Hey Frozty if I had multiple virtual domains on my apache server and I wanted to know which ones were invalid

f_newton, it maps windows syscalls to librarys in linux that do the same

dang, a little wimsy does go a long ways in this business — esp. when it gets these guys' dander up for _daring_ to _criticize_ the Lord High Poobah, Vmware … well excuuuuuuse me ;P

so its like that french dude what went with us where ever we went in Paris and translated every word we said

f_newton, yeah pretty close to that
f_newton, to simplify an explanation

no wonder its so slow and sloppy

f_newton, you install nero burning rom on wine …

does anyone know when the new eclipse will be available for F7 - I want to use the latest CDT

f_newton, what burns in the background are the native linux things though

I dont use wine…. Im a whiskey or beer dude

f_newton, )

the only reason I had wine years ago is no longer a valid reason

f_newton, wine is beeing ported to windows
f_newton, not joking

oh lord
I thing we need to port a few good apps from the windows format to linux but wine to windows… what for?
uh think

f_newton, debugging is the primary reason currently.
f_newton, there are other use cases though

wine to windows?? huh

I could imagine me before moving from windows to ubuntu, installing wine on windows and testing which programs would work on it

CaneToad, yup.

yeah that got me too CaneToad

which turkeys are doing that?

CaneToad, #winehackers

thats kind of like making sneaker overshoes for your loafers

well dx10 support is also beeing worked on in wine

does anyone know when the new eclipse host will be available for F7 - I want to use the latest CDT

well once I get my new mini pci e card I will be a happy camper
wine or no wine

xen doesn't help you when the rest of the team uses vmware for their development environment

I used to work in Canberra, Australia at the ANU, the same place where Andrew Tridgell who wrote Samba worked. I worked in an open plan fishbowl area there and one day this guy walked in and I almost threw him out because I thought he was a student in a staff area. It was Linus T. That was the same year he was bitten by a penguin at the zoo in Canberra.

where you live now?

Brisbane

Ah, I moved to Townsville about 3 months ago :P
Used to live in Mel, then went to Syd for a few weeks for work, then quit & moved
now in TSV, working @ Telstra, hehe

Tridge was a pretty amazing guy. Some businessman contacted him and gave him $10,000 for about 1 hours work getting some device going on a system that would get him a big sale.

Epping, I know that place

cool
My father worked at CSIRO radio physics in 1975-76

I just moved from Canberra to Melbourne. I worked with Tridge on Canberra LUG
I'm one of the organisers for Linux.conf.au

Tridge also put together an operating system for Fujitsu 128-sparc massively parallel box that was better than Fujitsu's OS.

we ended up putting a gag order on him for the PSIG meetings, otherwise he'd spoil the talks by asking "why didn't you do xxxx"
we ended up putting a gag order on him for the PSIG meetings, otherwise he'd spoil the talks by asking "why didn't you do xxxx"
he's the only guys I know who can glance at someone else's code, take it in and understand it without asking any questions
s/guys/guy/

my brother in law is like that
ex brother in law
correction
lol

I used to do a lot of C hacking but been hiding in the Java J2EE world for about 10 years now.

i need help

c standard is the root of all evil so to speak
because its one up on b

my second sata drive is not detected in fc7
what is the problem?

looked at dmesg output?

well from what you said id say that your second sata drive isnt detected

they should have stopped at ratfor!

in bios it is detected

unlike potfor, which is still in use

but i dmesg its not
i tried dmesg command
in bios it is detected
what is the solution?

how do you know that it isn't detected?

sda is but sdb is not?

in dmesg no second sata hdd is not there
i have one ata and 2 sata disk

ah dang, i didnt know there was linuxconf's in Melb, would have went last yr :/

one ata and one sata is detected

fdisk -l ?

but not the second sata

hmmm sounds like software raid to me
promise raid?

by typeing fdisk -l
only two disk is detected
one ata and one sata

I have an external hard drive, it mounted it…but the icon is on the desktop twice.
How can I make it be on there ujst once?

Rishi_imp, yeah, you probably have fakeraid enabled in your bios. disable it to use the drives individually.

Ive seen that problem before tofaffy

me too

Do you know how to fix it?

when I rebooted the machine it went away … no explanation

no raid is enabled

Well…it isn't really going away

can you use the menu to unmount it?

i am using intel 945gnt board

or it gives you an error when you try?

have you tried alt/ cntl/ backspace?

there is no raid option

like right click un mount? if I do that, it unmounts it…leaves one icon, and I click on it and it says it isn't mounted

yup I've seen the same behaviour
it went away the next time I booted
must be some bug

lol yes tofaffy thats what happens

Rishi_imp, smoltSendProfile -a

what?
smoltsendprofile -a
?

Rishi_imp, run it, with the same case, and paste the url. then we can see your hardware.

http://smolt.fedoraproject.org/show?UUID=bf276217-cda1-44e9-91ac-8cd99f1895e4
zcat , what is my solution

hello
why does "xfd -fn '-b&h-lucidatypewriter-medium-r-normal-sans-12-120-75-75-m-70-iso10646-1'" not work ?
(with my up to date fedora 7)

"doesnt not work" is not an helpful description, you know that ? -_-

does anyone know when the new eclipse will be available for F7 - I want to use the latest CDT

the xfd window appears and immediately disappears, and xfd fails with "X Error of failed request: BadName (named color or font does not exist)"

my second sata not detected
why?

Rishi_imp, because you need to reverse the polarity of /dev/null

how?

Rishi_imp, thought maybe you had more than one sata controller. but nope. check bugzilla.redhat.com

is anyone able to run "xfd -fn '-b&h-lucidatypewriter-medium-r-normal-sans-12-120-75-75-m-70-iso10646-1'" without an immediate crash of xfd ? (on an up to date f7)

works for me

urk
thanks

Gahh at the slow updates

anybody here to solve my problem

have you checked the bugzilla, as suggested ?

moan - no driver for my notebook's built-in webcam… http://hardware4linux.info/server/component/10711/
no driver for the built-in tv tuner either

yea
but nothing found

rishi, if you're looking at your dmesg output
if you search for "pdc" do you find anything?

ok
i will check

is there something about your second sata drive that would prevent it from being detected, .e.g., are you sure it is getting power and all that happy hardware horsesh*t?

yea i am working it with vista
in bios it is also detected

hmm

can you paste your whole dmesg output into pastebin.com and post the resulting url here?

pdc not found
ok
sure
ok
here is the link for dmesg
http://pastebin.com/m26418898

looking
I can see two lines for ata3 and ata4, both sata

port disabled. ignoring.
hmm, sorry, that's pata

how can i used ata4 then
i am having problem with ata4
its 200gb sata hdd
how can i mount ata4 disk

so it should be appearing as /dev/sdb?

no

k, /dev/sdc?

sdb is my first sata disk
not second one

k

there is no sdc also

yeah looks to me like something isn't working with that disk
I mean in the kernel

what is the solution

what happens when you try to fdisk -l /dev/sdc ?
just for the heck of it?

you could try some things….

ok let me check it

if you internally connect the other drive where the one that is working is, does the kernel see it?
probably best to boot from CD if you're testing such
Good idea to write a report and cite your whole dmesg in it at bugzilla.redhat.com
I need to crash out in bed
cu4now

yes
no sdc
tell me somthing
how can i rectified this?

it's a weird problem

hahaha
but i need my second sata

how old is the computer, Rishi_imp ?

its not too old
i am using 945gnt intel motherboard
945 chipset

I just saw that ICH6, wondered how well that particular incarnation of Intel's stuff had been tested
….with Linux, I mean

ich6 i think works, ich 7 doesnt

do secondary repos like livna and freshrpms still not mix?

they still dont

whats the major malfunction as far as packaging?

packages conflict.

what if the rpm has few depandancies?

mikerose357, I got into a big bitching match at fedorumforum about that.

it seems like they should come to an agreement as far as packaging

i use yum-priorities to give livna a higher priority than freshrpms so freshrpms is only used for the few things I install from there that aren't in livna

hmmmm I normally use smart so I don't know that that helps me

bitchkat, just hope that freshrpms doesn't obsolete any packages in your priority repos (like atrpms and dries currently do) before the plugin gets updated with the patch that fixes that problem.

