Red HatThis forum is for the discussion of Red Hat Linux.
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
I have a big problem concerning my display. Here i try to give as many details as possible, i hope someone can help me.
i have Red Hat 9.0, kernel 2.4.20-31.9 installed on a laptop: Dell inspiron 600m. My video card is an ATI rage mobility 9000 with 32 Megabytes.
i don't know Gnome version, but if i write "pkg-config --modversion gnome" i get as an answer "220.127.116.11"
Everything worked fine at the beginning, then i decided to install gxine from source.
I had installed the following libraries:
but gxine needed gtk+ >= 2.4.0 so I installed from source the following libraries
first i tried to install by using
it seemed to install everything smoothly, but it didn't work, gxine couldn't see the libraries, then i tried
and it worked fine, i have been able to compile correctly gxine and it worked perfectly, even with firefox plugins. I have to say that i called ldconfig here and there, i know that this is quite a "windows" way of doing things...
Then i shut down the computer, and the next time i turned it up it was unable to load the display... i can still log in as a single user in text mode, but the display can't be loaded. I see the usual log in screen for a fraction of a second, then i get a blank screen and the mouse pointer becomes a small black cross. if i try to move the mouse the blanck screen goes away and i get and ascii screen with a small smile in a corner, prompting me to choose yes or no, i think the system is trying to load the display. if i answer no, i get two or three times the same screen, then the system says that it tryed to install the display but failed and i get the blanck screen with the cross pointer and i am unable to do anything. If i answer yes i get the same screen, but with unreadable ascii characters.
Re: red hat 9: conflict between glib-2.6.3 and gnome?
first i tried to install by using
it seemed to install everything smoothly, but it didn't work, gxine couldn't see the libraries,
/usr/local is the default prefix for most source code tarballs. Effectively, you also installed pkg-config templates below /usr/local, where your system's pkg-config tool doesn't find them unless you adjust its search path.
then i tried
And with that, you overwrote your system's libraries and broke many dependencies.
I have to say that i called ldconfig here and there, i know that this is quite a "windows" way of doing things...
Calling ldconfig only affects run-time, not compile-time.
You can downgrade the listed packages with rpm --oldpackage -Uvh ... using the older versions which come with Red Hat Linux 9.
how can i install new libraries without breaking dependencies?
That depends on whether multiple versions of the same library can coexist and where you install them. Non-RPM based installation of files directly into your system's core directories is not recommended.
sorry, but i am a very very newbie, so you mean that i have to "upgrade" the system with the package manager, so i will overwrite the new bad libraries and reinstall the old ones, also restoring the dependencies.
and, if i want to install new libraries such as glib, i should also use rpm? did i understand right?
I tried to do rpm--oldpackage -Uvh ... but it doesn't work, i am told that packages are already installed. I tried also with --force, rpm installs, but if i try "pkg-config --modversion glib-2.0" i get always "2.6.3"
i also discovered that, if i login as root in text mode and launch "starx" gnome works... it seems that problems comes out with the login screen, is it possible.
To only replace existing package versions (instead of downgrading them), other options are needed, e.g. --replacepkgs -Uvh
Any files not installed via RPM, and not tracked in the local RPM database, would still be present in the file system. E.g. your newer glib pkg-config template in /usr/lib/pkgconfig. You would need to remove such files manually (if a make uninstall target is not available).
Whether the graphical greeter and its dependencies are damaged, is difficult to examine remotely. Did you use GDM? /usr/bin/gdm* and package "gdm". You could verify the binary and see with "ldd" what libraries it needs and which ones it finds.
I'm surprised as a newbie you upgraded core packages with software built from source tarballs.
I have a "make uninstall" target for each of the libraries: if i untar the sources and then:
pkg-config doesn't see glib anymore
do you think that uninstalling this way and then downgrading would work?
installing glib 2.6.3 from rpm would also be a correct option?
You are surprised i installed from source? i still didn't know of dependencies hell... i didn't know i was upgrading core packages... and finally i reasoned as i was working with windows... to bad ... but now i know something more...
i don't know what's gdm, and ldd, but i can learn quickly i hope...
To the left in the display are library names. To the right you see the path where they are found in the filesystem. Possible damage I could think of is that incompatible library versions are picked up. Hence I would run "rpm -qf" on every path displayed at the right and see whether the corresponding RPM package is okay, e.g.
You can also verify all installed RPM packages, e.g. with "rpm -Va" or "rpm -qa | xargs -n 1 -t rpm -V &> rpm-Va.txt" The output is displayed in "man rpm". You should not see many changes files. Mostly config files should be marked as '5' (i.e. MD5 checksum changed).
You could also look for log errors logged in the X server's log file below /var/log.