Arch gaming repositories

It seems that I'm getting errors with the two repositories listed in the stickied thread at the top of this subforum, as well as Twilight Lair. Anyone know what's going on with these servers?
# clyde -Syy
:: Synchronizing package databases...
core 36.0K 160.1K/s 00:00:00 [----------------------] 100%
extra 450.9K 601.9K/s 00:00:01 [----------------------] 100%
community 364.7K 567.1K/s 00:00:01 [----------------------] 100%
error: failed retrieving file 'arch-games.db.tar.gz' from arch.twilightlair.net : Connection refused
arch-games 33.9K 68.4K/s 00:00:00 [----------------------] 100%
xyne-any 5.7K 5.3K/s 00:00:01 [----------------------] 100%
And as for the two listed in the sticky, they just time out.

The main server is down at the moment, use a mirror.
Edit: Back up
Last edited by Daenyth (2010-04-06 04:27:15)

Similar Messages

Arch gaming

I'm interested in the idea of maybe having a sticky thread in this forum for all the Linux gamers out there. Specifically older Loki games such as Myth2 and Alpha Centauri. One reason, being a avid Linux gamer. I find it difficult to Google for information regarding installation of many older native games. For instance
Myth 2: Soulblighter
Install the game with the option FULL.
Myth2 requires the the latest unified patch:
http://lokifiles.tuxgames.com/updates/m … ed-x86.run
Myth2 has a couple of bugs regarding video playback. The only fix I could find was to move or delete the cutscenes folder, which is in the game directory. ie /usr/local/games/myth2/
There is also a bug that stops OpenGL from working. The workaround is here:
http://tobias.rautenkranz.ch/myth_memfix.html
It would be good if all the last problem could be automated. Ie, add a script like this
echo 'LD_PRELOAD=/usr/lib/fopenr.so myth2' > GL_Myth2 ; chmod a+x GL_Myth2
into say the /usr/bin folder?
This would be an example of a wiki entry for arch linux to get the game working?
BTW: this is just an idea / quick example of something I would find useful. So we can have all the native games up on one big wiki page for arch.

Oh and the sticky thread would be like a FAQ with wiki links etc?
http://wiki.archlinux.org/index.php/Games
This is the only wiki info i could find.
Last edited by B15HOP (2008-02-07 08:34:35)

Bluetooth USB Dongle works in Ubuntu but not Arch

My "Conceptronic Bluetooth 2.1 USB Nano adapter (CBT2NANO)" works in ubuntu correctly but when I try to use it in Archlinux the kernel says:
hub 6-0:1.0: unable to enumerate USB device on port 1
and the logs are flooded with this message.
it think that the problem is a bluez patch not included in the arch linux repositories.
(My first Lang is Spanish, so I can be wrong in the grammar, sorry)
I hope that someone know where is the problem and help me to fix it.

there is a problem with the kernel!, Actually I am using Ubuntu 11.10 and I keep having the same issue, but when I connect my dongle in the USB 3.0 port, it works and is recognized as a Full Speed USB Device:
usb 3-3: new full speed USB device number 37 using ohci_hcd
When I use a USB extension (Cable) or connect the device in USB 2.0 Port, dmesg says:
hub 3-0:1.0: unable to enumerate USB device on port 3
once, again, again, and so on.
And I think the problem is in ohci_hcd module.
If at first the dongle works in USB 2.0 Port I want it working in this Port too, couse some PC have not USB 3.0 Port, and my USB3.0 is in the back panel, so the signal is worse.
thanks for any help.
Last edited by jaao_deat (2011-12-22 11:10:37)

