From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20021003 Description of problem: The package manager is broken: it lists some installed packages as being not installed; lists others as being installed that I'm pretty sure are not. Then when I activate the "update" button by selecting and then deselecting one package, and click it, the package manager notifies that NN packages have been selected for deletion and MM packages selected for installation -- even though I've actually elected to make no changes! Unfortunately, the last time I tested this, it notified me that it was about to delete 32 components or packages, including gnome, the gimp, nautilus, metacity, file-roller, the control center and more. I have test this on all three of my RH 8.0 systems; the exact mistakes the Package Installer makes in identifying packages to install and delete have been different on each system; I have not made a record to compare whether the list of installed packages turns out the same each time the package manager is invoked. Version-Release number of selected component (if applicable): Not available viat the 8.0 menuing system How reproducible: Always Steps to Reproduce: 1. Foot -> 2. System Settings -> 3. Packages -> 4. Password for root -> 5. Look at the collected packages, compare with known installation -> 6. Activate the "Update" button by selecting or deselecting something 7. Click "Update" 8. Look at the results! Actual Results: Potential disaster Expected Results: Proper detection of installed and uninstalled packages Additional info: An unprepared person could trash their installation!
I inadvertently listed this as a redhat-rpm-config issue; it is a redhat-config-packages issue I apologize
Can you give an exact steps to reproduce this? I haven't seen it myself
I'm sorry, but I don't know how to be more specific than I was. This occurred on both of my Red Hat 8.0 installations, a Linux Hardware Solutions 1998 collage of hardware, and my IBM T-series 1.2GHz laptop (a 2001 collage of hardware). It has been consistently and reliably produced by going through the menus in the usual way, as I indicated in my first post. I have not compared the screen output from trial to trial. What other information would be helpful? Could I run a query of any sort and send you the output? If you need access to either machine, Michael K Johnson can provide it.
I've had the same problem as described here.
redhat-config-packages crashes before displaying list of available packages. To reproduce this, I do not touch the mouse or keyboard after invoking the command from the shell. $ redhat-config-packages Xlib: extension "RENDER" missing on display ":0.0". Xlib: extension "RENDER" missing on display ":0.0". Traceback (most recent call last): File "/usr/share/redhat-config-packages/MainWindow.py", line 163, in ? comps = im.readCompsViaMethod(hdlist, progress.update) File "/usr/share/redhat-config-packages/method.py", line 488, in readCompsViaMethod update_method = update_method) File "/usr/share/redhat-config-packages/comps.py", line 387, in __init__ self.readCompsFile(file, self.packages) File "/usr/share/redhat-config-packages/comps.py", line 471, in readCompsFile self.setupStateDicts(packages) File "/usr/share/redhat-config-packages/comps.py", line 420, in setupStateDicts for prov in h[rpm.RPMTAG_PROVIDES]: TypeError: iteration over non-sequence
I have had this same problem on my Sony VAIO FXA-36 laptop, while trying to install relatively insignificant packages with complicated sets of interdependencies (without a lot of interdependencies, it is still easier to use rpm) after the original install of 8.0. Config-packages trashed my installation to the point where I had to do a new install from the CD. It occurred following an upgrade from 7.3, and I wondered if it was the result of a poor upgrade. On other machines of mine, I have handled the redhat-config-packages program with kid gloves.
just signing up for cc
Yeah, I've watched the system threads while executing redhat-config-packages in 8.0. It exits out as soon as it calls on python. Rpm still works from the terminal,though; I've just fallen back on that.
*** This bug has been marked as a duplicate of 70972 ***
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.