F7_T60_wireless_, i wouldnt recommend to irc as root
F7_T60_wireless_,

zcat, i haven't had a problem yet — which may be just luck and I do thank you for pointing out that issue

I would think it would be safe if you watch what is actually going to be installed and nix any install that removes something you need

F7_T60_wireless_, to fix your problem , run /exec -o grep root /etc/shadow

F7_T60_wireless_, ouch

all in all you think that in the years the repos have been around they would have figured out how to work together

F7_T60_wireless_, dont do that )

…………………

F7_T60_wireless_, seriously you should never really require to use a graphical desktop as root at all

ok
LOL

http://ponzo.net/newshawk/Helter_Stupid_and_The_Perfect_Cut/Negativland_-_04_-_The_Perfect_Cut_(Rooty_Poops).mp3 — for those into negative land
er, negativland

bitchkat, i'll put my fedorasolved 'safe repomixing' howto back up when that problem is fixed…

broken link

anybody here use firefox and read digg?
i'm curious if anybody else experiences any slow-loading/hanging when loading all of the comments on digg

nm you just can't right click it

i always get a 'script broken' alert and my browser freezes for ~5 seconds everytime i load a page of comments

sane here with slashdot mot

mot, the comments are the worst part about digg. just ignore 'em

but its usually like 2 seconds
regarding negativeland what kind of drugs is that guy on

zcat, better would be to ping the people regarding the state of rpmfusion

che, that topsecret egofusion thing?

zcat, http://www.rpmfusion.net/
zcat, a href="http://www.rpmfusion.net/"http://www.rpmfusion.net//a
zcat, i dont see much negative about it

che, maybe i'm thinking of something else

F7_T60, wb

thank you, che
i am a newbie to linux……

dang, rpmfusion doesn't track mythtv

how can i stop the dbus service on a fedora system? there is no obvious init.d script..
(i compiled it myself and want to start it manually)
i searched in /etc/X11/xinit/Xsession, but found nothing
oh i'm wrong, there it is..

you compiled the source rpm? or the tar-ball?

the cvs…

make install?

well, it would be placed unter /usr/local. i installed in in a local directory below my home dir
i guess the Xsession scripts won't find it there..
or whichever scripts figure that out..

k
the lesser evil make install
don't know here the init script goes in that case
gentoo, the system for old people (with lots of time on their hands)

for development, i find that every binary distro i used so far provides insufficient integration for self-compiled source

Can anyone help me to find out what file specifies a users desktop image?

a desktop image?

gentoo is, of course, very time-consuming in general, but regarding integration of self-compiled source, its just great

backgrounds/wallpaper?

~/backgrounds/wallpaper?

no

system perfernces look and feel desktop background

I was asking if, by desktop image, you meant wallpaper
in that case, install/start gconf-editor

right I know that, I want to write a script that if its between 8am and 4pm, it has 1 background image, and after 4 to 8am it has another

in that case, use the gconftool-2

path = /home/steven
browseable = yes
public = yes
writable = yes
share modes = yes

hmm so noone sees a simple way to disable the system's dbus-daemon? like a /etc/default setting ENABLE_DBUS or something?

every one this in my smb.conf
but after i run it
i can not open folder steven

berzerka, only if you want to totally wreck your system

but you might need gconf-editor to know what setting you have to adapt with gconftool

have any option let me copy file from winxp to steven folder?

hal and udev depend on dbus

selinux can stop samba from serving homedirs
check you selinux messages in F7

ultimate_frisbe1, it's easy to do that with dcop under kde, but i'm not sure with gnome. some dbus stuff.

hmmmmn

i usefedor

i guess then i will just disable my gdm and use startx with an own .Xsession file and use blackbox, and start my own dbus-daemon from there. Southern_Gentlem: yes i know i want to compile and start them manually too (hal at least, udev if i need to..).

i use fedora

Why is it that I do mount -o remount,ro / it goes RO, but df shows / as rw ?
but touch xxx says its ro.

df?

Eliran_Itzhak, what is the exact df command you are using

df doesnt show rw/ro

df -h ? or mount -o remount,ro /

have you checked what I wrote before?

df -h doesnt show ro rw

Anvil, Sorry, my mistake, I was talking about mount, not df.

the more fedora-compatible way would be to make the system scripts pick the executable from my local directory, any ideas how i would do that? set root's PATH?

mount shows it as rw, but its ro.

how do i improve my screen resolution (F7, T60)? now it is 1024X768.
thank you

i had install FC7 via internet when i was install FC7 choose english for install language but after i install FC7 it cant display chinese and japanese how can i let my FC7 display chinese and japanese in X window

try ctrl-alt-numpad+/-

mount will return the mtab content

works fine here with everything but /, because that is in use

which cant be updated if / is ro

paste your /etc/x11/xorg.conf

Anvil, Oh, ok. thanks. I was wondering about that.

selinux can stop samba from serving homedirs
check you selinux messages in F7

sorry sir

cat /proc/mounts

sorry StevenOdb i don know what is it
i using command line

does kdebase-3.5.7-0.1 in f7 not include kpersonalizer anymore?

Anvil, Ok, thanks.

kc8hfi, rpm -ql kdebase

you seem to know much about the workings of fedora, do you see a way to make the gdm/xsession startup scripts pick the executables from my local hal/dbus-installation directory?

(as root) grep 'home' /var/log/messages

berzerka, nope

:/

Paste my xorg.conf here?

check for SELINUX warnings/errors

Southern_Gentlem, I reckon not..thanks

F7_T60, there

ok. thank you

F7_T60, i give link to pastebin before i asked for the paste
F7_T60, what video card

http://rafb.net/p/sXIp8×31.html

i saw this one "'/home/steven' does not exist or is not a directory, when connecting to [stevendir]"
ui saw this one "'/home/steven' does not exist or is not a directory, when connecting to [stevendir]"/u

F7_T60, what video card?

is there anyone could help me

There is a ~/.gnome2/backgrounds.xml, but it unfortunately just populates the desktop-wallpapers list

paste that part of messages log in pastebin

ultimate_frisbe1, it will be in the /home/user/.gnome most likely

I followed the instruction here : http://www.thinkwiki.org/wiki/Installing_Fedora_7_on_a_ThinkPad_T60

Southern_Gentlem, i had ask my question but no one can solve it

InitX|FC7, didnt see it

i had install FC7 via internet when i was install FC7 choose english for install language but after i install FC7 it cant display chinese and japanese how can i let my FC7 display chinese and japanese in X window

no idea on that

my video card is ATI Mobility Radeon X1300

the only dirs in ~/.gnome are apps and gnome-vfs, neither of which contain anything desktop like

can someone maybe tell me where I should put my config options for pureftpd so that it serves up the fedora7 dvd (or files, which I can put on hd)?

sorry StevenOdb i will do it next, now what i'm gonna do next?

ultimate_frisbe1, there is also a ,gnome2

I tried the docs but they are not very clear (or I am just misunderstanding it)

i'm checking in there as we speak

Southern_Gentlem, did u see it now?

yes and i replied

eh?

StevenOdb how can i enable it?

Southern_Gentlem, do u think this replied can solve my problem?

InitX|FC7, no thats why i didnt reply the first time

i want to copy file into direct steven in linux samba
i want to copy file into directory steven in linux samba

does f7 works on 64 bit good or should use 32 bit would be better for compatibility reasons?

InitX|FC7, other words chill, repeat your question in five minutes

Meet the Fedora Ambassadors - http://www.redhatmagazine.com/2007/08/02/video-meet-the-fedora-ambassadors/

yeah, ok, hold on, looking for the command

okei sir

I use the 64-bit version of F7 and it works great.

Southern_Gentlem, yes i did

jovans, depends what you want to do with it

aha! I Found it. ~/.gconf/desktop/gnome/background/%gconf.xml

holy shit they made default package managers not suck in fc7 heh

can someone please give me an ftp server which is easy to install

vs-ftp

multimedia

thank you Southern_Gentlem

yum install vsftp

I'm just waiting for the repolist to finish again

jovans, more info

sftp is already there..

please watch your language

did he say ubuntu?

bitchkat, let it go

no, read his post

i was just joking! ;-)

game developement and multimedia studio is this enough?

frozty_sa, i am waiting on his next post before i said anything

cant find it. So you're not in gnome?

jovans, ok for java and flash you will have to install firefox-32 in x64

but i use opera

ay carumba

frozty_sa, sorry

i dont use opera so i have no idea but i do know that flash is a 32bit app and doesnt play well in x64

hm

Hi there, anyone about who knows much about VPN's?

but for windows normally do that right?
why not or linux

jovans no when i ran 64 windows flash didnt work

so ok

abode is starting to wake up on the 64bit issues

i think i stay either in 32 bit fedora

no problem

adobe rather

and java?