"Multiseat" gaming with multi-pointer X

Hello Arch gaming community,
recently I've discovered a decent way to play multiplayer games in multiseat (or pseudo split-screen) manner. It really works - I use it to play Myth II: Soulblighter (on Wine) with my brother
Short explanation:
In recent versions of Xorg (>= 1.7), there's a nice feature - it lets you plug another mouse and make it control a second pointer on the screen. Plug another keyboard, assign it to this pointer, and you can independently work with two windows at a time (each pointer getting it's own focus). Now, if you connect a second physical display and configure it to be a separate X screen belonging to the same X server, you can send your second pointer to that screen, virtually allowing both sets to be used independently. Now, run some game on first screen, then another instance on another, and you can play it multiplayer with your friend
What makes this approach better than a "true" multiseat:
Configuring a proper multiseat system (where each "seat" get's it's own login screen) is not easy, to start with. Often it won't let you use hardware acceleration on both displays unless they're connected to two separate video cards. Mutli-pointer approach lets you use both screens with hardware acceleration with just one dual-head graphics card (like most laptops have). Besides, it's much less painful to set up.
The downsides:
The multi-pointer feature of Xorg is still fairly new, and is not supported by any window manager that I know. Thankfully, it doesn't have to be explicitly supported by a WM in order to be used with it, but most WMs tend to get confused by it. Also, sometimes keyboard focus is not automatically assigned for the second pointer, and you have to explicitly assign it via command-line for the keyobard input to work.
So, how to do it?
The basic steps are:
Prepare an alternative xorg configuration file for both displays (if you don't use both already)
Start another X server with "xinit /usr/bin/xterm -- :1 vt8 -config [your dual screen xorg config file]" (so that when something goes bad, your main desktop won't suffer)
Set up a secondary pointer with xinput utility
Launch two instances of some game, one on each screen
Use xinput with "set-cp" option to assign a second pointer to the window that will be operated by it
Play!
You'll have to repeat all steps (except the first one) each time you want to play in this configuration. As you can see, it requires some tedious typing each time, so right now I'm writing scripts that automate the setup.
Here's one that lets you choose a mouse and keyobard and assign a second pointer to them (uses zenity, but can be adapted to use CLI only):
#/bin/bash
# vim: si ts=4 sw=4
shopt -s extglob
IFS=$'\n' DEVICE_LIST=( $(xinput list --short) )
for (( I=0; I<${#DEVICE_LIST[@]}; I++ )); do
IFS=$'\t' INFO=( ${DEVICE_LIST[$I]} )
[ "${INFO[2]:1:1}" = "s" ] || continue
NAME="${INFO[0]:6}"
NAME="${NAME%%*([[:blank:]])}"
ID="${INFO[1]#id=}"
[[ "$NAME" != *XTEST* ]] || continue
case "${INFO[2]:8:1}" in
"p" )
POINTERS+=( "$ID"$'\t'"$NAME" )
"k" )
KEYBOARDS+=( "$ID"$'\t'"$NAME" )
esac
done
POINTER="$( IFS=$'\t' zenity --title="Select device" --list --text="Choose a pointer" --column="Id" --column="Name" ${POINTERS[@]} )"
KEYBOARD="$( IFS=$'\t' zenity --title="Select device" --list --text="Choose a keyboard" --column="Id" --column="Name" ${KEYBOARDS[@]} )"
[ -n "$POINTER" ] && [ -n "$KEYBOARD" ] || exit 1
MASTER_NAME="second"
xinput create-master "$MASTER_NAME"
xinput reattach "$POINTER" "$MASTER_NAME pointer"
xinput reattach "$KEYBOARD" "$MASTER_NAME keyboard"
Don't ask me for more detailed instructions - I'll provide them when I have time.
Now I'm looking for suggestions where I could place them. Maybe a wiki page would be a good idea? I think it would also be easier to join efforts there. What do you think?

Yes, most of WMs get confused by multiple pointers, because they don't recognize mutliple focuses. That's one of the reasons why I run games in a dedicated X server with a bare-bones TWM setup. I launch it with:
xinit ./xinitrc.dual -- :1 vt8 -config xorg.conf.dual
xinitrc.dual:
twm -f twmrc &
DISPLAY=:1.1 xterm &
xterm
twmrc (compacted):
NoDefaults
NoIconManagers
NoTitleFocus
NoTitleHighlight
NoHighlight
RandomPlacement
UsePPosition "on"
OpaqueMove
DontMoveOff
NoTitle
BorderWidth 0
NoMenuShadows
# Bindings
Button1 = m4 : window|icon : f.move
Button3 = m4 : window : f.resize
Button1 = : icon : f.deiconify
"1" = m4 : all : f.warptoscreen "0"
"2" = m4 : all : f.warptoscreen "1"
"f" = m4 : all : f.focus
"Tab" = m4 : all : f.circleup
"Tab" = m4|s : all : f.circledown
"n" = m4 : window : f.iconify
In this config there are no window decorations, and "Super" (the windows-logo key) is the general window-handling modifier.
Bindings explained:
Super+LMB - drag windows around
Super+RMB - resize windows
Super+1/2 - warp pointer to screen 0 or 1
Super+f - lock/unlock focus
Super[+Shift]+Tab - cycle between windows (more precisely, cycle their stacking order)
Super+n - iconify (minimize) window; iconified windows are restored by clicking on icons.
Why do I see two entries for the USB keyboard ? (ID8 / ID9)
It seems that some devices report themselves that way. Maybe they have some functions that should be logically separated, or something like that. Then, I think it's best to keep them under the same master device to avoid confusion. I've updated my zenity script to allow selecting more than one device:
#/bin/bash
# vim: si ts=4 sw=4
shopt -s extglob
IFS=$'\n' DEVICE_LIST=( $(xinput list --short) )
for (( I=0; I<${#DEVICE_LIST[@]}; I++ )); do
IFS=$'\t' INFO=( ${DEVICE_LIST[$I]} )
[ "${INFO[2]:1:1}" = "s" ] || continue
NAME="${INFO[0]:6}"
NAME="${NAME%%*([[:blank:]])}"
ID="${INFO[1]#id=}"
[[ "$NAME" != *XTEST* ]] || continue
case "${INFO[2]:8:1}" in
"p" )
POINTERS+=( "$ID"$'\t'"$NAME" )
"k" )
KEYBOARDS+=( "$ID"$'\t'"$NAME" )
esac
done
MASTER_NAME="$( zenity --title="Enter name" --entry --text="Enter name for new master" )"
[ -n "$MASTER_NAME" ] || exit 1
POINTER="$( IFS=$'\t' zenity --title="Select device" --list --multiple --text="Choose a pointer" --column="Id" --column="Name" ${POINTERS[@]} )"
[ -n "$POINTER" ] || exit 1
KEYBOARD="$( IFS=$'\t' zenity --title="Select device" --list --multiple --text="Choose a keyboard" --column="Id" --column="Name" ${KEYBOARDS[@]} )"
[ -n "$KEYBOARD" ] || exit 1
xinput create-master "$MASTER_NAME"
IFS="|"
for ID in $POINTER; do
xinput reattach "$ID" "$MASTER_NAME pointer"
done
for ID in $KEYBOARD; do
xinput reattach "$ID" "$MASTER_NAME keyboard"
done
@Darksoul71
This should be all you need. You can adapt my xorg.conf.dual (it's in one of my previous posts), but the way of configuring multiple screens can differ between display drivers, so you'll have to find out how to do it in your case.
EDIT: Forgot to mention:
@Darksoul71
As to you keyboard problem - do you set the client pointer for the second instance's window?
Last edited by Rad3k (2010-09-29 22:06:08)

