- Arch-general - lists.archlinux.org

[arch-general] [PATCH 1/4] Add skeleton for hooks support
by Victor Lowther 23 Jul '10

23 Jul '10

This begins work on adding some sort of hook infrastructure to pacman. There has been much discussion and little code that I have seen, so I figured I would go ahead an implement a very basic skeleton for transaction hooks support. --- lib/libalpm/Makefile.am | 1 + lib/libalpm/{sync.h => hooks.c} | 34 +++++++++++++++++++++++++--------- lib/libalpm/{add.h => hooks.h} | 19 ++++++++----------- 3 files changed, 34 insertions(+), 20 deletions(-) diff --git a/lib/libalpm/Makefile.am b/lib/libalpm/Makefile.am index 3473a73..b4fdadf 100644 --- a/lib/libalpm/Makefile.am +++ b/lib/libalpm/Makefile.am @@ -44,6 +44,7 @@ libalpm_la_SOURCES = \ sync.h sync.c \ trans.h trans.c \ util.h util.c \ + hooks.h hooks.c \ version.c libalpm_la_LDFLAGS = -no-undefined -version-info $(LIB_VERSION_INFO) diff --git a/lib/libalpm/sync.h b/lib/libalpm/hooks.c similarity index 65% copy from lib/libalpm/sync.h copy to lib/libalpm/hooks.c index 000a09c..9630c70 100644 --- a/lib/libalpm/sync.h +++ b/lib/libalpm/hooks.c @@ -1,9 +1,10 @@ /* - * sync.h + * hooks.c * * Copyright (c) 2006-2010 Pacman Development Team <pacman-dev(a)archlinux.org> * Copyright (c) 2002-2006 by Judd Vinet <jvinet(a)zeroflux.org> * Copyright (c) 2005 by Aurelien Foret <orelien(a)chez.com> + * Copyright (c) 2005 by Christian Hamar <krics(a)linuxforum.hu> * Copyright (c) 2005, 2006 by Miklos Vajna <vmiklos(a)frugalware.org> * * This program is free software; you can redistribute it and/or modify @@ -19,14 +20,29 @@ * You should have received a copy of the GNU General Public License * along with this program. If not, see <http://www.gnu.org/licenses/>. */ -#ifndef _ALPM_SYNC_H -#define _ALPM_SYNC_H -#include "alpm.h" - -int _alpm_sync_prepare(pmtrans_t *trans, pmdb_t *db_local, alpm_list_t *dbs_sync, alpm_list_t **data); -int _alpm_sync_commit(pmtrans_t *trans, pmdb_t *db_local, alpm_list_t **data); +#include "config.h" -#endif /* _ALPM_SYNC_H */ +#include <stdlib.h> +#include <stdio.h> +#include <string.h> +#include <unistd.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <sys/statvfs.h> +#include <errno.h> -/* vim: set ts=2 sw=2 noet: */ +/* libalpm */ +#include "hooks.h" +#include "trans.h" +#include "alpm_list.h" +#include "package.h" +#include "util.h" +#include "log.h" +#include "handle.h" +#include "add.h" +#include "remove.h" +#include "sync.h" +#include "alpm.h" +#include "deps.h" +#include "cache.h" diff --git a/lib/libalpm/add.h b/lib/libalpm/hooks.h similarity index 73% copy from lib/libalpm/add.h copy to lib/libalpm/hooks.h index eb37dc7..d60b1f4 100644 --- a/lib/libalpm/add.h +++ b/lib/libalpm/hooks.h @@ -1,8 +1,11 @@ /* - * add.h + * hooks.h * * Copyright (c) 2006-2010 Pacman Development Team <pacman-dev(a)archlinux.org> * Copyright (c) 2002-2006 by Judd Vinet <jvinet(a)zeroflux.org> + * Copyright (c) 2005 by Aurelien Foret <orelien(a)chez.com> + * Copyright (c) 2005 by Christian Hamar <krics(a)linuxforum.hu> + * Copyright (c) 2006 by Miklos Vajna <vmiklos(a)frugalware.org> * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by @@ -17,15 +20,9 @@ * You should have received a copy of the GNU General Public License * along with this program. If not, see <http://www.gnu.org/licenses/>. */ -#ifndef _ALPM_ADD_H -#define _ALPM_ADD_H +#ifndef _ALPM_HOOKS_H +#define _ALPM_HOOKS_H -#include "db.h" -#include "alpm_list.h" -#include "trans.h" +#include "alpm.h" -int _alpm_upgrade_packages(pmtrans_t *trans, pmdb_t *db); - -#endif /* _ALPM_ADD_H */ - -/* vim: set ts=2 sw=2 noet: */ +#endif /*ALPM_HOOKS_H */ -- 1.7.1.1

