt + shift + arrows) for switch to another screen 0/1 but doesn’t looks like to work on virtualbox... will need to do test in a real machine (and implement the hotkey in the default configuration)

Notes: if any of these betatests are not working as expected, please include the details of the “steps” about how to reproduce it, and the exact (first words) of the “conflicting” message found

Ubuntu doesn’t set any password to root by default, this is a very good idea (ironic), so that the migration mode of the installer works perfectly but when the user want to do anything with root... can’t do anything, cant login, nothing... since there’s no password

Add a verifier of root passwords in the migration/upgrade mode so that will request it if is empty

I will need a sample line of the root file / etc/shadow in order to know how to detect it

I just tried to use Google Talk for videoconference (calls) but seems to not work, I tried it on chrome and even installed a .deb package provided by google, it complains about that is not correctly installed or something similar, I will need to make future look to it

Lastest e17, after to finish wizard go fast to the applications menu and scroll it, e17 freezes

This bug reported by AlexAnteMachina has nothing to do with composite, not even with vbox, I tried it in laptop and e17 simply gets frozen, seems like X11 is died, very strange bug! even with my libs updated without the optimize options (please somebody confirm that we have debug messages now), there’s simply no way to get a backtrace because everything is blocked (well, you can kill X11, but you lose everything)

Maybe there’s not even needed to select German language, it looks like an efreet bug

This is another bug, but seems like caused by the same reason, sometimes, and just sometimes, e17 starts looking like this (some hang in the startup of efreet, or uncomplete or similar, idk)
http://main.elivecd.org/tmp/enlightenment/shot-2013-06-01_22-24-45.jpg

A restart of e17 will solve the problem, but is not a good way to start e17 of course

Another bug: if you install eltrans (here is elive in version 2.1.40), and you try to open the menus, it becomes frozen

• If you create a new user, it can’t login to entrance since you dont wait that the selector of graphical system appears (at least having other graphical systems in the list like e16 listed in the list of window-managers to select (top-left corner))

Is not possible to erase a cdrw device since thunar automounts it directly and only has the “eject” option, maybe fix that in ivman ? (if cdrw, popup to ask if umount)

There’s needed some extra modules in the initrd in order to boot elive from an usb that are plugged in (some) pcmcia devices

Not much info, i just know that it not works, so:

• see if the module loads correctly (modprobe nvidia and run dmesg too)
• see if the version of the module package and the glx package are the same version
• read X logs
• use the xorg.conf of the installed system

Installing in /dev/hdb1 for example, grub generates a configuration using (hd0,0), where it should be (hd1,0)

The questions is...

• with which disk you boot the computer ? hda or hdb ?
• where you have selected to install grub ? (if the installer has asked you this)

I think that the selector of grub needs to be rewrited... in case that is not hda or sda, ask allways to the user about which/where to install grub, a better look needs to be made on this side

please comment this ticket, i need your info for fix it :P

The actual version of flash 10.0.12.36-elive1 is unstable, firefox crashes (like in badoo.com), for the stable version of Elive we need to use a newest (stable) version or go back to a backup

Aparently if you install Elive with / as JFS and rebooting from the e17 menus, the JFS root partition is mounted as RO in the next boot

Please do details about how to reproduce this bug, in a updated version of elive (< 1.9.30)

run this command from the live mode using CDROM (no usb): dmesg | grep -i error | grep “I/O” and please paste in the comments a block of ‘code’ mode the result

smbmount is required for smb-browser, but the package smbfs is not available in debian wheezy right now, i’m not sure why, maybe is replaced by the package cifs-utils, or maybe it enters back in the future (release of wheezy stable), in any of the cases i leave this report opened to know that something needs to be fixed around :), which means adapt smb-browser to the new tool or get some kind of port

PS: in any of the cases, smb-browser is not a strict requirement for Elive since thunar is able now to browse samba networks in a friendly way

PrinceAMD has a computer which when kms loads, the console turns blank, user can’t see anything, but the system continue booting correctly (1-2 minutes in blank)

Maybe should be good to add a boot option in the live menu with this entry: nomodeset that solves the problem, for those cases

Seems like wifi doesn’t automatically connect when i come back from suspension and im in another location (and I enable the wifi power-button later too), i need to open the internet configurator emodule for to start/open the Network-manager applet (note that by default is removed in order to save some resources consumption and memory usage.