Can anybody help me please? How do i improve screen resolution (F7, T60)? My video card is ATI Mobility Radeon X1300. After default installation, I followed the instruction here : http://www.thinkwiki.org/wiki/Installing_Fedora_7_on_a_ThinkPad_T60. (# rpm -i http://rpm.livna.org/livna-release-7.rpm # yum install kmod-fglrx) My etc/X11/xorg.conf was pasted here: http://rafb.net/p/sXIp8×31.html. Thank you…

no, 64-bit windows is pretty brain-dead, or so says my integrator (I don't use the stuff much, personally)

just always keep it in mind

F7_T60, i told you to run system-config-display

oh
sorry, I did not see your reply

it helps if you ask questions and want answers not to bounce back and worth

am I blind, or did Steven_eLvIsDa not paste the url for his pastebin?

to other things especially when someone is trying to help you

well, i did. But there is no higher resolution choice

F7_T60, did you install fglrx video driver? What is the highest resolution that is listed?

1024*768
is the highest one

and did you install fglrx from livna?

i have installed fglrx video driver

F7_T60 did you reboot after installing the kmod

yes, i have rebooted

Hello! How do I check the verson of my video card? What's the command
I checked dmesg but didnt see anything
i grep'd ati

look in /etc/X11/xorg.conf and see what driver X is using

boot into runlevel 3 login as root then do a system-config-display –reconfig configure the adaptor and monitor
bitchkat, you are way behind

Will that show me the model?
sorry, that wasn't for me I guess

thank you, Southern
i will try that

Southern_Gentlem, not surprised. Anyways, I need to start getting to the airport so that we can kick some Canadian butt in the first women's international match of Australian Football

bitchkat, i had him paste his xorg.conf which was he relisted in his post

i'm in command line of fedora

What is the command from the terminal to open the desktop wallpaper changer?

try (as root) setenforce 0
and see if it works

FYI to anyone who might wanna know, kpersonalizer wasw moved to kdebase-extras

I've got the script to write the correct image filename to the xml file, I just need to find out what the Desktop Background Preferences calls to update the desktop

if you would use the tools that i provided you it would work automaticly

i try setenforce 0

you shouldn't be manually updating the xml

i did not see any happend

but i don't want any user interaction

is the game sauerbraten and warsow not opensource license?

that;s ok, try reconnecting

fedora core 6, vncserver running using startkde… kde comes up fine but i dont get any fonts (only those odd square boxes instead of fonts) anyone know why?

gconftool-2 doesn't require user interaction

why is not in f7?

ultimate_frisbe1, gconftool-2. i updated my random wallpaper script so it works in both kde and gnome

reconnect what StevenOdb?

to your samba

by
/etc/init.d/smb restart

what were you using before to connect to samba?

ultimate_frisbe1, http://rafb.net/p/Ste82423.html

winxp

\\192.168.100.132

ok is fc7 totally broken with nvidia?

mikerose357, works great for me

fc7 is a big steamy turd
acpi/pcmcia handling is broken completely.
i'd wait for 8 if i were you.

mot on your hardware

I just installed fc7 on a new system
installed nvidia via livna

gconftool-2 –type string –set /desktop/gnome/background/picture_filename /path/to/image.jpg

mikerose357, uname -r

x fails to start

we're talking 1-year old laptop architecture
not 8 year old junk.

thanks for your help! sorry to have gone about this the long way

ok, and, does it work now?

mikerose357, boot into runlevel 3 login as root system-config-display –reconfig

yeah
it is work now
why it is work?
i did not do anything
why it work?

2.6.21-3194

anyone know about my realvnc issue?

why you know it is working StevenOdb?

mikerose357, and rpm -q kmod-nvidia

it only wants to start at all with the nv driver

it's working now ? why the "" then?

mikerose357, the kmod-nvidia you installed is for the current kernel not the default kernel

yeah StevenOdb it is work now

it was selinux interfering for your security

yeah StevenOdb it is working now

mikerose357, yum update

what selinux mean StevenOdb?

you are wrong Southern_Gentlem

their is a way to allow home-dir access for samba, without switching selinux off

mikerose357, rpm -q kmod-nvidia

security enhanced linux

ohhh

….updates take forever…

there

Hi there, anyone about who knows much about VPN's?

hold on you may be right

not many

mikerose357, answer my question
rpm -q kmod-nvidia

I telinit 3 and then did system-config-display –reconfig. The highest resolution is 800*600…

if you're in gnome, a yellow icon will appear in the systray, informing you of the way to allow homedir access

setenforce 0 = mean?

F7_T60, config the adaptor and the monitor athen go back to general

i just went on my first adventure with yum, now i am trying to do damage control

setenforce=0 turns selinux off

ok you are right in fact it upgraded my kernel to match BUT

my goal was to install adobe flash player, but i didn't succeed

I haven't rebooted
thanks

now i want to remove what i have installed

linux_stu, what did you do

i had done yum install swfdec-mozilla.i386

I have returned to original config.

and it installed 14 things, which i assume are dependencies

so yum remove it

ohh
okei
thanks

will it remove the dependencies it installed also that no other program needs

but if we selinux off
what wrong with our linux based hosting services OS?

no, yum doesn't do that

would it be safe to remove them all myself?

linux_stu, you yum.log should have log of what it installed and you remove each package it install

i still have the list of things it installed
ok
i will do that

linux_stu, would it be safe just to leave it alone
is the question

potentially
i don't care if i break my system
i want to figure this stuff out
thanks for the url Southern_Gentlem

linux_stu, x64 or 32 bit system

well its easy to reinstall linux_stu

StevenOdb can you tell me what wrong with our linux if we selinux off?

nothing's wrong, it's just less secure

Steven_eLvIsDa, is it on a private network

i have 32 bit system

linux_stu, then that link will work fine

How do I tell my wireless card to quit automatically connecting to the AP

heheheh Southern_Gentlem sometime we do it on the our vps or dedicated server

kc8hfi, tell it not to start on boot?

is there anyway to do without selinux off?

i removed them all successfully!

there is but I forgot the command, and I don't seem to find it with google

yes you have to edit your selinix policies
system-config-selinux

that works, but not for the purpose I need it now though. I had it working before, just can't remember. Its connecting to the strongest open AP…

Steven_eLvIsDa, you didnt answer my question is the network its on a private jvm java server hosting network
kc8hfi, mine connects to the strongest ap on the network i have used before
but of course i use netwrokmanger as well so its easy to switch

yeah Southern_Gentlem it is on LAN
ohhh well thanks Southern_Gentlem
but system-config-selinux command in GUI

ok, you can do it with de gui tool Southern_Gentlem suggested

can i see it in command line?

or run as root: setsebool -P samba_enable_home_dirs=1
after that: setenforcing 1
and try again

okei

Steven_eLvIsDa, does it have a public or private ip

yeah StevenOdb it still work
you mean my IP right?

ok great

the buy dedicated ip of the samba box

is LAN IP
it is LAN IP Southern_Gentlem

which means what
is it a public addressable ip

it start with… 192. ?

or 10.

10. ? or something else?

192
it is class C IP

thats is a private ip

cannot get exclusive lock on /var/lib/rpm/Packages" when i try to do step 4 here: http://fedorasolved.org/browser-solutions/flash/

use for local

running as root?

yes StevenOdb

linux_stu, try going to su - first

setsebool -P samba_enable_home_dirs=1
what does it mean?

what do you think it means

it tells selinux to allow samba to access home directories

command not found

"command not found?" A: Run the command AS ROOT after switching users with "su –login" or "su -" in order to get root's command PATH so the command WILL be found; don't run "su" by itself. If it's still not found, then either it's not installed in a standard path location (echo $PATH), is not executable (chmod a+x), is in your current dir (./runme), or you've spelled it incorrectly.

linux_stu, read zcats post

ohhh well
thanks StevenOdb

i did
i still have the problem when i do su –login or su -

do you have yum running somewhere else
linux_stu, uname -r

2.6.20-2925.9.fc7xen

service yum-updatesd status

running another yum updte or rpm install/upgrade process?

not that i know of

StevenOdb, yum-updatesd

yum-updatesd (pid 13892) is running…
ah hah

service yum-updatesd stop

is that a service that starts automatically?