2 6

[arch-general] A question about using tar... ( -p & --numeric-owner)
by Joe(theWordy)Philbrook 23 Jul '10

23 Jul '10

I'm a multi-linux-distro/multi-boot guy. I always have more than one configured linux distro on my grub menu. Whenever I'm considering radical changes to one of my working linux I like to back it up. Quite some time ago I learned to do this with tar. For example I've got both Arch Linux and Xubuntu on this desktop pc. I've decided to attempt using a network upgrade process to upgrade the Xubuntu from Karmic to the current Lucid release. So first I boot Arch, mount a storage drive on /mnt/other, & mount the Xubuntu partition on /mnt/xubuntu. Then cd to /mnt/xubuntu. (with a root shell of course) # tar -czf /mnt/other/xubuntu.tgz -p --numeric-owner . And of course if it's necessary I can restore the backup to an empty partition with: # tar -xzf /mnt/other/xubuntu.tgz -p --numeric-owner . But I only partially understand the -p and --numeric-owner options. I do know that -p is supposed to be default for root but I like to make sure... I think I understand that the reason somebody told me (quite some time ago) to use --numeric-owner is because (in this example) Arch may have different user id numbers than Xubuntu which could have bad results if it tried to apply permissions based on user names (especially if it involves a "system account")... What I don't know is if either of these options are necessary during the "tar -czf" operation, or if I only need to bother with them when/if I restore a file system with "tar -xzf" ??? I know it does work the way I expect it to when I use the options for both the "-czf" & the "-xzf" operations. And that I don't get any error messages from tar when I do it that way. But I don't know if using them for the "-czf" operation makes any difference or not. Would someone be so kind as to set me straight on this??? -- | ~^~ ~^~ | <?> <?> Joe (theWordy) Philbrook | ^ J(tWdy)P | \___/ <<jtwdyp(a)ttlc.net>>

1 0

[arch-general] Clarfication request regarding mkinitcpio code
by Dave Reisner 22 Jul '10

22 Jul '10

There's a few things in mkinitcpio I'd like to cleanup, but I need to have one issue clarified before I proceed. The header of mkinitcpio itself states that various parts need to run under Busybox's ash, and thus various standards need to be adhered to. However, the file itself sports a /bin/bash shebang, and makes mild use of parameter expansion, e.g. APPNAME=${0##*/} on line 44. I understand that hooks themselves need to run in a more POSIX environment, but would it be safe to use more Bash syntax in the code that generates the initcpio image? I would be sticking to a feature set compatible with 3.2. thanks, dave reisner

1 0

[arch-general] Hard lockups during boot
by Chris Bannister 22 Jul '10

22 Jul '10

Hi, I used to use a zen kernel but that stopped working which is an unrelated story so I switched back to the Arch kernel, but im getting hardlocks during boot. It happened twice just now, first time was after ATA got init'd but before the SCSI came up. The second time it got to the waiting for udev .. during userspace init. Heres the attached working dmesg log [1] in case its of anyhelp. I didnt want to post this to the bug tracker because of the lack of logs I can provide so ill put it here and if anyone has any ideas please shoot them over. Thanks! [1] http://zariel.pastey.net/138823

1 0

[arch-general] simple iptabels question
by aerospace1028＠hotmail.com 22 Jul '10

22 Jul '10

Hello, I have a question on iptables. On the arch wiki, the tutorial on a simple stateful firewall includes the creation of the following rule. # iptables -A INPUT -p tcp --syn -m state --state NEW -j OPEN-TCP Which, to my understanding, meant only pass new TCP streams with the syn flag (the initialize connection or handshake part 1 of 3) to the user defined chain. When I checked the output of iptables, I noticed the following rule. # iptables -L INPUT --line-numbers -v 7 OPEN-TCP tcp -- anywhere anywhere tcp flags:FIN,SYN,RST,ACK/SYN state NEW Why is iptables also using the FIN, RST and ACK/SYN flags? Did I not request SYN only like I thought the tutorial said? _________________________________________________________________ The New Busy is not the old busy. Search, chat and e-mail from your inbox. http://www.windowslive.com/campaign/thenewbusy?ocid=PID28326::T:WLMTAGL:ON:…