Letter of Appreciate to Devs, and two requests

Now that the nouveau and NVidia junk has been sorted out, I'm back to Arch. (I use it for my consulting business, when the graphics work that is)
I want to say Thank You, to the developers for all their hard work. They get yelled at all too often, and get thanked way too little.
I'm an older guy (Vietnam Era vet), and I appreciate this distro, more than the others because of it's rolling release nature, and by it's giving control to us, the users, in what is or is not included in our individual systems. As the date to the left of this post uindicates, I am no noob to Arch.
The only times I've had to use something else is when the graphics sub-system was broken. I had to work, and the CLI did not allow me to do that work. I freelance to make up the difference in income that VA disability doesn't cover.
I know about the finger pointing at NVidia, Xorg, and choose not to assign blame. I just don't have enough technical info, and I really hate pissing contests anyway, so I choose not get involved.
The Nouveau drivers in my experience have been unusable in a production environment (especially CAD). I have an NVidia GTX-560Ti card, and on the Nouveau site they don't support it well. 400's and 600's, but the 500's seem to have been passed over for what I will assume are legitimate resons within their dev community.
You see I use a commercial CAD package I purchased for Linux. I need the graphical sub-system to not break, or I have a work stoppage. Yes, I now dual-boot with Winsucks with the same CAD package installed therein, so work can continue in case of breakage......now..... ( at one time Arch was the only OS installed, until an NVidia driver update broke the entire system and I missed work deadlines). I would welcome the day when I can return to the paradigm of only Arch being installed. Windows 8 is a disaster (in the preview edition I VM'd). I will refuse to use it, as I do Gnome 3.*. (I need a traditional desktop where I can work, with multiple workspaces, and not have to play with a smarphone. I acknowledge that some users prefer this. So be it. I just don't want this paradigm shift shoved down the throats of traditional users.) I prefer MATE over XFCE. More flexibility is the reason, more choices for me the user without having to do a lot of things by hand (hint = gtkrc editing). For now, I use XFCE.
Request #1 is this:
As updates go, I humbly request that a little more time and QA happen on the graphics level so "pacman -Syu's" don't break the GUI because of some issue with either NVidia, Xorg, or whomever. And in saying QA, I do not implicate the package maintainer, but am suspicious of NVidia releasing an update that screws the pooch, so QA means a little more testing on NVidia's/Xorg's latest updates before being released to the wild.
I would welcome the day when nouveau is ready for primetime, and I am saddened that NVidia is not cooperating with them. Considering the Nouveau dev's have had to reverse engineer everything, I'd say they've done a stellar job......but they are not there yet for those of us in a production environment, regardless of which distro one chooses.
I use Arch also becasue of it's currency, and therein lies the catch 22. Currency vs stability. I also use Arch so I do not have to reinstall every 3-6 months to get current packages, and wasting all the time it takes to install a new release and get it all setup the way one likes. I use Arch also because it gives me control over my system.
I'm just asking for a tad more attention to detail if possible, to the graphical subsystem when updates occurr. If I had the cash, I buy a "butcher machine" for the sole purpose of being a testing resource for Arch, but seing that I am disabled, cash is tight so I can't help in a meaningful way. And I am NOT blaming the NVidia maintainer. NVidia is just as guilty for not keeping current with xorg developments, etc. I just ask that if an NVidia update breaks things in testing. Warn us on the main page as you do for init updates, etc. Then don't release the problematic update to the wild, or if possible, give us workarounds to implement on our systems to keep us online.
Request #2 is:
Please bring the MATE desktop mainline. I've used it in Mint before returning to Arch, and with the GTK 3 libs, it performs well, is compatible with newer GTK3 based apps, yet maintains the traditional desktop for those of us who are business oriented. I left Mint becuase I could get the most current libs needed to run the latest updates of my CAD package, not at least, without butchering the system, and causing a lot of downstream headaches for myself. Ubuntu, Mint's parent is getting prety wierd in how they do things compared to a mainstream distro. Compiling many things manually and adding them to your system causes more downstream issues, than the ones you fix for your immediate problem.
I know that the latest packages may break something....That's part of the risk of currency.....I am just asking that new packages don't take the system down to it's knees. My definition of "System" = base system plus functional GUI in XOrg.
Yes, I review the main Arch news before I update. So I am not a "stupid" user of the distro where warnings are ignored, like some I read in the forums.
I realize that others may have had different experiences than me, and I by no means want to minimize those experiences in this post.
Again, I appreciate Arch...And I thank all the devs for their volunteer time and effort. Sincerely so.
Respectfully,
Dave
PS: The new install method: Not that hard. I used my netbook so I could read the wiki, while I was reloading my main workstation from scratch to check it out. And in about the same time, I had a new Arch installed. A few rough edges, but all easily fixed. Once the wiki gets fleshed out, I suspect all the rough edges will vanish into thin air.