yep

ok
thanks all

can someone help me with a net install? I'm trying to install f7 via ftp, and it bombs out saying it can't retrieve the "images/minstg2.img" file

it's the backend for the update popup tool in gnome

oh no
it still didn't work

I've checked, and this file exists

ps axu | grep [r]pm

no return

ps axu | grep [y]um

no return

rpm -qa ?

rpm -qa gave a huge output

wget http://macromedia.mplug.org/FEDORA-GPG-KEY
rpm –import FEDORA-GPG-KEY

cannot get exclusive lock on /var/lib/rpm/Packages

disingenuous
weird

frozty_sa, you probably have your path wrong

linux_stu, exit the terminal and restart it
su - (notice the minus)

I've ftp'd into the server host from another pc and I definitely have it right

yeah

(which stands for –login)

it's basic..ftp://192.168.0.4/f7

i did that
just now
but it is also what i had done before

still doesn't work..?

same error?

frozty_sa, (or anaconda couldn't setup networking)

nope, doesn't work, same error

telinit 1 , telinit #previous

reboot

and after tha a reboot

is there a way I can check? this is I think still before anaconda loads….

that error is usuall not root as su -

even though, in this case.. it should even work without -

according to my dhcpd leases information, it's got an ip…

anyway, got to go

i did telinit 1
and then i tried telinit #previous, but #previous wasn't valid, so i did telinit 5, thinking that was probably the right number

does anyone know the command to edit the gconf junk in the most recent gnome?

linux_stu:

now you said to reboot?

if nothing else works…

ok

try kicking it

frozty_sa, what publish your images online are you booting from?

lol
i tried it, it didn't work

gl, bye

same error
bye

the boot.iso from the f7 dvd

hm.. could it be that wget downloads the first html page more than once?
(eg to check for the links)

hi everyone

frozty_sa, so what's in your f7 dir?

i seem to have a problem

the entire f7 dvd
that dir is a symlink to the dvd's mount

running fedora 7 , compiled my own kernel (like i have done always without problems) , now my samsung ide hdd isn't recognized(new hdd) , anyone know which modules i def need?

frozty_sa, you're probably missing something simple. use http instead, and watch the access_log

could you step me through the apache config?

frozty_sa, sorry :/

darn nvidia is still broken

k. I'll try to get it up

heh

what's viagra gonna help me?
btw, hello opsec

Does anyone else have trouble playing vorbis files with gstreamer lately?

hello, froz

kernel is 2.6.22.1-41 kmod nvidia is for 2.6.22.1-41

mikerose357, what video card

Is the CFS schedular in 2.6.23 or is it already in the Fedora kernels?

ah hell this is probably an old nvidia maybe I need legacy

what video card
check the hcl link on that howto

I don't know some nvidia

lspci

maybe geforce 2

stop guessing and check

"instead of giving you what you asked for, i'll just make vague generalities" -yours truly mikerose357

Hi, is there a way to . . .check my network card HOST ID on Linux ?

host id?

I didn't buy or put together this computer no need to be insulting

http://rafb.net/p/8yf8fZ24.html

Like. . .each network has a special number which identifies it e.g. 00004dac7e . . .etc.

ma caddress?

i wasnt being insulting nor was i the one cursing

Ethernet address.

chewed-on, each card has a different mac address

ifconfig\
ifconfig

Is there a way to find out the Eternet address ?

whatever, thanks for the help Southern_Gentlem , no thanks for the lip opsec

ifconfig is tha ta command ?

yes

/sbin/ifconfig

yeah I'm reading it now

chewed-on: yes, ifconfig will show it to you as "HWAddr"

thanks

how is it helpful to complain generally, then when asked for information and told how to get it.. you don't provide it.. is that fair to Southern_Gentlem or anyone else for that matter?

frozty_sa, 403 = permission denied

Oh, that's not the address I was looking for. . . .

I couldn't give a precise answer right on the spot because I didn't know, however the link he provided told me what I needed to know

….any idea how do I make it work?

It's something like "0040c78133a6"

frozty_sa, you said you were using a symlink? that's probably why

7:ac

the relevant info being that nvidia dropped support for early cards in the newest series of drivers

chewed-on, what kinda network card is this… sounds like you are looking at a token-ring card

chewed-on, it is the same address in more human readable form

I'll try to serve from that root directly then

oh is it ?
it was a D-link network card [:

7… blah is one way to show the hardware address of the
the colons are optional

ARGH!! I wanna headbut apache…

if it is a dlink, then it is most likely a ethernet card

anyone??

chewed-on, if you remove the ard the mac address is usually on the card somewhere

Alrighty then.

fdlinux, try nodma=ide on your kernelline in grub

no need to compile hosting your own kernel .. libata

Hey Frozty, if I had multiple virtual domains on my apache server and I wanted to know which ones were invalid, how would I go about making a list of those virtual domains which were invalid?

can i config DNS in my LAN
by take my linux is DNS server

prestwick, depends on how you define an 'invalid' virtualhost. documentroot doesn't exist? nothing logged for X months? doesn't resolve to you? what?

I found the problem, thanks

sure, why not?

HI, does ATI drivers still broken?

juannicolas, there is one that is suppose to work

via livna or rpm?

the newest one rather

ohhh

livna

in my document did not tell me that
how many package that i need to do DNS
how many package that i need to do DNS services and DNS services yet

Steven_eLvIsDa, are you the network admin for your lan

rpm -Uvh http://rpm.livna.org/livna-release-7.rpm ?

yeah sir

you should talk to them before you suddenly put a dns server

yeah sir, but my skill to low i need to improve

on their network
notice the requirements

too*

Thanks for help guys, [:
Everything's working fine

you around?

what is bind in linux (DNS)?

Steven_eLvIsDa.. this is a nice package in fedora called 'caching-nameserver' that works well as a basic setup. It installs everything you need

I have a quick question. While trying to download some Fedora 7 ISO images, they ALWAYS download corrupt. On a laptop I have, they downloaded and the SHA1 check sum came out good. I loaded up a ftpd on my desktop and transferred it over. SHA1 checked out bad
Any ideas?

you didn't transfer in binary mode

now i can /etc/init.d/named restart
lolz
now i looking for resolv.conf

But why would a download corrupt when downloading from the website as well? And only on my Desktop. It downloaded fine on the laptop.

no idea, but no transfering in binary mode is the obvious one

Anyone have any experiences with Intel ICH8M chipset (laptops)

search angkor.com

maybe you have bad hardware on the Desktop

can i change angkor.com to other?

Ok. I'll try transferring it via FTP again.

rfu_ make sure you enable binary mode on the FTP transfer

Yeah. I set it to binary mode.
ben@Tizzle:~/Fedora-7-x86_64$ sha1sum -c SHA1SUM
F-7-x86_64-DVD.iso: FAILED
F-7-x86_64-rescuecd.iso: OK
That's how it always happens
The CD is ok, but the DVD file fails.

how big is the dvd.iso file? what os/host are you running? Perhaps you are hitting the 2GB limit?

ubuntu. 3.3gb

you have disk space available? did the ftp complete? or error out?

FTP completed.
Size on disk is exactly the same on both the laptop and on the desktop

and you got the file from the same mirror for both?

Yeah.
The laptop has a good copy, but no DVD burner. The desktop when downloading from the web is corrupt, and when receiving via FTP is corrupt.

can you transfer from the laptop to the desktop?

Thats what I was doing.
FTP is running on the desktop, and I FTPd it from the laptop.

when my fedora is start up
i saw one services is fial
i saw one services is fail
it is smartd
what is that?

odd. Not sure what could be going on. ;(
smartd is a daemon that monitors the SMART info on your hard drives. It can report when they are going bad or having errors or whatever. Not all drives are supported however.
can you try transfering it via ssh or rsync?

never used either.
I suppose I could learn. Haha.

learning is always good. rsync is very easy to use and handy. see the man page…

rfu_, keep the 'bad' dvd iso there (so it keeps using the same bad sectors), and ftp it AGAIN to another dir. could be that your desktop drive is going bad.

i am getting squares instead of fonts i should normally get… i am using fedora core 6 … anyone know how to fix this?
this would be for KDE desktop

I'll try that.
On one hand I hope you're right zcat. And on the other, bigger hand I hope you're wrong. Haha.

rfu join the torrent and point it at your dvd location and the torrent will repair the file

if i wanted to install gnome on a fedora box, what would i type?
yum install
yum install gnome ?
that doesnt appear to work

yum grouplist

why paste?

pste the output of yum grouplist to the pastebin

oh gotcha

i really need another varable to that script

found it… gnome-session

no

wow that is obvioius … not

do as i asked please
the following is a test please ignore

word brother

I think you may be right. Leaving the old ISO there, and re-FTPing it worked.

rfu_, ruh-roh

Which really sucks pretty bad. The drive is only around 7-8 months young!!!

rfu- bad packets happen

3-4 times on a LAN FTP Transfer?

yep they can
i would start looking for nicks in cables

there will be a big pile of 0's and 1's where the nicks are

ahaha

yes
cool

Southern_Gentlem, tcp corrects for bogus packets

if it can
not always
sometimes it just drops it

yeah

would seem unlikely that it would fail transfer several times, then work.

howdy all

tcameron, that my line

Southern_Gentlem - heh
y'all are not the only one from the South, sir

you might check /var/log/messages for any disk or network errors… also look at '/sbin/ifconfig' and see if there are errors or retransmits listed.
you might check /var/log/messages for any disk or network errors… also look at '/sbin/ifconfig' and see if there are errors or retransmits listed.

so I have a mirroring question for you

Shalom everyone ! I started getting these msgs (http://pastebin.com/m2f1776d3) and the system is far from stable. I've suspected the disk (Segate 20G) so moved to a new disk, but get the same error msgs. Help would be appreciated.

