- Pacman-dev - lists.archlinux.org

[pacman-dev] test/sync999 - pacman segfault
by Aurelien Foret 23 Feb '06

23 Feb '06

Hi, Running pactest today showed that sync999 test case no longer works. Eventually, pactest return on investment does not seem null anymore :) It seems the flaw appeared with lib/libalpm/conflict.c:1.15. The same test with conflict.c:1.14 is ok. Here's the log from pactest: flow2: PM_OPT_DBPATH set to 'var/lib/pacman' flow2: PM_OPT_CACHEDIR set to 'var/cache/pacman/pkg' debug: opening database 'local' debug: opening database 'sync' :: Starting local database upgrade... flow1: checking for package replacements debug: loading package cache (infolevel=0x3) for repository 'sync' flow1: checking for package upgrades debug: loading package cache (infolevel=0x3) for repository 'local' debug: CACHE: fetching 'sync/pkg1' flow2: pkg1-1.0-1 elected for upgrade (1.0-1 => 1.0-2) debug: CACHE: fetching 'sync/pkg2' flow2: pkg2-1.0-1 elected for upgrade (1.0-1 => 1.0-2) resolving dependencies... flow1: resolving targets dependencies done. looking for inter-conflicts... flow1: looking for unresolvable dependencies debug: CACHE: fetching 'local/pkg1' debug: CACHE: fetching 'local/pkg2' flow1: looking for conflicts debug: targs vs db: found pkg2 as a conflict for pkg1 debug: targs vs targs: found pkg2 as a conflict for pkg1 debug: db vs targs: found pkg1 as a conflict for pkg2 flow2: package 'pkg1' is conflicting with 'pkg2' debug: CACHE: fetching 'local/pkg2' debug: package 'pkg1' provides its own conflict debug: resolving package 'pkg1' conflict flow2: electing 'pkg2' for removal flow2: removing 'pkg2' from target list flow2: package 'pkg2' is conflicting with 'pkg1' debug: CACHE: fetching 'local/pkg1' When handling conflict pkg1, pkg2 is removed from targets. Then, the library tries to handle pkg2 conflict... but find_pkginsync() returns NULL... I've committed a fix (sync.c:1.66) to skip conflict resolution in such a case. Thoughts? The test suite is ok with it ;) -- Aurelien

2 4

[pacman-dev] Sorry, test mail for me :)
by Christian Hamar 23 Feb '06

23 Feb '06

Hey. Just a test email, sorry from list members, i need to figure out why i got bounced on this list :( Regards -krix-

1 0

[pacman-dev] makepkg: removing empty directories?
by VMiklos 21 Feb '06

21 Feb '06

from makepkg: 1.12 (aurelien 16-Feb-06): # remove empty directories 1.12 (aurelien 16-Feb-06): if [ "`check_option NOEMPTYDIRS`" ]; then 1.12 (aurelien 16-Feb-06): msg "Removing empty directories..." 1.12 (aurelien 16-Feb-06): find pkg -mindepth 1 -type d -empty -exec rmdir {} \; 1.12 (aurelien 16-Feb-06): fi cmiiw, but openssh creates an empty /var/empty directory and it _does_ need it. any idea how to solve this problem? for now, i haven't merged this hunk to our tree udv / greetings, VMiklos -- Developer of Frugalware Linux, to make things frugal - http://frugalware.org

2 2

[pacman-dev] Fwd: [ Re: Patch queue ]
by VMiklos 21 Feb '06

21 Feb '06

----- Forwarded message from Christian Hamar <krics(a)linuxforum.hu> ----- > Date: Mon, 20 Feb 2006 01:27:19 +0100 > From: Christian Hamar <krics(a)linuxforum.hu> > To: VMiklos <vmiklos(a)frugalware.org> > Subject: Re: [pacman-dev] Patch queue > X-Mailer: Evolution 2.4.2.1 Frugalware Linux Distribution > > 2006. 02. 18, szombat keltezéssel 16.38-kor VMiklos ezt írta: > > On Fri, Feb 17, 2006 at 11:37:46PM +0100, Aurelien Foret <aurelien(a)archlinux.org> wrote: > > > Hi all, > > > > > > Following today's discussion on the ML, here is the list of pending > > > patches, with an updated status. > > > If one got lost during the struggle, thanks to let me know. > > > > > > > > > [6] doxygen: support for autotools > > > http://www.archlinux.org/pipermail/pacman-dev/2006-January/000170.html > > > > > > Commited. But I had to comment the part in the file doc/Makefile.am for now. > > > It generates the following error: > > > doc/Makefile.am:22: man_MANS must be set with `=' before using `+=' > > Hm. I don't understand why this problem occurs at you. If you view the > Makefile.am then you can see that there is a man_MANS = (without +) > > That man_MANS "runs" everytime. So there is no way to just only man_MANS > += "run" because the man_MANS = isn't in any ifdef define. > > Ps.: Why i'm banned from the pacman list ?? I saw that > krics(a)linuxforum.hu too many bounced messages ... But my email works > fine. Any idea ?! > > Regards > -krix- ----- End forwarded message ----- udv / greetings, VMiklos -- Developer of Frugalware Linux, to make things frugal - http://frugalware.org