dcbdbis wrote:I humbly request that a little more time and QA happen on the graphics level so "pacman -Syu's" don't break the GUI because of some issue with either NVidia, Xorg, or whomever.
You can do a "test update" without spending any money: Create a 10 GB partition and install Arch Linux on it with the same packages you have on your "real" installation. Just wait to do an update until you have time to deal with possible errors.
dcbdbis wrote:Please bring the MATE desktop mainline.
MATE is available through the AUR and through other repositories, as is mentioned on the wiki page. Is there a benefit to having it as part of the official Arch Linux repositories?
Last edited by drcouzelis (2012-07-27 20:37:51)

Just here to say thanks

and best wishes to the development team.
I was a Slackware user for the past several months...the first distribution that stayed on my PC for more than two weeks before I wanted to try something different. Arch combines everything that I loved about Slackware with everything I loved about Gentoo with everything I loved about Debian The init scripts are simple, yet elegant. The package management system is simple yet robust, and the ABS is great for custom builds and for using software not available in Arch's repositories.
It's not perfect, of course, but it's the closest thing that I have yet to find when it comes to distros.
My only complaints were that Mozilla was compiled with --enable-xft but not --disable-freetype2...disabling freetype2 forces Mozilla to use only your xft configuration and thus fonts look better and follow the settings in /etc/fonts/fonts.conf and Vim was not compiled for GTK2. ABS made fixing those things on my system a breeze though.
Keep up the good work, and good luck with this distribution!

y'know that was what really brought me over to Arch--I had left Linux almost completely for FreeBSD, and then begun playing with Gentoo which brought me back to Linux. But the gcc upgrade from 2.95 to 3.x was a blip--one of the developers did some scripts, but the whole thing was quite iffy and most of us (it seems to me that this was back when the user base was still small) tarred up /home and reinstalled.
With Arch, typed pacman -Sy gcc and in two minutes it was done.
As someone posted in another thread somewhere, despite the claims of source based distros about how much better they work being individually tweaked, I've found Arch to be as fast or faster than most. (There was an interesting comparison recently, twixt Gentoo, Deb and Mandrake--where they found, to their surprise, that MD was almost as fast as the others.)

