Description of problem: I've installed texlive-2009, which still uses libpoppler.so.5. This causes a lot of packages that can't be updated, and yum chokes (see part of the error message below). Version-Release number of selected component (if applicable): yum-3.2.27-11.fc14.noarch How reproducible: Almost always. Steps to Reproduce: 1. yum update --skip-broken 2. 3. Actual results: Lots of messages in the line of --> Finished Dependency Resolution Skip-broken could not solve problems Error: Package: texlive-pdftex-bin-2010-5.20100430.fc13.16392.fc13.x86_64 (@texlive) Requires: libpoppler.so.5()(64bit) Removing: poppler-0.12.4-2.fc14.x86_64 (@rawhide) libpoppler.so.5()(64bit) Updated By: poppler-0.13.3-2.fc14.x86_64 (rawhide) Not found Error: Package: gnome-panel-2.30.0-1.fc14.x86_64 (@rawhide) Requires: libedataserverui-1.2.so.8()(64bit) Removing: evolution-data-server-2.30.0-1.fc13.x86_64 (@rawhide) libedataserverui-1.2.so.8()(64bit) Updated By: evolution-data-server-2.31.1-1.fc14.x86_64 (rawhide) Not found Expected results: Detect non-updatables, update the rest... Additional info: texlive-2010-5.20100430.fc13.x86_64 (and a selection of others from there) installed. Installing a lot of non-conflicted packages by hand ("yum upgrade --skip-broken list-of-packages") usually then makes the next try at "yum upgrade --skip-broken" go through.
could you please attach the output from 'yum -d9 update --skip-broken'
Created attachment 414377 [details] Requested output from "yum -d9 -y update --skip-broken"
Created attachment 414470 [details] patch to fix skip-broken loop detection issue
Look like the skip-broken loop detection bails out to early. Could you please try to add the attached patch to /usr/lib64/python2.6/site-packages/yum/__init__.py to see if it solves the issue.
Seems to fix it (no new updates, but yesterday after updating everything that could be it complained also). Will report back after today's round of updates. BTW, here the file is /usr/lib/python2.6/site-packages/yum/__init__.py (plain "lib", not "lib64"). Luckily, patch(1) did "get" your git(1) patch, if it hadn't I'd been at a loss. Now go on and file a bug against bugzilla, as it isn't exactly easy to get the raw patch for testing...
Yep, fixed. Another machine showed the same problem with today's round of updates, after the patch the update goes through.
Survived today's round of updates too.
Thanks for testing :) The real test is in cases where there is a lot of updates and multiple packages with despolve issue, but that don't happen very often. The patch has been pushed upstream, I will close for now, please reopen if you hit any issues again