The question is... do we need this nm-applet running all the time so that the network will be correctly connected/updated all the time? this sounds for me non-sense since there’s a daemon of NM, the applect should be only for manual selection or configuration of things, otherwise a daemon should works a daemon.

It happens to you too ?

By some reason the first boot doesn’t allows the user to suspend/hiberante, what is the cause?

* debug pm-suspend
* try with e23 first boot
* try with reboot after post-install of eliveinstaller

simple test of mail server

a patch from debian (about xscreensaver) made it that doesn’t run, rebuild mplayer (patch to remove patch actually added to the builder of packages of elive)

• If you plug a linux* disk, thunar (exo-utils) mount it without uuid=user mode (so root owner),

this not looks like to be very correct, user can’t access/modify files

• If you insert a Cdrom to record it (like cdrw), brasero (burner) doesn’t has permissions to do anything with it,

thunar doesn’t has also a “umount” option but that’s other case

• remember that big problems requires big solutions

Looks like grub is not installed correctly on the MBR, in a clean hard disk, automatic+easy mode install, (automatic grub install), elive 1.9.14

There’s a kind of mess between these 2 interfaces (wifi and wired), i removed all the netconnector configurations and then started the daemon and configured only for my wired connection, that is a static ip with autoconnect=yes, after to reboot i have:

A different ip (a dhcp one) for my wired interface, if I run netconnector (gui), in the Wired part i have my static ip set/configured, and in the Status values says that i have a different one (this dhcp obtained one), if i run ifconfig i have the static ip set to the WIFI interface instead of the wired one. Again in the GUI, in the Wifi part, i have the “autoconnect” box set in ON (of course i have not set it, maybe the problem is from here?), and the status for the wifi says that I have the static ip.

A more descriptive image can be found at http://thana.no-ip.org/shots/deleteme-netconnector-strange.jpg

My configurations are:

root@Babylon:~# cat /etc/network/interfaces-eth0 
iface network inet dhcp
root@Babylon:~# cat /etc/net-connector/profiles.conf 
[ath0]
[eth0]
[[casa]]
wireless = No
wireless-mode = None
essid = None
encryption_key = None
wpa-driver = None
encryption = None
dhcp = No
gateway = 192.168.2.1
address = 192.168.2.166
netmask = 255.255.255.0
auto_connect = yes
dns1 = None
dns2 = None
root@Babylon:~# 

There’s a big problem with the updates of the system that will broke the graphical system, this bug has Critical severity

The problem is that, if you upgrade your system, it will update the GLX package of your ati/nvidia drivers, this means, that you will have a X version of the kernel with an ati (or nvidia) module in the version Y, that is incompatible with the updated version Z of the GLX package

To solve that problem, I have used all this time a block (hold mark) of the GLX package (the user should have the same kernel all the time)

But the problem is, for example, that if the user update the system that include an updated kernel, and an update of the ati kernel module, the GLX package for ati is not updated so it is incompatible and the user can’t reach the graphical system

Actual case: updated system: update the kernel, but the packages fglrx-driver, fglrx-glx and fglrx-kernel-2.3.26.8-elive-686 are holded, so not updated, the kernel can’t simply load that old version of the ati driver and the user can’t reach the graphical system

If we hold the packages, there’s the problem that everything needs to be updated (so, do not hold anything) in order to work correctly. But if we don’t lock them, also if the user upgrade the system, will use the same version of the kernel, so then the ati GLX package will be updated and not work with the old kernel/module, the only solution for that and everything is that update the version of the kernel too (with a virtual-package for example)... this was requested by some users but, it is a good idea ? what happens if a user dont want to update his kernel ? what about if a version of the kernel mades the computer of a user to NOT boot or not work ? and finally... the grub menu needs to be updated on that case then

Comments about this problematic issue ?

After the verifying speeds steps of the installer, if you have a NTFS partition, the /root files will be vanished (see how your gtk theme of the installer looks ugly now)

Solutions: update aufs version (kernel) or copy the root files back

Note: This DOESN’T affect to the installed system, it is just a bug for the live mode, installed system is perfectly clean and stable... if will be not the case, Elive won’t use this way :))