I have rsync'ed the trees for FC for ever, but now with the new directory structure changes I have been hesitant
I am up to date on FC5 and 6, 32 and 64 bit
also extras for those two distros and those two arches

might be cables? did you re-use the same cable? also might be the disk controller on your MB.

why are you mirroring EOL releases

how much bigger is the F7 Everything repo?
I will be dropping the FC5 stuff soon

nirik, Funny you ask, but yes, I've used another (new) cable with the new disk.

tcameron, figure core + extras

nirik, I've also moved the disk from the Primary to the secondary, but same msgs.

tcameron, figure core + extras= everything

FC6 I keep around cause the extras repo has a lot of yummy stuff that Just Works(TM) in RHEL 5

Everything is 56G

Southern_Gentlem - so do I need to mirror *only* Everything?

for all arches

Eliran_Itzhak, are you using a 40 wire IDE cable or an 80 wire…

tcameron, i would mirror everything and updates

in other words

http://pastebin.ca/644501 has the sizes of things under Everything.

http://mirrors.kernel.org/fedora/releases/7/Fedora/ and also http://mirrors.kernel.org/fedora/releases/7/Everything/
do I want both of those, or only one or the other?

RX packets:6078671 errors:0 dropped:0 overruns:0 frame:0
TX packets:3496055 errors:0 dropped:0 overruns:0 carrier:0

Fedora is a subset of Everything.

kdekorte, Got me there. All I know is IDE cable, didn't know there are 2 types. anyway, the system started sending me these msgs without any provokation (so to speak) its a server that worked fine until this morning.

ah, OK

fedora is just was released on the dvd

I'd have to go so far as to say that I got no bad packets…

we ain't got no stinkin bad packets!!

rfu_, so check memory or harddrives

Eliran_Itzhak, the 80 wire is needed for ATA/66 mode. what is looks like is the driver is trying to find the fastest speed and then dropping back if it can't do it

thanks for the info, all, much appreciated.

I suspect the message is harmless, unless you are getting it hundreds of times

Alright. I'll be back in a bit.

80 pin EIDE cable has a blue end

kdekorte, Ok about the cable, I can try another one. I had no problems with this server until today.

Eliran_Itzhak, you may also need to tell Fedora to use the older IDE driver. I'm sure someone here has a pointer on how to do that
Did you upgrade your kernel?

kdekorte, by the "old" driver you mean the one that calls devices "hda" and NOT "sda" ?
kdekorte, The one from Fc6 ?

correct the older driver is the one that uses hdX over sdX

I don't think there is a way to go back to the old ide drivers without an old fc6 kernel… but I could be mistaken.

nirik, or custom kernel

can't you blacklist 'libata' and it will use the older driver?

kdekorte, dont think so
how is it going to boot in the first place
you are talking alot of things to change to even boot

ok, so another question about mirroring the repos
if I explode the DVD isos, is rsync smart enough to not overwrite the whole damned tree?

Fedora 6 loads and starts the install process on my laptop fine, however under Fedora 7, it doesn't read my DVD drive (Intel ICH8M chipset), is there a way around this?

if you put the files in the right place it should work…

k

jnanney , you may not have a thinkpad, but some of these note may apply http://www.thinkwiki.org/wiki/Problems_with_SATA_and_Linux

jnanney - what is the error you see?

Unable to find any devices of the type neede for this installation type Would you like to manually se;ect upir driver or use a driver disk?
the installation type I chose was local cdrom

did you try all-generic-ide irqpoll
at the install command line?
when the installer starts, you type:
linux all-generic-ide
or
linux all-generic-ide irqpoll

no, wasn't aware of that, lemme try this now

you might try 'noprobe' on the command line and then manually select the ata_piix driver and see if that has any luck.

k
jnanney - Google is your friend… http://www.google.com/search?hl=en&q=Unable+to+find+any+devices+of+the+type+neede+for+this+installation+fedora+7&btnG=Google+Search

I have a redhat 9 manual/book…would it be likely I coudl put some of the stuff from it into use with fedora 7?

crap, rsyncing all the packages even though I exploded them off the CD
dammit

some of the general concepts might be the same… specific commands are unlikely to match up.

Okay, so after I backup the information that I want to keep from my NTFS drives, what's the best way to completely go from Vista to Fedora 7 without hiccups?

ah, I think I figured out why
I did cp -ru from the loopback-mounted ISO
instead of cp -a

zcat you around today?

3 of my fc6 systems fails to update via yum, anyone know how to fix it ?

did you "yum clean all"

nope, why?
would i need to..

how would we know how to fix it
you've provided us with no information other than "it doesn't work"

cargohold - a lot of times you can resolve weird yum stuff by cleaning the yum cache
since you've given no real info, I just figured you would like the sledge-hammer approach to repairing it

http://mange.dynalias.org/linux/misc/bugs_yum/Yum_Looptrix.txt
i think that might just me it.. thx, will test it iaw

no go on eitherall-generic-ide or the noprobe with ata-piix driver (or the ahci driver)

hrm

7, or a fresh install of 7. any advice on why I should do one or the other ? Will that many sequential in place upgrades take longer than a fresh install anyway ?

jnanney - do you have any other fedora linux boxes around you can use as an NFS server?

anyone really familliar with mdadm and software RAID?

CoffeeIV_ - fresh install is almost always preferred method
GreenJelly_Linux - sure, what's up?

I have a large post in paste bin at http://rafb.net/p/s6C3dr12.html with details and questions, which is a RAID based issue!

can you paste the yum run you are seeing the problem with? it's hard to say what the issue is without more info

ill test it asap and let the channel know how it went. I upgraded to fc7 on those 3 systems and the problem was solved though.

jnanney - try linux askmethod nodmraid all-ide-generic

didn't you get your raid working the other day?

CoffeeIV_, fresh install after backing up would be best

how can i provide nice debug for you ?

yeah, I'll load up my local server with Core 7 and try using http method

CoffeeIV_, take notes on customizations you make to the system

nirik yep… GOt the data, just a few little details I noticed on close expection of the array!….

cargohold, on those fc6 systems disable yum-updatesd service and do a yum clean all

run yum with say -d 5 and pastbin the results?

GreenJelly_Linux, so you didn't partition sde. you don't HAVE to,b ut it's a good idea.

tcameron - I gather that from web searching — does a yum upgrade take hours of downloads ? I have DSL, but I am thinking it may be faster to download FC7 cds and go from there

i have that disabled by default, testing it now

can you add 'cat /proc/mdstat' ?

nirik its clean..

ok, was looking for a –verbose or something… brb

CoffeeIV_ - I would go the DVD route (there are no CDs for F7, only DVDs)

nirik I mean nothings going on… the rebuild was done yestereday

tcameron - then I would have to buy a DVD burner

zcat, your saying fdisk sde?

CoffeeIV_ - not necessarily

zcat as Linux Auto Detect RAID

do you have a Linux box already up and running on your network?
you could use it as an install server

looks like http method next, askmethod nodmraid all-ide-generic failed too

jnanney - what kind of motherboard?

GreenJelly_Linux, you don't need to. and if you don't have a spare to sync to you'd be taking a chance while in degraded mode

[raid6] [raid5] [raid4]
active raid5 sdb1[0] sde[3] sdd1[2] sdc1[1]
1465151232 blocks level 5, 256k chunk, algorithm 2 [4/4] [UUUU]

it's a dell laptop intel board

paste bin?

jnanney - laptop model?

zcat what do you mean degraded mode?