2 1

[arch-general] Archlinux Wiki Backup
by Francesco Turco 22 Jul '10

22 Jul '10

Hello, As I said in the forums (https://bbs.archlinux.org/viewtopic.php?id=94201) I started contributing to the wiki (minor edits for now), but I'm concerned about the absence of a proper way to obtain a snapshot of it for backup reasons. I looked at the arch-wiki-docs package, but it's only a collection of HTML pages, with no source code, no discussions, no revisions, and not working hyperlinks. This way, it would be hard to restore the wiki after a crash. Other wikis provide a backup: please see http://download.wikimedia.org/ and http://en.gentoo-wiki.com/wiki/Gentoo_Linux_Wiki:Backup. To reduce bandwidth usage, we can use Bittorrent, just as with ISO releases. What do you think? Thanks.

1 0

[arch-general] All kdemod3 menu apps in "Lost & Found" menu folder + control panel apps gone - how to get back?
by David C. Rankin 21 Jul '10

21 Jul '10

Guys, Help! Something (probably that fricking kde4) has caused almost every menu item to end up in the "Lost & Found" kde3 menu folder off the main menu. I don't know where that folder is so I can try an move the items back. Also, what ever happened also caused all kde3 control panel entries to disappear from the control panel. Now I don't have a control panel any more: 23:22 alchemy:/opt/kde/share/applications/kde> kcmshell --list The following modules are available: Huh? The only thing I have done menu wise is to edit the Gnome menu, but that shouldn't have trashed my kde3 menu and control panel. Can anyone offer any ideas where all the missing apps are. Where is this "Lost & Found" place all the menu items went? Here is a screenshot of what I see when I move over the "Lost & Found" menu item. (131k) http://www.3111skyline.com/dl/Archlinux/bugs/kde3-lost+found-menu.jpg It is frustrating as hell to have everything setup just the way you like it and then have some unrelated process fsck it all up. I've been in gnome for the past week or so and this is the first time I've booted back to kde3. Go figure. Any ideas? I have had the disappearing menus and control panel issue before in kde3 on SuSE, but never in Arch - now Arch has been bitten by this bug. Hopefully we can find the cause. Surely alacarte is to blame... Is it? -- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com

2 8

[arch-general] $BROWSER variable and Google Earth
by F. Gr. 21 Jul '10

21 Jul '10

Hi, Google Earth can't recognise the $BROWSER environment variable. I've set the one in my '~/.zshrc' as follows: " export BROWSER="/usr/bin/icecat" " When I try to open an external link I get the following error: <http://www.imagebam.com/image/a79e8989327021>. That Italian message means: "cannot start up a web browser. Check to have set the $BROWSER environment variable to the web browser filename you want to start up." I don't know why. Cheers