Problem with local mirror syncing

Hi mens! And sorry, my English is bad.
I rewrite script from here for syncing mirror for my local network.
That's what happened http://ix.io/1aK or http://paste.pocoo.org/raw/263591/ or permanent http://ix.io/user/atommixz
It support http (maybe ftp?) mirroring other repos. Using lftp for it.
#!/bin/bash
# The script to sync a local mirror of the Arch Linux repositories and ISOs
# Copyright (C) 2007 Woody Gilk <[email protected]>
# Modifications by Dale Blount <[email protected]>
# and Roman Kyrylych <[email protected]>
# and Vadim Gamov <[email protected]>
# and Aleksey Frolov <[email protected]>
# Licensed under the GNU GPL (version 2)
USECOLOR=yes
. /etc/rc.d/functions
# Filesystem locations for the sync operations
SYNC_HOME="/home/mirror"
SYNC_LOGS="$SYNC_HOME/logs"
SYNC_FILES="$SYNC_HOME/files"
SYNC_LOCK="$SYNC_HOME/mirrorsync.lck"
SYNC_REPO=(core extra community multilib iso archlinuxfr catalyst)
#SYNC_REPO=(archlinuxfr catalyst)
typeset -A REPO_URL
REPO_URL=(
[archlinuxfr]='http://repo.archlinux.fr/x86_64'
[catalyst]='http://catalyst.apocalypsus.net/repo/catalyst/x86_64'
#SYNC_SERVER=distro.ibiblio.org::distros/archlinux
SYNC_SERVER=mirror.yandex.ru::archlinux
RATE_LIMIT=50 # in kb/s
TIME_OUT=5 # in sec
# Set the format of the log file name
# This example will output something like this: sync_20070201-8.log
#LOG_FILE="pkgsync_$(date +%Y%m%d-%H).log"
LOG_FILE="pkgsync_$(date +%Y%m%d).log"
#Watchdog part (time in seconds of uninterruptable work of script)
# Needed for low-speed and/or unstable links to prevent
# rsync hunging up.
# New instance of script checks for timeout, if it occurs
# it'll kill previous instance, in elsecase it'll exit without
# any work.
WD_TIMEOUT=10800
# Do not edit the following lines, they protect the sync from running more than
# one instance at a time
if [ ! -d $SYNC_HOME ]; then
printhl "$SYNC_HOME does not exist, please create it, then run this script again."
exit 1
fi
if [ -f $SYNC_LOCK ];then
OPID=`head -n1 $SYNC_LOCK`;
TIMEOUT=`head -n2 $SYNC_LOCK|tail -n1`;
NOW=`date +%s`;
if [ "$NOW" -ge "$TIMEOUT" ];then
kill -9 $OPID;
fi
MYNAME=`basename $0`;
TESTPID=`ps -p $OPID|grep $OPID|grep $MYNAME`;
if [ "$TESTPID" != "" ];then
printhl "exit";
exit 1;
else
rm $SYNC_LOCK;
fi
fi
echo $$ > "$SYNC_LOCK"
echo `expr \`date +%s\` + $WD_TIMEOUT` >> "$SYNC_LOCK"
# End of non-editable lines
# Create the log file and insert a timestamp
touch "$SYNC_LOGS/$LOG_FILE"
printhl "Starting sync on $(date --rfc-3339=seconds)" | tee -a "$SYNC_LOGS/$LOG_FILE"
for repo in ${SYNC_REPO[@]}; do
repo=$(echo $repo | tr [:upper:] [:lower:])
printhl "Syncing $repo" | tee -a "$SYNC_LOGS/$LOG_FILE"
NEXT=false
for i in ${!REPO_URL[*]}; do
if [ $i = $repo ]; then
mkdir -p "$SYNC_FILES/$repo"
cd "$SYNC_FILES/$repo"
lftp -c "\
set xfer:log no; \
set net:limit-rate $[RATE_LIMIT * 1000]; \
mirror \
--delete \
--only-newer \
--verbose=3 \
${REPO_URL[$repo]}" | tee -a "$SYNC_LOGS/$LOG_FILE"
date --rfc-3339=seconds > "$SYNC_FILES/$repo.lastsync"
NEXT=true
#sleep $TIME_OUT
fi
done
if $NEXT; then continue; fi
rsync -rtvHh \
--bwlimit=$RATE_LIMIT \
--no-motd \
--delete-after \
--delete-excluded \
--prune-empty-dirs \
--delay-updates \
--copy-links \
--perms \
--include="*/" \
--include="latest/*x86_64.iso" \
--include="latest/*sum*.txt" \
--include="archboot/latest/*.iso" \
--include="os/x86_64/*" \
--exclude="*" \
$SYNC_SERVER/$repo "$SYNC_FILES" | tee -a "$SYNC_LOGS/$LOG_FILE"
# Create $repo.lastsync file with timestamp like "2007-05-02 03:41:08+03:00"
# which may be useful for users to know when the repository was last updated
date --rfc-3339=seconds > "$SYNC_FILES/$repo.lastsync"
# Sleep 5 seconds after each repository to avoid too many concurrent connections
# to rsync server if the TCP connection does not close in a timely manner
sleep $TIME_OUT
done
# Insert another timestamp and close the log file
printhl "Finished sync on $(date --rfc-3339=seconds)" | tee -a "$SYNC_LOGS/$LOG_FILE"
printsep >> "$SYNC_LOGS/$LOG_FILE"
# Remove the lock file and exit
rm -f "$SYNC_LOCK"
unset REPO_URL
exit 0
But I'm have problem. If I'm run
sudo pacman -Syu
on my server, it's fine work.
[atommixz@fileserver ~]$ date; sudo pacman -Syu; echo "------"; date; sudo pacman -Syu
Сбт Сен 18 19:55:47 MSD 2010
:: Синхронизируются базы данных пакетов...
core 35,7K 14,2M/s 00:00:00 [#############################################################] 100%
extra 465,9K 189,3M/s 00:00:00 [#############################################################] 100%
community 383,1K 198,6M/s 00:00:00 [#############################################################] 100%
archlinuxfr не устарел
:: Запускается полное обновление системы...
нечего выполнять
[atommixz@fileserver ~]$ date; sudo pacman -Syu
Сбт Сен 18 19:55:48 MSD 2010
:: Синхронизируются базы данных пакетов...
core не устарел
extra не устарел
community не устарел
archlinuxfr не устарел
:: Запускается полное обновление системы...
нечего выполнять
But if I'm try it on my desktop, it work wrong. Always reget base. But it is updated properly.
[atommixz@relentless ~]$ date; sudo pacman -Syu; echo "------"; date; sudo pacman -Syu
Сбт Сен 18 19:58:42 MSD 2010
:: Синхронизируются базы данных пакетов...
core 35,7K 34,0M/s 00:00:00 [#############################################################] 100%
extra 465,9K 58,7M/s 00:00:00 [#############################################################] 100%
community 383,1K 57,9M/s 00:00:00 [#############################################################] 100%
multilib 19,3K 34,7M/s 00:00:00 [#############################################################] 100%
archlinuxfr 18,6K 42,6M/s 00:00:00 [#############################################################] 100%
catalyst 2,2K 67,7M/s 00:00:00 [#############################################################] 100%
:: Запускается полное обновление системы...
нечего выполнять
Сбт Сен 18 19:58:43 MSD 2010
:: Синхронизируются базы данных пакетов...
core 35,7K 34,0M/s 00:00:00 [#############################################################] 100%
extra 465,9K 58,7M/s 00:00:00 [#############################################################] 100%
community 383,1K 64,8M/s 00:00:00 [#############################################################] 100%
multilib 19,3K 48,9M/s 00:00:00 [#############################################################] 100%
archlinuxfr 18,6K 38,9M/s 00:00:00 [#############################################################] 100%
catalyst 2,2K 55,5M/s 00:00:00 [#############################################################] 100%
:: Запускается полное обновление системы...
нечего выполнять
What am I doing wrong?

