Hi. I had a firefox installation problem which I have now fixed, but don't really understand why my solution worked.

My question is, "Do you get different rights if you log in as root, compared to logging in as a normal user and running a terminal and su ing to root? If so, what is different? If there IS a rights issue, could I have got around it by opening a new session with <ctrl><alt><Fx> and logging in as root to perform the install?"

The reason I ask is because of the following experience:

I run SuSE 9.1 and use firefox. A while ago, I downloaded an updated version of firefox and installed it to /usr/bin/firefox/, then ran it with no problems. (As an aside, is /usr/bin/ where I should be installing software to?)

I recently performed an online update through YOU, and everything was updated except for the firefox installation, which failed. I downloaded a newer version of firefox and tried to install it, and got the message:

xlib: :connection to :0.0" refused by server"
xlib: protocol not supported

All my kernel sources are installed, and xhost+ etc didn't help.

I then connected YAST directly to SuSE.com and removed firefox. This removed the original, but not the one I installed myself earlier, which continued to work fine. I reinstalled firefox using YAST and failed again to update it using YAST. I uninstalled it again (my seperately installed copy still worked after this) and once again failed to install the newer copy I had downloaded.

The failures happened when I was logged in as a normal user and ran the firefox-installer as root from a terminal.

I then tried logging in as root and ran the installer from the terminal and it worked.

Could anybody shed some light on this for me, please?

Mike.

If you installed Firefox manually (as in not via YAST or as an rpm) then YAST won't pick it up - it only sees rpms. Firefox normally installs into it's own directory, deleting that directory should uninstall that version of FF. Then you'll need to manually remove any icons.

To give su full root rights, issue and use that. Although normal su is usually sufficient for my needs.

Ah! On further investigation, it seems that 'su -l' resets all enviroment variables and the path. I take it that the reason I got my errors before is because after running su on it's own, there was no path to the kernel sources?

So would it be right to say that the permissions are the same if you use 'su' or 'su -l', but functionality may be lost due to the configuration of environment variables?