Is there a way to make CDs for FC7 ? most of my servers have only CDROMS, not DVDROMS

Any good guides on how ot convert from Vista to F7?

GreenJelly_Linux, degrades as in running on 3 out of 4 drives in a raid5 set
degraded*

format vista partition, install fedora

bpat1434, what does that even mean?\

CoffeeIV_ - you can use the boot.iso file from (for instance) http://mirrors.kernel.org/fedora/releases/7/Fedora/i386/os/images/ and do a network install
it is easy

zcat did you look at my pastebin, cause I have a feeling Im either using RAID-6 or that I am in degraded mode as it is

bpat1434 - what do you mean?

are there any good guides to help me along the way from Vista to Fedora 7 so I don't have any hiccups.
see above

are you looking for a table of apps on Linux which match up with what you use on Vista, or are you asking how to keep Vista but also use Fedora or what?

tcameron — thanks, I will probably do that

bpat1434 - you probably want to look at http://docs.fedoraproject.org/

How to properly get completely rid of Vista

there is a ton of good documentation there

format your vista partition

what you are asking is waaaaaaaaaaaaaaaaaay too broad to be able to give you a real answer

GreenJelly_Linux, your paste SAYS raid5. and the total size will be the smallest drive * (Numberofdrives-1)

install linux

clean install

http://catb.org/~esr/faqs/smart-questions.html

naturally back up any data you want to save like images of the dogs/kids, and your music collection, + anything else

yes yes, I know

zcat then I am not using sde
zcat for each drive is 500mb

my question was only because of yesterdays issues with GRUB not booting

zcat and I have 1000mb

did you make your partition when you didn't have sde in there?

brb

bpat1434 - would love to help but I have *no* idea what you are talking about

Array Size : 1465151232 (1397.28 GiB 1500.31 GB)

nirik no, but sde was damaged…

the array is the full size, but the filesystem on that /dev/md3 is the smaller size. You need to resize it, or reinstall it

GreenJelly_Linux, 500*(4-1) = 1500

GreenJelly_Linux - what did you do to create your partition? Did you use fdisk or parted?

zcat zcat df-h states "/dev/md3 903G 525G 331G 62% /home/Media"

GreenJelly_Linux - if you used fdisk you probably did not get the full partition size you could have
fdisk seems to choke on really large partitions
wait, never mind

GreenJelly_Linux, umount /home/Media; resize2fs /dev/md3; mount -a

tcameron this is a recovery of an array… I never re-fdisk the messed up drive…

right, just read that, hence my "wait never mind" statement.

zcat should I fdisk sde first?

GreenJelly_Linux, like i said, running in degraded mode is risky. lost one more drive and the array is gone.

I wound up blacklisting the iwl drivers, and installed the ipw ones for the 3945 card. it works perfectly now

GreenJelly_Linux, so, just leave it

I wanted to locate the problem so i skipped "yum clean all". Suprisingly its working now (Atleast downloading the packages)

k

glad you got it working, sorry you had to go back to the ipw… hopefully iwl will get working for you at some point.
might be you were hitting a broken mirror?

yeah, its all good though, its gonna be awesome when the iwl drivers get perfected

I got some mirrors that where unavailable but not all

ahh. rawhide and things actually started working
7 is like 3. should've skipped it

zcat I guess I dont understand why the df -h doesnt agree with a raid 5 array with 4 devices (500mb) and why sde is listed instead of sde1…. I can run that unmount, if that will fix the problem
zcat but ultimately I want 1500mb

I had some issues with yum before and the "db" where you had to set KERNEL_VERSION_DB= or something (you know what i mean) but this time it just seemed to be looping endlessly taking up all CPU power (could hardly move the mouse)
Maybe its a bug in yum when no mirrors can be used ?

odd. If it happens again let us assist you in gathering data for a bugreport.

i will let you know

okay… so zeroing the drives should be good enough right?

your /dev/md3 filesystem is not using all the size of your raid… see what zcat said above on how to resize it.

bpat1434 - what do you mean "zeroing the drives?" Using dd to overwrite the first couple blocks?
you can do that
or use badblocks -vfw
that will do it too

_llseek(11, 7503872, [7503872], SEEK_SET) = 0 (Means seek beginning of something, it did this over and over)

zcat I will be expanding this drive every few months or so… Therefore, I am prepared of the risks if fdisk sde as raid array, then rebuild… I feel I am probably in a degraded mode as it is, due to the data… but I am still learning how to interpret all of these reports… I have gotten this far… in addition, a mdadm -E /dev/md3 returns mdadm: No md superblock detected on /dev/md3.

no, I mean the drive zeroer from Western Digital (Data Lifeguard Tools)

nirik ok… will do…

"Write Zeros to Drive"

sorry had phone call and was distracted, laptop model: Dell Vostro 1500

no worries

you could remove sde and make a sde1 and add that, but if sde1 is the entire drive anyhow, it won't matter.

is newest line from dell, so I know I am going into uncharted territories
it's just odd that fc6 installer loads fine, but f7 installer doesn't

really?
hrm

I got the http load going, but will only do text based
fc6 goes graphical

well, that's what you want anyway
text based is ~ 30-40% faster
than GUI

GreenJelly_Linux, if you're going to be growing/reshaping the array every couple months you'll need do that resize2fs each time after.

jnanney - looking at the specs, wait 1

I cant unmount because its "online"

okies

can I do a online rebuild?

I tried to go: yum -y update on 3 different computers but all it did was read primary then consume all resources (On fc7 its fine now though)

GreenJelly_Linux, mdmadm /dev/md3 -f /dev/sde -r /dev/sde

I wont be home for a few days

updating the same computer now seems to be ok

odd. oh well.

GreenJelly_Linux, that fails then removes the drive from the array. you would do the same for a failing/failed a drive. after you partition it, add it back with -a /dev/sde1
GreenJelly_Linux, then cross your fingers and hope none the drives die before the resync completes

jnanney - and you say that FC 6 loads just fine?

the problem after getting in text based is with partitioning, runs into error with partitions, I may end up scrapping windows to try and get past that problem
yes, fc6 loads graphical installer just fine

well, does Windows own the whole disk?
and is it a dynamic disk?
cause if so, you're hosed

I used gparted to resize the windows partition to 1/2 the drive size

ah
do you know if Windows thinks it's a "dynamic disk?"
you can view it in disk administration
inside Windows

but fc6 installer and f7 installer choke on the partition
lemme check

boot into 'doze and see if it is dynamic
if it is you're hosed

why?

jnanney, because you need to reverse the polarity of /dev/null

http://rafb.net/p/sWEHgN30.html

GreenJelly_Linux, yeah? you're in degraded mode now. hurry up.

hehe… definately a fdisk job I guess…

Something makes it loop endlessly (thats for sure). Hmm, maybe if i install a new fc6 and test it again but ive gotta deal with a "local bug" in an app

jnanney - when Windows marks a partition as "dynamic" it is really just changing the partition table so that no other OS can play nice with it

I am assuming zcat is a bot

jnanney, tell me about your mother.

help

tcameron, index out of range

LOL

lots of dum dum pills being taken again i see
good idea to login to irc as root

ok, off the subject question. anybody here hafta share an office with someone else?

jnanney - so anyway, if the installer is barfing on partitioning, it might be that Windows has spewed it's proprietary shit on the partition table

there must have been a sale somewhere
language please

probably

that's a pretty verbose and highly technical explanation you have there

I'm just going to have to wipe windows to be sure

k

Command Not Found

"command not found?" A: Run the command AS ROOT after switching users with "su –login" or "su -" in order to get root's command PATH so the command WILL be found; don't run "su" by itself. If it's still not found, then either it's not installed in a standard path location (echo $PATH), is not executable (chmod a+x), is in your current dir (./runme), or you've spelled it incorrectly.
ima gunna blacklist some people

I noticed another thing on one computer at work. If i had fc6 installed on it wintendoh xp installer refused to work. My guess is it tried to check installed oses and failed miserably (Gotta love it

lol

why?

opsec, because you need to reverse the polarity of /dev/null

cat it to urandom

the windows installer cannot do anything with the HD if it has any linux partitions

or wise versa

yeah, the xp installer is not very smart. It will also let you install to a external usb drive, and then fail to boot because it didn't install usb drivers.

therefore, the installer hangs up right aftr it says setup is detecting hardware

trust me it is more fun to just abuse zcat

jnanney - now that opsec is /ignored, moving on

hehehehehe

there's a simple solution for anyone having windows issues .. **don't use it**

me resists the temptation to abuse frozty_sa …

amen

Bwahaha, totally awsomely coded that pice of junk eh ?

