Description of problem: Fresh mga5 install on new Lenovo W541. Trying to install wine; but get: # urpmi wine The following packages can't be installed because they depend on packages that are older than the installed ones: libdri-drivers-10.5.7-3.mga5.tainted libmesagl1-10.5.7-3.mga5.tainted libmesaglu1-9.0.0-5.mga5 wine32-1.7.30-1.mga5 wine-1.7.30-1.mga5 # urpmi wine64 The following packages can't be installed because they depend on packages that are older than the installed ones: libdri-drivers-10.5.7-3.mga5.tainted libmesagl1-10.5.7-3.mga5.tainted libmesaglu1-9.0.0-5.mga5 wine32-1.7.30-1.mga5 Trying this on my daughter's new Lenovo. Wine{,32,64} are installed on my Dell M6800; though it shows: libdri-drivers-10.5.9-3.mga5.tainted** v. libdri-drivers-10.5.7-3.mga5.tainted ** This is in the updates; but never seen in mcc. Version-Release number of selected component (if applicable): wine-1.7.30-1.mga5.i586 How reproducible: always Steps to Reproduce: 1. Install mga5 2. try to install wine. Wine64 installed on first fresh install of mga5, though I didn't install it yet after mga5 re-install due to a great many packages removed while uninstalling wine64 to retry -- this actually killed system (mcc window went mostly black during that aggressive uninstall. So did firefox textareas, etc; destroying initial mga5 install. I've done dozens of distro installs since 1998 (Mandrake, Mandriva, Mageia, and others); this is the most disastrous ever.... Did a new mga5 install just a few weeks ago on a friend's laptop. Reproducible: Steps to Reproduce:
That's because you started doing an update to the new mesa 10.5.9 but it was not completed. You must keep Tainted Updates activated if you use Tainted Release, else you will end up in such situations where Core Release/Updates and Tainted Release (without Updates) conflict with each other.
Status: NEW => RESOLVEDResolution: (none) => INVALID
See http://madb.mageia.org/package/show/application/0/name/libdri-drivers And this comment to get the full list of packages that have to be updated if they are installed on your system: https://bugs.mageia.org/show_bug.cgi?id=17384#c0 (but of course just enabling Core Updates and Tainted Updates and doing the regular updates afterwards should work fine).
(In reply to Rémi Verschelde from comment #1) > That's because you started doing an update to the new mesa 10.5.9 but it was > not completed. You must keep Tainted Updates activated if you use Tainted > Release, else you will end up in such situations where Core Release/Updates > and Tainted Release (without Updates) conflict with each other. That seems like a reasonable guess; but not what happened. I installed mga5 on the new laptop (resized Windows 10) partition) using default KDE install. Wine is not part of that install. Installed wine64 (non-tainted), then tried to install notepad++ as a test and that failed because win/wine32 were not installed. Tried installing those and it failed. Checked wine on my laptop (installed a LONG time ago) and discovered the tainted packages were installed, so I enabled ALL local and remote "release" and "update" package sources and tried again which became a disaster by auto-removing tons of packages until the system became corrupted. Re-installed from scratch; enabled ALL local and remote "release" and "update" package sources which resulted in the initial report... > (but of course just enabling Core Updates and Tainted Updates and doing the > regular updates afterwards should work fine). Are you implying that I should have only enabled Updates and disabled the {Core,Tainted} Release? For my first time ever, mcc could never select wine or wine32 for install. http://advisories.mageia.org/MGAA-2016-0002.html shows that libdri comes in Core an d Tainted versions-- so I'm puzzled why 10.5.7 was available; but not 10.5.9 even though 10.5.9 was in the repos... Could this be due to another occurrence such as in https://bugs.mageia.org/show_bug.cgi?id=13461#c1 where I got burned by MD5SUM being wrong for a short period of time? Well... looks like that bug bites again: The file in mirror.nexcess.net mirror is: -rw-r--r-- 1 pfortin pfortin 576 Nov 7 15:10 MD5SUM [1] and the he file in my repo (never seen by the install I was doing) -- only useful to prove that the file was updated, then reverted rather than corrected: -rw-r--r-- 1 pfortin pfortin 576 Jan 15 01:52 MD5SUM [1] Needs to have more current date in source and mirrors. So once again, a newer MD5SUM is replaced by an older one and any local mirroring will never get older files... This is a race condition which may not affect all people; but one I got caught in again due to timing. :( Will check later when the laptop arrives back in Canada...
Status: RESOLVED => REOPENEDResolution: INVALID => (none)
This package has no registered maintainer so assigning to packagers collectively.
Assignee: bugsquad => pkg-bugs
It's clearly not a wine issue though. Comment 3 shows that it would be a local mirror syncing issue, I think it would be worth opening another issue with a clear description of what the issue is if it can be solved on Mageia's side, and close again this one as INVALID as I did previously.
Source RPM: wine-1.7.30-1.mga5.i586 => (none)
BTW if you are using 32-bit applications, make sure to have Core 32bit Updates and Tainted 32bit Updates enabled too (+ the Release repos of course)
Hi Pierre, Thank you for having taken the needed time to report this issue! I'll close this report as OLD, because Mga5 stopped getting any other updates than security updates. Even if there might be a real bug underneath what seemed like a wine issue, it is indeed better to open a separate report, as Rémi suggested in comment #5, if you can pinpoint (maybe with help from the forums) what that is and it still happens in Mageia 6 or later. Kind regards, Marja
CC: (none) => marja11Status: REOPENED => RESOLVEDResolution: (none) => OLD