summaryrefslogtreecommitdiffstats
path: root/perl-install/pkgs.pm
Commit message (Collapse)AuthorAgeFilesLines
* kill some "used undefined variable" warningsPascal Rigaux2007-09-141-1/+1
|
* re-sync after the big svn lossPascal Rigaux2007-05-181-1486/+31
|
* re-sync after the big svn lossPascal Rigaux2007-05-181-0/+1598
|
* re-sync after the big svn lossPascal Rigaux2007-04-251-1598/+0
|
* fix selected_leaves() dropping all packages in a circular referencePascal Rigaux2006-01-241-6/+11
| | | | | (including short circular references like amarok-engine-arts -> libamarokarts.so -> amarok-engine-arts)
* package lsb is requiring many locales though we don't want it to be installedPascal Rigaux2005-12-131-2/+9
| | | | | | only if those many locales are chosen. So discarding those locales requires (bugzilla #20183)
* (bestKernelPackage) install i586 kernel flavor for live systemsThierry Vignaud2005-12-071-0/+1
|
* fix rpmsrate negations when not installing a live systemOlivier Blin2005-11-211-4/+4
|
* (read_rpmsrate) do not match CAT_(KDE|GNOME|...) and ignore ! whileThierry Vignaud2005-11-211-5/+4
| | | | | installing a live system
* (read_rpmsrate) let "build_live_system" mode be selected by auto config fileThierry Vignaud2005-11-211-1/+1
|
* (read_rpmsrate) install all hardware related packages when building a live ↵Thierry Vignaud2005-11-211-0/+1
| | | | system
* Fix test for dangling symlinkRafael Garcia-Suarez2005-10-201-1/+1
|
* Comment fixRafael Garcia-Suarez2005-09-261-1/+1
|
* don't warn lisa (which is installed by default)Pascal Rigaux2005-09-201-1/+4
|
* URPM/Resolve.pm diff 1.109: "Remove the return value of ↵Pascal Rigaux2005-09-161-1/+9
| | | | | | | compute_installed_flags"... but alas install still use it, so doing here what was done in compute_installed_flags
* - call installCallback() on "open" tooPascal Rigaux2005-09-151-0/+2
| | | | | | | - on "open" we are not chrooted - setting advertising when chrooted could cause havoc if pango wants to load a new font (eg: advertising #13 in pt_BR)
* add upgrade script for conectivaPascal Rigaux2005-09-141-10/+21
|
* do have all useful info in selected leaves, including base packages,Pascal Rigaux2005-09-141-1/+1
| | | | | it won't be much longer but we can precise choices like lilo vs grub
* do log the prefered choice (esp. to debug lilo vs grub)Pascal Rigaux2005-09-141-1/+1
|
* have same rpm config as when installing pkgsPascal Rigaux2005-09-091-1/+4
|
* At some point, an empty hashref is autovivified in $packages->{medium}.Rafael Garcia-Suarez2005-09-051-1/+2
| | | | | | So, protect the loops that iterate over this hash, in application of the belt and suspenders doctrine.
* - new functionality: upgrade_by_removing_pkgs, enabled when upgrading redhat ↵Pascal Rigaux2005-09-021-0/+81
| | | | | | | | | | | and conectiva distributions - add file upgrade-map.conectiva.10 for precise choice of packages - save /etc/xxx-release into /root/drakx/xxx-release.upgrading when starting - release_file(): look for xxx-release.upgrading first - find_root_parts(): better logging about upgrade_by_removing_pkgs, and factorize code - when all packages are installed, remove xxx-release.upgrading and rename pkgs::removed_pkgs_to_upgrade_file()
* Workaround : Don't call method directlyRafael Garcia-Suarez2005-09-021-1/+1
|
* rename pkgs::rpmDbOpenForInstall() to pkgs::open_rpm_db_rw()Pascal Rigaux2005-09-021-3/+3
|
* allow verbose removing of packagesPascal Rigaux2005-09-021-1/+1
|
* opening rpm db in selectPackage(), so remove some rpmDbOpen()Pascal Rigaux2005-09-021-0/+2
|
* new function select_by_package_names()Pascal Rigaux2005-09-021-0/+9
|
* create remove_raw(), remove() now retries with option noscriptsPascal Rigaux2005-09-011-4/+14
|
* preferred packages:Pascal Rigaux2005-09-011-1/+1
| | | | | | - remove packages not existing anymore - add nail and glibc-devel
* help debugging packageCallbackChoices()Pascal Rigaux2005-09-011-2/+3
|
* instead of dirtying pkgs::installCallback, use install_steps::installCallbackPascal Rigaux2005-08-311-9/+9
| | | | | (still not clean, but better)
* fix typoPascal Rigaux2005-08-311-1/+1
|
* create remove_marked_ask_remove() and use itPascal Rigaux2005-08-311-26/+21
|
* on 2002/07/10 (1.347), selected_leaves() behaviour was broken, listing all ↵Pascal Rigaux2005-08-291-4/+11
| | | | | | | packages. restoring it (bugzilla #18000)
* remove now unused variable (cf previous commit)Pascal Rigaux2005-08-251-1/+0
|
* don't kill "runaway" processes anymore, it should not be needed for ejecting ↵Pascal Rigaux2005-08-251-14/+0
| | | | cd (?)
* Strange typing bug workaroundRafael Garcia-Suarez2005-08-251-1/+2
|
* for bestKernelPackage(), sort kernels to have higher version firstPascal Rigaux2005-08-231-1/+3
|
* update naughtyServers for new distroPascal Rigaux2005-08-181-6/+15
|
* (naughtyServers_list): those packages don't exist anymorePascal Rigaux2005-08-181-8/+0
|
* don't fork anymore to install rpmsPascal Rigaux2005-08-171-71/+25
|
* during install, use "nofsync" for rpm database (=> speedup x2)Pascal Rigaux2005-08-171-0/+1
|
* cleanupPascal Rigaux2005-08-171-8/+8
|
* on upgrade, have not only the upgraded packages, but also the installed ↵Pascal Rigaux2005-08-101-1/+1
| | | | packages in package_list.pl (bugzilla #15296)
* move is_xbox from common to detect_devicesOlivier Blin2005-08-051-1/+1
|
* This function was renamedRafael Garcia-Suarez2005-06-301-1/+1
|
* Remove dependency of c::stuff on rpmlibRafael Garcia-Suarez2005-06-301-7/+6
|
* try to cleanup fs.pmPascal Rigaux2005-06-281-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | (to have simpler dependencies between modules, esp. have some modules only required by diskdrake): - move some functions from fs to fs::mount (most keep their name, except mount_part and mount_usbfs) - move formatMount_part and formatMount_all from fs to fs::format - move some functions from fs to fs::wild_device (part2wild_device_name -> fs::wild_device::from_part) (subpart_from_wild_device_name -> fs::wild_device::to_subpart) (analyze_wild_device_name -> fs::wild_device::analyse) - formatMount_part(), formatMount_all(), fs::mount::part() don't take a prefix anymore the current situation was quite muddy we now rely on fs::get::mntpoint_prefixed() which will maybe depend on a field in $part for now, we mount every part in chroot, it seems to be what's wanted - fs::format::part() now expect $all_hds instead of $raids - fs::type::carryRootLoopback is now fs::get::carry_root_loopback() - in fs::loopback, most functions don't want a prefix anymore
* Fix install_urpmi with nfs suppl mediaRafael Garcia-Suarez2005-06-281-1/+2
|
* Remove DB logs before doing an install.Rafael Garcia-Suarez2005-06-241-0/+6
|