6 8

[arch-general] Archweb package update script
by Dan McGee 21 Jul '10

21 Jul '10

On Wed, Jul 21, 2010 at 5:45 AM, Netanel Shine <netanelshine(a)gmail.com> wrote: >> >> Date: Sun, 18 Jul 2010 21:36:42 -0500 >> From: Dan McGee <dpmcgee(a)gmail.com> >> Subject: Re: [arch-general] RSS feed - missing feature|problem >> On Mon, Jul 12, 2010 at 3:59 AM, Netanel Shine <netanelshine(a)gmail.com> >> wrote: >> > Hey all, >> > >> > My name is Netanel Shine, and a few days ago i began the process of >> building >> > an israeli community of archlinux. I used a git clone to set up the main >> > site. (like some other great projects file that archlinux git have to >> offer >> > like the wiki and the forums) >> > Everything went smoothly and just works great, except for one problem - >> the >> > package rss feeds. >> > >> > The site address is: http://archlinux.org.il - under heavy development. >> > >> > The problem: As you all can see, the left rss box with the latest updates >> > dont work and the list is empty. - because the system comes with an >> > integrated "packages" part (you can see it here: >> > http://archlinux.org.il/packages - not gonna use it), so the rss list >> > updates from here (i guess) >> > >> > After discussions with some great people (in the irc) we concluded that >> the >> > rss box pull`s the data from the 'models.Package' object - that connected >> to >> > archlinux mainsite DB (Package.objects=PackageManager()), so if i dont >> have >> > any access for the database - thats not gonna work. >> > >> > All sorts of ideas were came up like bulding a script that will do the >> job >> > or, a an iframe that will copy the box from the mainsite (but the problem >> it >> > will not be in hebrew) and an idea to replace the table that has for >> update >> > in pkg_updates. (with logic that pulls from the feed) - but thats not >> easy >> > because the logic is a wrapped up in django MVC. >> > >> > So dear friends, i turn to you for help - and I`m sure I would find it. >> >> Check out the reporead command baked into manage.py- you can very >> easily populate the package DB yourself. >> >> I have a small script I can send you for this, but I don't have it >> handy at the moment. Let me know if that would be useful to you. >> >> -Dan > > First of all, I sorry for seeing this message just now. > Thanks Dan, I appreciate your comment, and i would like to have the script > you're talking about , please send him to: netanelshine at gmail dot com ;) Something like the following: $ cat projects/archweb/fetch_update_package_db.sh #!/bin/bash -e # run at nice 3. it can churn quite a bit of cpu after all. renice +3 -p $$ > /dev/null # setup paths DLURL="ftp://ftp.archlinux.org" SPATH="/path/to/archweb" SCRIPT="${SPATH}/scripts/reporead.py" TMPPATH="/tmp/updaterepos" ARCHES="i686 x86_64" REPOS="core extra testing community community-testing" startdir="$(pwd)" [ ! -d ${TMPPATH} ] && mkdir ${TMPPATH} for arch in ${ARCHES}; do [ ! -d ${TMPPATH}/${arch} ] && mkdir ${TMPPATH}/${arch} for repo in ${REPOS}; do #[ ! -d ${TMPPATH}/${repo} ] && mkdir ${TMPPATH}/${repo} #cd ${TMPPATH}/${repo} echo "Updating ${repo}-${arch}" cd $TMPPATH/${arch} wget -N ${DLURL}/${repo}/os/${arch}/${repo}.db.tar.gz #${SCRIPT} ${arch} ${TMPPATH}/${arch}/${repo}.db.tar.gz cd "$startdir" ./manage.py reporead ${arch} ${TMPPATH}/${arch}/${repo}.db.tar.gz echo "" done done

2 2

Re: [arch-general] arch-general Digest, Vol 69, Issue 37
by Netanel Shine 21 Jul '10

21 Jul '10

> > Date: Sun, 18 Jul 2010 21:36:42 -0500 > From: Dan McGee <dpmcgee(a)gmail.com> > Subject: Re: [arch-general] RSS feed - missing feature|problem > On Mon, Jul 12, 2010 at 3:59 AM, Netanel Shine <netanelshine(a)gmail.com> > wrote: > > Hey all, > > > > My name is Netanel Shine, and a few days ago i began the process of > building > > an israeli community of archlinux. I used a git clone to set up the main > > site. (like some other great projects file that archlinux git have to > offer > > like the wiki and the forums) > > Everything went smoothly and just works great, except for one problem - > the > > package rss feeds. > > > > The site address is: http://archlinux.org.il - under heavy development. > > > > The problem: As you all can see, the left rss box with the latest updates > > dont work and the list is empty. - because the system comes with an > > integrated "packages" part (you can see it here: > > http://archlinux.org.il/packages - not gonna use it), so the rss list > > updates from here (i guess) > > > > After discussions with some great people (in the irc) we concluded that > the > > rss box pull`s the data from the 'models.Package' object - that connected > to > > archlinux mainsite DB (Package.objects=PackageManager()), so if i dont > have > > any access for the database - thats not gonna work. > > > > All sorts of ideas were came up like bulding a script that will do the > job > > or, a an iframe that will copy the box from the mainsite (but the problem > it > > will not be in hebrew) and an idea to replace the table that has for > update > > in pkg_updates. (with logic that pulls from the feed) - but thats not > easy > > because the logic is a wrapped up in django MVC. > > > > So dear friends, i turn to you for help - and I`m sure I would find it. > > Check out the reporead command baked into manage.py- you can very > easily populate the package DB yourself. > > I have a small script I can send you for this, but I don't have it > handy at the moment. Let me know if that would be useful to you. > > -Dan First of all, I sorry for seeing this message just now. Thanks Dan, I appreciate your comment, and i would like to have the script you're talking about , please send him to: netanelshine at gmail dot com ;) -- Netanel Shine

1 0