I'm not sure, but your script may be too old. That is, it may no longer work.
I believe that creating a local mirror is discouraged due to the high bandwidth needed to do that.
Perhaps you could try this instead.

[network: messenger] gaim 1.1.0

It's my first pkgbuild, but it seems to work for me...
PKGBUILD:
pkgname=gaim
pkgver=1.1.0
pkgrel=1
pkgdesc="Gaim instant messenger"
url="http://gaim.sourceforge.net/"
license=""
depends=('startup-notification' 'gtk2' 'libao' 'gnutls' 'tk' 'gtkspell')
conflicts=()
replaces=()
backup=()
install=
source=(http://heanet.dl.sourceforge.net/sourceforge/gaim/$pkgname-$pkgver.tar.gz)
md5sums=('6f08dc6f20fd788588a1313e50c638fa')
build() {
cd $startdir/src/$pkgname-$pkgver
./configure --prefix=/usr
make || return 1
make prefix=$startdir/pkg/usr install
"pacman -A gaim-1.1.0-1.pkg.tar.gz" for new install
"pacman --upgrade gaim-1.1.0-1.pkg.tar.gz" for upgrade
I'll hope it works for you to

tip: read about ABS, it is a good method for upgrade the package
when it is not in the arch binary repositories
PD: the pkgbuild works

SOLVED .deb installation help

Hi, I want to install google chrome on my arch/openbox and I don't want to use another browser, I don't want chromium. I downloaded google chrome as a .deb and I don't know how to install it.
I already read a bunch of topics about this subject and I couldn't find a clear answer. I found two options that I could try :
1) PKGBUILD, I understand it's made to make packages to install with pacman, but still I have no clue how to convert a .deb with pkgbuild, I'm pretty new with linux.
2) dpkg, that's how I installed my google chrome on my lubuntu, but I can't find dpkg with pacman when I type : pacman -Ss dpkg
Can you help me a little bit here? Maybe you know a simple guide to how to install .deb ...It'd be nice to have something clear since there is many confusing thread.
Thank you and merry xmas.
Last edited by trixrabbit (2012-12-25 07:52:37)

A friendly advice: generally it's best to stick with the package format of your distribution, in this case, Arch Linux packages, especially if you're new to Linux. DEB is the package format of Debian based distros like Ubuntu and should best be used for such distros only that support it. dpkg is a program specific to such distros.
Usually you should find all you need either in Archs official repositories or in the AUR. If not, it's best to learn how to make an Arch package yourself which you'll be able to install with pacman, so it can keep track of everything that's installed on your system and help you maintaining clean installations and preventing chaos.