I agree to what opsec said too

Booh!

i wasn't kidding.

abuse? /me acts like a windows user and installs an antivirus out of fear…

^^ ti's why I'm switching

gonna try to find an easy way to copy the partition off, cause another 12 hours reinstalling windows, then wipe the partition and try installing

oh wait….I run linux, no AV necessary!! :-D

why?

why people insist on using outdated software that interferes with your right to use your computer as you see fit is beyond me

some of us don't have that luxury opsec

We wouldnt be in this channel if we liked wintendohmofones (well some but they arent very smart

slavery, it's not just for africa anymore

haha

zcat, now this is probably not the smartest thing in the world to do, but can I use that drive as it rebuilds…

nice lines there

I know you weren't kidding. I really agree with what you said

GreenJelly_Linux, you can use the array as it rebuilds, sure
GreenJelly_Linux, would be kinda dumb to be stuck offline during rebuilds

windows is simply the modern digital equivalent to the slave ships of old

You can thank the US government for that one…. stupid billion dollar endorsement

employers that require windows is my biggest problem. I use a vmware windows install for most of my work, but some hardware doesn't translate through vmware well enough

I guess Im going to have to change my mdadm.conf… to be sde1 instead of sde….

jnanney - does the Linux installer see the disk as one big blob with no partitions?
or what?

it sees all the partitions, but says not enough free space

GreenJelly_Linux, your mdadm.conf should be using uuid's, not partitions

hrm, then that *probably* means that Windows has not "converted" the disk to dynamic

there is a 14MB free space blob before the ntfs partition and a 50GB Free Space after

so did you create a Linux partition in the free space?

it creates the swap partition fine, but chokes on creating an ext3 partition
saying not enough free space
with fc6 installer and with f7 installer

are you defining a specific size or saying to use all remaining space?

zcat cool!

tried both

The Debian people found a bug in one of my apps a month ago (local, lowrisk) still, I as in upstream wasnt informed!!! (Not a single email). This really bugs me right now *grr*

Comments

hi I have a friend using gentoo and he cant change the root password using the init=-bin-bash argument to the

P

AllenJB, hold on..
AllenJB, 4×1000 kbit =~ 576 kb..
kb=kilobyte
1 byte= 8bit

you meand download would be 576Kbps jsbach ?

Oliver, jah

jsbach, actually its kB = kilobyte

err - no - dialup runs at 33600bit/s
4000bits / 8 = 500 BYTES

AllenJB, well therefore put a =~

approximate? lol - what's your error range?

AllenJB, wide enough to give quick info's

lol
*jsbach

yo

anyway… back on topic

anyone knows, if its possible, that the xfs implementation in kernel 2.6.22-r5 got a odd bug ?
for sime reason it shuts itdelf down for internal data coruption on regual basis since i updated the kernel
couldnt figure the reason

sb 10

You might want to try #gentoo-kernel for that question

has anyone gotten a gentoo minimal install cd to work behind a nat'd firewall box ? my firewall keeps droping the packets and i cant figure out why

What specific task are you trying to do when it's "dropping packets"?

AllenJB, im trying to ping google, tried by hostname and address

I've done all my Gentoo installs from behind NAT, so NAT on its own certainly isn't a problem
It sounds like your network setup isn't correct.

AllenJB, you wouldnt happen to have the iptables rules that enable your NAT handy would you ?
AllenJB, well, i would agree, but, ive tried various NAT rules from various websites, and my old NAT rules which worked fine on my old fedora install, and no matter what i try it drops the packets

I needed no rules. The only thing you would need rules for is if you were SSH'ing through the NAT (from outside) to the box you're installing on

not rules for the gentoo install client box, rules for the firewall box … i know NAT doesnt just work automatically without some commands to enable it
the firewall box is a 3 day old gentoo install, with kernel 2.6.22-something
i have "[*] FULL NAT" enabled in the kernel
i have MASQUERADE loaded as a module
i have "1" echoed to /proc/sys/net/ipv4/ip_forward

What iptables rules do you currently have?

i have, input/output enabled to localhost, which shouldnt affect anything
then i have
/sbin/iptables -A INPUT -i eth1 -m state –state ESTABLISHED,RELATED -j ACCEPT
/sbin/iptables -A OUTPUT -o eth1 -j ACCEPT
to let my firewall talk to the outside world
eth1 being the external interface
which works
then i have:
/sbin/iptables -A INPUT -p all -s 192.168.2.0/24 -i eth0 -j ACCEPT
/sbin/iptables -A OUTPUT -p all -d 192.168.2.0/24 -o eth0 -j ACCEPT
to let my client box talk to my firewall … using ETH0 internal interface, which works
and then i have:
/sbin/iptables -A POSTROUTING -t nat -o eth1 -j MASQUERADE
/sbin/iptables -A FORWARD -i eth0 -j ACCEPT
/sbin/iptables -A FORWARD -o eth1 -j ACCEPT
did you get that ?

Don't know how much of what you posted that I got

3 lines

teach headmonkey about pastebin

A pastebin is a web site where you can paste many text lines or sometimes files, to avoid flooding a channel. Try to avoid pastebin.com as it is very slow, consider using rafb.net/paste instead. If you don't have a browser you can "emerge nopaste" or "emerge wgetpaste" to send text to rafb.net/paste from the command line.

Yes, got 3 lines

i know about pastebin
but noone else is talking, and i didnt figure 3 lines of paste would kick me
except perhaps its cumulative
anyway, thats it

nice monologue

fragalot, thanks

http://forums.gentoo.org/viewtopic-t-578248.html

eth1 is the Internet (red) or LAN (green) interface?

AllenJB, eth1 is the internet

Hello

which version of gcc did you upgrade from?

AllenJB, http://rafb.net/p/HIHxVG28.html — the entire ruleset

sheldonh, it was from gcc 3.3.6

so you rebuilt the toolchain, then did a full emerge -e world?

sheldonh, yes i've done it

Have you checked /var/log/messages or "dmesg" for error messages?

the revdep-rebuild approach is not safe when going from 3.3.x to 4.x

still here?

gentoo abba gcc upgrading

http://www.gentoo.org/doc/en/gcc-upgrading.xml

1. update gcc, 2 rebuild toolchain and then emerge -e world

That document is the official upgrade path from gcc 3 to gcc 4

hmm, well that was my long shot, sorry

http://www.gentoo.org/doc/en/gcc-upgrading.xml

Hmm. The livecd has some dbus problems.

that's the one he's used

sheldonh, u re welcome !!!

AllenJB, no, i havent, i wasnt aware they showed up there, let me look

should I compile gcc with openmp on a celeron d ? any benefits on that ?

everything but printing is running good

is there any thing special for making the sound work on a new imac (and i don't find any pages on the imac)

that's strange

AllenJB, no, nothing, just some link is up, link is down messages from when i was messing with ifconfig

] -A POSTROUTING -o eth1 -j

AllenJB, and thats all your firewall box needs to let the client through ?

sorry, POSTROUTING table
Yes, and no rules at all on the client

AllenJB, can you paste the complete rule, i'll try it

http://rafb.net/p/KqSOqB98.html is the entire *nat section from my iptables. There's nothing relevent to nat in any of the other sections

AllenJB, is that an iptables command ?

what does emerge -X flag do?
install the latest version?
it's not in man

It's an iptables script yes (it's actually from the save/restore file) but you can convert it to a command easily

AllenJB, maybe you can … i cant

Where did you see it? As far as I know, there is no such option for emerge

t35t0r I get a 'no such option' .. so nothing?

probably leftover from the revdep-rebuild for some reaon
ahh right that new revdep-rebuild changed the -X option to something else or doesn't have it anymore
revdep-rebuild passes any flags it doesn't recognize to emerge

If you're using the ~arch _pre6 version it's been completely rewritten and the -X option no longer exists (it does that by default and there's another option to use specific versions) - any unrecognized options get passed through to emerge

AllenJB, it doesn't do -X by default
AllenJB, it goes for the specific version

Err, it did when I did the expat upgrade

hi! "After my last emerge –update –deep –world" my system is completely hosed and I'm helpless, can somebody help me?