2 4

[pacman-dev] Little question regarding alpm
by Joao Estevao 20 Feb '06

20 Feb '06

I'm currently trying to write a front end in python and a python wrapper to alpm and I'm running into some problems. I can't seem to be able to install a package twice. All goes well the first time, but then I get this: DEBUG: closing database 'extra' DEBUG: freeing package cache for repository 'extra' Then I tought I might have to re-open the 'extra' db, so I close it and then I re-open it and get this: FLOW2: target 'alsaplayer' not found -- looking for provisions or sometimes I get : RuntimeError: transaction not initialized or sometimes it works. I guess my question is this. Why does alpm close the db and free the pkg cache when one installs some package(s)? Couldn't it refresh the db and the pkg cache? Thanks in advance and please correct me if I'm wrong. Joao -- A mind all logic is like a knife all blade. It makes the hand bleed that uses it. -- Rabindranath Tagore (1861-1941)

2 2

[pacman-dev] Patch queue
by Aurelien Foret 20 Feb '06

20 Feb '06

Hi all, Following today's discussion on the ML, here is the list of pending patches, with an updated status. If one got lost during the struggle, thanks to let me know. [1] libalpm: symbols naming Just commited. [2] dependsonly feature http://www.archlinux.org/pipermail/pacman-dev/2005-December/000039.html Rejected for now. It will be handled as a feature request (see bug #3220) Add the patch as an attachment to the bugtracker? [3] sync_commit: pass callbacks to the internal transaction http://www.archlinux.org/pipermail/pacman-dev/2006-January/000157.html Commited. I've extended it so that logs and events from the removal part of a sync transaction got notified too. [4] sync_prepare: re-order w.r.t. dependencies http://www.archlinux.org/pipermail/pacman-dev/2005-December/000074.html To be reviewed. [5] sortbydeps: handle provisions http://www.archlinux.org/pipermail/pacman-dev/2006-January/000169.html To be reviewed. [6] doxygen: support for autotools http://www.archlinux.org/pipermail/pacman-dev/2006-January/000170.html Commited. But I had to comment the part in the file doc/Makefile.am for now. It generates the following error: doc/Makefile.am:22: man_MANS must be set with `=' before using `+=' [7] makepkg: partial support for version comparators Commited. -- Aurelien

1 1

[pacman-dev] [PATCH] pacman manpage fixes
by VMiklos 20 Feb '06

20 Feb '06

hi http://frugalware.org/~vmiklos/patches/libpacman-proposed/pacman-manfix.diff somehow these hunks were reverted during the pacman 2.9.8 merge udv / greetings, VMiklos -- Developer of Frugalware Linux, to make things frugal - http://frugalware.org

2 1

[pacman-dev] test suite pactest 0.2 ready
by Aurelien Foret 20 Feb '06

20 Feb '06

I've updated the test suite with more features and test cases. It is available here: http://aurelien.foret.free.fr/archlinux/pactest-0.2.tar.gz It should be almost usable as is by frugalware guys (just update the PM_EXT_PKG variable in the util.py file?) -- Aurelien

2 3

[pacman-dev] [PATCH] makepkg: partial handling of version comparators
by VMiklos 09 Feb '06

09 Feb '06

hi http://darcs.frugalware.org/darcsweb/darcsweb.cgi?r=pacman;a=plain_commitdi…; the old situation: if for example a module depended on kernel=2.6.15-1 and pacman -S installed 2.6.14-1, then this problem was silently ignored the new situation: now the wrong version (ie. kernel=2.6.14-1) is still installed, but at least after that makepkg exists with an error udv / greetings, VMiklos -- Developer of Frugalware Linux, to make things frugal - http://frugalware.org

1 0

[pacman-dev] download code in pacman
by VMiklos 05 Feb '06

05 Feb '06

hi is there any reason why the download code is in pacman and not in the library? yes, i know, this way the frontends are not forced to use libftp, for example they can use curl, if you want. but i think this is not the same situation as the list handling functions / GList (ie. at a gtk frontend) imho placing the old download code in pacman, and not in the library is a disadvantage, and not an advantage or we could allow to use NULL as the download cb function pointer, and then use the our own download code opinions? if i would implement this then theoratically you wouldn't have objections? udv / greetings, VMiklos -- Developer of Frugalware Linux, to make things frugal - http://frugalware.org

4 11