Using repo filelists (like with "pkgfile")

Arch's repositories provide lists of the files contained in each package. As far as I know, the only thing making use of that data is the "pkgfile" tool from community/pkgtools. Are there other such tools that I don't know about? It seems strange that there is a feature implemented in the repos that no "official" package make use of.
I also remember that, a while back, there used to be a link from each package's description page to its file list. Where did that feature go and why? I miss having it there.

There are no other tools that I know of.
The website feature disappeared ages ago... I do not know why, but given we generate file lists already, it should not be hard to re-add. Maybe file a feature request.

About post install scripts...

hi,
i wonder if there is a solution for the post-install script problem to prevent some commands like "rm -rf /" to be executed. I know that Gentoo hasn't got this problem because ebuild scripts use a particular language that doesn't implement a such command. But for Arch which uses bash, could this problem be really resolved ?
Comete.

Toth wrote:[...]
A package is software, and you have to trust that the software you are downloading is legit. If it is, and you trust the packager, you have no reason not to trust the package.
exactly ---> do not download from sources you dont trust
(if you ever worked on a windows, you know that already, because there is a lot of evil things around (spyware, virus, bacteria, nice VB-scripts, nice JS-scripts ...))
-> the problem about security: if your OS becomes popular, then the bad guys see it as enough popular to do something bad with it
Toth wrote:I assure you, if the package maintainers ever run across a submitted package with "rm -rf /" in the post install script, they won't add it to the Arch Linux repositories. If/when you find a 3rd-party Arch package for download, that's a risk you are going to have to take. Check the scripts, check the source of the program (if you're paranoid). That's about all you can do.
what about a check (in pacman), to give warning, that something will be done with / ... as i know, there is no line to run with / as parameter for a normal reason during an installation or upgrade
if you download a non-archlinux.org package (downloaded from somewhere else), the install procedure is now this:
> pacman -A bad-0.1-1.pkg.tar.gz
loading package data... done.
checking for conflicts... done.
installing ... done.
>
the check would look like this:
> pacman -A bad-0.1-1.pkg.tar.gz
loading package data... done.
checking install/upgrade scripts ...
Warning: Found lines in Install-Script that contain / :
line 3: do something
line 4: cd / <---
line 5: rm -rf *
line 6: echo "I was a bad script"
Continue with installation [y|N] y
checking for conflicts... done.
installing ... done.
I was a bad script
>
this will at least give the user a chance to look at code that contain "/ " in the install/upgrade/remove-scripts and have some control
of course this will not check the package-content :-( ... but at least you will have one leak closed particulary

Wine/Winetools: IE6SP1 installation stalls

I searched but did not find an answer, but if this has been answered anyway, I apologize for this unnecessary post.
I'm trying to set up Wine version 20050628-1 from Arch's repositories using Winetools.When I get to IE6SP1, the installation stalls at about 44%. Wine doesn't freeze. I can press cancel button on the installation window, and it exits. The installation just doesn't proceed from 44-47%.
Winetools notifies that it hasn't been tested with that version. Does this happen because of the version incompatibilities, or because of something else? What am I doing wrong?

this has been reported in other forums (for other distro)
You might want to check this link
http://sidenet.ddo.jp/winetips/config.html
the web master also provides a tool allowing ppl to install IE with wine, and from what I heard, it works fine with wine0725

