summaryrefslogtreecommitdiffstats
path: root/perl-install/install/pkgs.pm
Commit message (Collapse)AuthorAgeFilesLines
...
* (_cleanHeaders,extractHeaders) drop themThierry Vignaud2009-03-021-32/+0
|
* (_installTransactionClosure) drop itThierry Vignaud2009-03-021-80/+0
|
* (install) reuse urpmi's higher level code (drop librpm ordering hackThierry Vignaud2009-03-021-78/+55
| | | | and retry logic since urpmi should behaves better)
* (empty_packages) use urpm rather than URPM (thus enabling us to reuse urpmi'sThierry Vignaud2009-03-021-1/+2
| | | | | higher level code) (no changes for current callers since urpm inherits from URPM)
* (computeGroupSize) make it shorter through spliting out internal helper ↵Thierry Vignaud2009-02-201-41/+43
| | | | functions
* (rpmDbOpen) make it more readableThierry Vignaud2009-02-181-2/+8
|
* (_filter_packages) split it out of _installTransactionClosure()Thierry Vignaud2009-02-181-12/+17
|
* prefix internal functions with "_"Thierry Vignaud2009-02-181-31/+31
|
* - do not warn about "servers which open ports on outside by default"Pascal Rigaux2009-01-301-107/+0
| | | | | (the list of servers is not maintained, and the policy for those servers has changed)
* perl_checker cleanupsThierry Vignaud2009-01-261-2/+2
|
* (bestKernel_extensions) do not install kernel-server on x86_64 whenThierry Vignaud2009-01-091-1/+1
| | | | | there's more than 4Gb of RAM (#44709) ; use better heuristics
* cleanup: remove support from migrating to a different rpm dbapiPascal Rigaux2008-12-161-8/+2
| | | | | (it never worked and it was never used)
* - generate /var/lib/rpm/installed-through-deps.listPascal Rigaux2008-09-251-0/+1
| | | | | | | | (to initialize urpmi orphans) - for this, we must ensure $packages (which is similar to urpmi's $urpm) has field {root}
* allow to prefer kernel-server on x86_64 tooOlivier Blin2008-09-161-1/+1
|
* (bestKernel_extensions) do not uselessly install kernel-server onThierry Vignaud2008-08-201-1/+1
| | | | | x86_64 if more than 3.8Gb are availlable
* remove install::pkgs::read_rpmsrate proxy functionOlivier Blin2008-08-181-6/+0
|
* (packageCallbackChoices_) 'strict' pragma nicely show that a bug wasThierry Vignaud2008-08-131-1/+1
| | | | | | | | | introduced in r231074 (pixel, 2007-10-15) which was: "- in case of ftp/http media, we select basesystem before doing bestKernelPackage so we must ensure the choice callback selects the good kernel (reported by Thomas Spuhler on cooker mailing list)"
* - fix reading rpm macros from /etc/macros.d/ (esp. %_filetriggers_dir)Pascal Rigaux2008-06-261-1/+1
|
* kernel-laptop is no moreThierry Vignaud2008-05-041-1/+0
|
* - when computed install size in group selection (#39303)Pascal Rigaux2008-04-031-2/+2
| | | | | (it didn't take suggests into account)
* in the log, display reached size in MB, not GB through formatXiB (not ↵Pascal Rigaux2008-04-031-1/+1
| | | | precise enough)
* be a little less verbosePascal Rigaux2008-03-031-1/+0
|
* use fs::anyOlivier Blin2008-02-281-0/+1
|
* use newly moved fs::any::create_minimal_files (oops)Olivier Blin2008-02-281-1/+1
|
* add a progression log message (pkgs installation) as used to bePascal Rigaux2007-12-131-0/+1
|
* - do not rebuild rpmdb if it is ok Pascal Rigaux2007-12-111-2/+4
| | | | | (will workaround #32547) (need perl-URPM 3.01)
* - in case of ftp/http media, we select basesystem before doingPascal Rigaux2007-10-151-6/+15
| | | | | | bestKernelPackage so we must ensure the choice callback selects the good kernel (reported by Thomas Spuhler on cooker mailing list)
* - add /etc/rpm/macros.cdbPascal Rigaux2007-09-281-1/+1
|
* - translate the pkg summary when mdv-rpm-summary info is availablePascal Rigaux2007-09-211-0/+17
| | | | | | (for the "Details" mode when installing packages) - ensure mdv-rpm-summary is installed first (if selected)
* allow to specify keep_unrequested_dependencies in auto_instOlivier Blin2007-09-191-2/+4
|
* do not use $::o in install::pkgs (to be consistent with next commit and to ↵Olivier Blin2007-09-191-2/+2
| | | | please pixel)
* - when basesystem can't be selected, display the reasonPascal Rigaux2007-09-181-9/+18
|
* cleanup: use URPM::packages_providing()Pascal Rigaux2007-09-141-8/+2
|
* - packageCallbackChoices: use prefered packages given by perl-URPM 2.00 Pascal Rigaux2007-09-141-13/+6
| | | | | and perl-URPM now handles nicely locales-* as prefered packages
* - create resolve_requested_and_check() which nicely check selection failuresPascal Rigaux2007-09-141-17/+19
| | | | | - keep_unrequested_dependencies is now a global flag in perl-URPM 2.01
* selectPackage:Pascal Rigaux2007-09-141-4/+2
| | | | | | - stop reporting false errors - much nicer error message
* kernel-desktop586 is only on i.86Pascal Rigaux2007-08-311-1/+1
|
* kernel-latest is no morePascal Rigaux2007-08-311-1/+1
|
* (bestKernelPackage) fallback to kernel-desktopThierry Vignaud2007-08-271-1/+1
|
* drop globetrotter support (unused for quite a long time and superceeded by ↵Thierry Vignaud2007-08-241-1/+0
| | | | | | | Mdv One) we should just drop Mdv Move code btw...
* (bestKernelPackage) adapt to new kernel flavorsThierry Vignaud2007-08-241-4/+5
|
* - internalPascal Rigaux2007-08-121-9/+8
| | | | | | o use urpm::select::get_preferred to choose preferred packages (need urpmi library >= 4.10.1)
* - internalPascal Rigaux2007-08-121-3/+4
| | | | | | | | | o computeGroupSize: do direct requires before choices (eg for epiphany: require libmozilla-firefox2.0.0.6 before choosing package providing libmozjs.so) ps: is such a thing already done in perl-URPM's resolve_requested() ?
* computeGroupSize(): use packageCallbackChoices()Pascal Rigaux2007-08-121-19/+11
|
* perl-URPM callback gives us the virtual name, do display it instead of a ↵Pascal Rigaux2007-08-121-2/+2
| | | | mess of things
* fix typo (from 2003-02-25)Pascal Rigaux2007-08-121-1/+1
|
* rename {mediums} field into {media} to be compatible with urpm:: modules ↵Pascal Rigaux2007-08-121-5/+5
| | | | (from urpmi)
* - allow using sqlite for rpmdb when using option rpm_dbapi=4Pascal Rigaux2007-07-111-2/+8
|
* - bundle /etc/rpm/platform for rpm 4.4.8Pascal Rigaux2007-06-281-0/+6
|
* really install kernel-enterprise (not entreprise)Olivier Blin2007-05-231-1/+1
|