[ANNOUNCE] pacboy: automatic repository creation

EDIT: The name is now pacboy.
Hey everyone.
Just wanted to share with you a new piece of software that I am writing that will allow repositories to be created and managed for Arch Linux repositories.
This software, called "pkgman" which is written in PHP and Bash will allow new packages to be compiled, uploaded, verified and offered for download.
pkgman will allow for many really cool things, as you will learn as you read on
pkgman works by having a slightly modified makepkg script, that will log the whole makepkg process. The log file that is outputted will then be md5sum'd, with the md5 hash being stored in the .pkginfo file.
Once makepkg has been completed, the user then runs the bash script 'pkgman-push', which will upload the following three files: PKGINFO, the log file, the built package and a PGP signature.
Once pkgman-push has completed the upload of the three files, the pkgman server will verify the integrity of all four files. If all the files are found to be valid, then the new package will go into a stage of having to be 'signed off' by a specified number of users. Once it has been signed off, the repository database is recreated, the old package deleted and the new one will take it's place.
Now here comes the *really* cool part. Packages are actually never really deleted off the database. Because of this, a user can specify a repository from any point in the past. For example, say if I want to stick with the [core] repository from 2nd Feb 2009, all I would need to do is find what the UNIX time is for that time (in this case, 1233562755) and stick that in my repository path like this:
[core]
Server = http://pkgman.local/repo/core/+1233562755/
When this is done, all packages offered by pacman will only be the packages from that particular point in time. Of course, this would only work on the actual pkgman website, and not on any of the mirrors (unless if Arch does take this up we are lucky to find a mirror that will want to mirror many past packages)
On the pkgman administration site, the admin specifies which users are 'trusted'. Only 'trusted' users can upload packages to the pkgman database.
For the weary user, they can install the view-build-log package, and then view the build log for a specified package(s). For example
view-build-log pacman
will ask pacman what the currently installed version of pacman is, what repository it is from, and then download the build log that matches that version.
Also, on the administration server, the admin can create new architectures, and specify which users have push permissions to what architectures. Only one install of pkgman is required for multiple architectures and multiple repositories.
pkgman will be able to provide dependency tracking, so that for example, a package can not be uploaded to core if it depends on a package in extra.
I will upload some screenshots when I get home from work.
Last edited by timgws (2009-02-02 22:43:46)

A screenshot of the repository overview. Notice how on the bottom it lists the different repositories, and shows how they are linked to each other. At the moment, a new repository can only depend on one repository (for example, you can't have a 'bob' repository that depends on both 'james' and 'extra')... I hope to add this feature soon.
This image shows me searching for the package 'bash'. Notice how it shows the current version in the repository, along with information provided by my pkgdesc.

What to do about the Catalyst unsupported Radeon X600?

So I have an older Radeon X600 PCIe card I'm using in a server right now, and unfortunately it appears there is no Catalyst support from ATI.
I need this card to support 3D acceleration because I need 3D acceleration in VirtualBox, which I run headless from the server. I tried using the open source drivers, but they don't appear to allow me to enable 3D acceleration in virtualbox (or even 2D acceleration).
Does anyone know of a way I could get 2D/3D acceleration working on this card? Would I need to drop back to a much older version of Catalyst (this card has been unsupported for a couple of years now)? Or do I simply have the open source stuff configured incorrectly?
Thanks!

Hi,
It is good to read that you have got it working. I dropped in here to give some information.
My older system had this card (X600 SE), and it worked perfectly with xf86-video-ati driver, with full 3D acceleration. This was using the one provided in the default Arch Linux repositories.
xf86-video-ati also works well for many RadeonHD cards (for example, my HD5670).
As for AMD's binary drivers,
AMD dropped support for this card in their FGLRX/Catalyst drivers some years ago. There is a legacy driver, but it will require an older version of Xorg.
You can find more information about it here, if you wish to use that driver:
http://support.amd.com/us/gpudownload/l … ng=English (x86 and x86_64 drivers)
I am not sure what could have changed in xf86-video-ati that drops support, natively, for this card. Perhaps it was removed and added to an alternative driver (xf86-video-radeon?).
I still have within my vicinity the card in question, but the machine that it is in, I gave to my family for shared use. That machine is currently running Windows XP, their favourite operating system (thus, I cannot test it to find out more).
Last edited by NightCrawler03X (2012-01-15 18:50:34)

Advice on Linux financial software

What do people recommend for home budgeting financial software on Linux and why do you think it stacks up better than others? (KMyMoney2 and Gnucash 2.0 are both in the Arch Linux repositories, and others are ready to be installed in the user supported repositories.) I'm just curious on how people feel about these packages, what there experiences have been like, are they stable, and so on.

skottish wrote:That was actually my first thought, but my thread is intentionally a little vague. I'm not the one who needs the budgeting software and she's the Anti-Nerd!!!! I'm curious: did you do it from scratch or did you work off a tutorial of some sort?
I made it up from scratch, it isn't anything fancy though. Here is a quick explanation of what it is:
I have a sheet for each account with 5 columns: "Date of Transaction", "Transaction amount", "Running total", "Transaction Type" and "Comment", so I get something like "16/07/2006, -£20, [running total here], ATM Withdrawl, Tesco cash machine"
I then have a sheet that gives me the current balance for each account, and then adds the current accounts (I have two), adds the savings accounts (I have two also), and subtracts the credit card balances, and gives me a final grand total.
The final sheet is a chart, and what I do is every Monday, I enter the current balance of each account. This way, I can see where my money is going on a weekly basis (weekly is enough for me) and I can see if I'm spending too much
Hope that is of some help, I would show you it but I don't really wanna send my bank balances around

Arch gaming repositories

Similar Messages

Maybe you are looking for