AllenJB, well it's still trying to compile hosting spefici vercis

did you remember the fix_libtool_files.sh 3.3.6 step? also, i didn't ask and you didn't say whether you rebuilt libtool before emerge -e system

sheldonh, yes I did

looks like you are discussing something similar

And why would it have the following option if it did it by default: -e, –exact Emerge based on exact package version

AllenJB, whenever you get a second, if you can provide that rule in an iptables command line format, id appreciate it, since i dont know how to interpret it

oh well. as a long shot, you could try emerge -e cups, but that really shouldn't make a difference
oh, and you did emerge –oneshot sys-libs/libstdc++-v3 ?

Sorry, I'm not going to read the iptables man page for you. You can do that yourself.

AllenJB, but i wont know if i interpreted it correctly if it doesnt work, will I ? fuck

sheldonh, I did emerge sys-libs/libstdc++-v3

Language please

AllenJB, thanks for the help

i _highly_ recommend firehol for iptables newcomers

sheldonh, that would just add more complexity to a script that used to work but now doesnt since switching to gentoo

so much so that we use it on corporate firewalls, where ruleset readability is more important than a 10% optimization
*shrug* suit yourself

sheldonh, im simply trying to find out why my packets are getting dropped
sheldonh, im not wanting to install firewall management software, its not going to tell me why my packets are getting dropped

you need your drop rules to be preceeded with LOG or ULOG rules

AllenJB tried to shed some light on the situation, then gave me some shit i dont understand at the end
which isnt helping any

sheldonh, oups I didn't such a thing

aha! that's good news. any mistake you made is something you can try

sheldonh, I'll try to emerge it so

language

i really don't see why irc people are so tight about swearing. tech people understand swearing, it's an outlet in crisis times. and as for kids, have you heard what they're listening to these days, grandma?

All you have to do is read the man page to find out what the parameter to specify the nat table is. I don't know about anyone else, but I'm not here to read man pages for people. If you don't understand the iptables format, then maybe that's half your problem and you need to do some more reading.

agreed, but those are the rules here

give #iptables a try if you haven't, but logging is a good first start
bgive #iptables a try if you haven't, but logging is a good first start/b

AllenJB, you said "i could convert it easily" … and i pointed out "you can convert it easily…and i cant" … now your saying you cant convert it without doing research first …. make consistant statements and you wont confuse people
Fieldy, i tried in there, noones talking, how do i log iptables rule actions ?

I personally don't consider reading a man page hard.

i think LOG is a target, and you just have ot make sure the rule passes through it before it's dropped or any other actions. else the packet won't get to LOG

AllenJB, reading a man page isnt hard, ive read the iptables man page, and by everything it says, my rules should work, but they dont, so reading it again isnt gonna help, and converting formats only serves to complicate the problem more

And I was making an assumption that since you insist on messing with iptables at a low level without helper tools like firehol, you understood what it's doing.

AllenJB, if you cant convert it, fine, dont tell me how easy it is if you cant do it yourself

"tell me how to use this complicated thing without understanding how it works, _and_ without using the tools designed to help me do so"

guys if you can't get along just please don't talk to each other, or ignore each other

Fieldy, im down with that … do you know how i can enable logging of rule actions, or can you point me to something that explains it ?

AllenJB, i see what I see and it's doing an emerge of exact versions

Must be me then, but it definitely upgraded every package for which a more recent version was available on my system

and it's a version that doesn't have -X

(on 2 of my systems, in fact)

nope i forget

Fieldy, thanks

all i know is it involves the LOG target, thtt's it

Fieldy, i messed with ulogd and the log target earlier, couldnt get it to work either

-j LOG ?

does python-updater have an option to upgradE?

Vinky, what does that do ?

-i

http://gentoo-wiki.com/HOWTO_Iptables_and_stateful_firewalls#Logging_techniques

Ehnvis, thanks man, i'll check it out

wtf is it with /etc/cups.orig/ and cups complaining about not finding /etc/cups/ ?
bad ebuild?

this expat really messed things up
all for a stupid xml lib
xml is retarded

libxml2 \o/

t35t0r, ln -s /usr/lib/libexpat.so.1.5.2 /usr/lib/libexpat.so.0
and everything will work again

MrSassyPants, that's a hack

no. that's bad advice, which I have asked you not to tell people before. so please stop.

t35t0r, it works

bad hack!

Please stop telling people to do that.

DO IT

auto-rejoin–

yep

’cause it forces ops to ban you instead

You were warned before, that one sticks for 2 weeks. Please grow up and stop advising people to do horrible hacks.

Can I do a quick straw-poll: who has the "tetex" package installed? And if yes, do you actually use it?

yes

What do you use it for?

LaTex?

yes and no

mostly as a backend for lyx

I really mean, what bloody good is it? Can I live without it?
And how can I figure out which packages are depending on it?

that depends on whether you have stuff that uses *tex for print-quality output

yes, its for documentation if I dont remember right

Well, this computer has no printer attached…

revdep-rebuild -p -i

emerge -n gentoolkit; equery depends tetex

hello, can someone help me with a strange emerge - not working problem?

where does gentoo-s iptables save the tables when you do /etc/init.d/iptables save ?

it's used, for example, to produce PDFs from texinfo files
probably in /var/lib somewhere

Hmm…

thank

use quickpkg to create a backup package of it. then remove it and see if you miss it

so … what's the correct solution to the expat problem ?

I haven't installed it yet…

i tried rebuiold-revdep -X
no luck

whenever i try to emerge a package, emerge is stopping with no error message but is not unpacking. last line is checking the and i see the "ok" but thats is

/topic

sheldonh; I'm hoping there's a iuse flag that I can disable for whatever packages are depending on it…

hi all how may I search an ebuild for cups 1.3 in th layman overlays?
must I add each overlay?

emerge -s ?
niv_one_three, yes

:-\
thnaks
did any one find an ebuild fro cups 1.3?

MicroMole, -X doesn't do anything anymore

check bugzilla first

MicroMole, i.e. if you're using a new enough revdep-rebuild

oh ya sure!

app-portage/gentoolkit-0.2.4_pre6 (/etc/revdep-rebuild)

This is a bit weird… I don't have "tetex" or "latex" in my USE flags, so according to "equery d tetex" there is no reason for it to be emerged.

thaks

hi
qmake_image_collection.cpp still segfaults while compiling qt-3.3.8-r4
anyone know why?

wrong file, correct is app-portage/gentoolkit-0.2.4_pre6 (/usr/bin/revdep-rebuild)

yes its in bugzilla - this 1.3 offer zeroconf..

maul, what flags and which gcc ?

bye bye

gcc-4.2 and only the basic cflags
I tried changing the cflags and it didn't work either

hrmm i'm still using 3.3.8-r3

12
fail

3.3.8-r3 fails too

maul, i'm using gcc 4.1.2
maul, USE="cups gif ipv6 mysql opengl -debug -doc -examples -firebird -nas -nis -odbc -postgres -sqlite -xinerama -immqt -immqt-bc"
maul, CFLAGS="-O2 -march=pentium4 -fno-strict-aliasing"
from genlop -i qt

should I buy zyxel or linksys switch ?

#hardware

I think it has nothing to do with gcc, as I found the same errors on the web with really old versions of qt and gcc too

maul, try those cflags and use flags

my use flags are pretty much the same, I'll try different cflags

amar-ze: might wanna try #networking too

for some reason, my desktop aint limited by the monitor limits, so the right side is acting like my monitor is twice the size, any ideas?

do you have two monitors in Xinerama or something?

dont know what Xinerama is, but i definatly aint got 2 monitors

try setting the resolution in your desktop manager's config

check your xorg.conf and see if you have any line saying VirtualSize

kk i will, thanks alot

hi

is tarpit patch applied to gentoo-sources?

i try install ltsp server on gentoo but when i try to start client to boot from the network i get this error on the client side PXE-T01: file not found

have a look here http://dev.gentoo.org/~dsd/genpatches/

thx Ehnvis

hi, can someone help me emerge eix?
i've been told it's good for searching portage but, i can't seem to build it

what's the error message?

http://pastebin.com/m569a5f9f
i can get the full log if you want

I dunno, did you try another version?
seems like your gcc doesn't like it

i didn't no
i figured as it was stable it should work, heh :]
i'll give 9.10 a go

I have 9.10 and it compiled with gcc-4.20

I just compiled a kernel manually instead of using genkernel, and when I boot, some devices are not available (my external hd and pen drive). But once udev-postmount is run, they show up. How do I get them available in /dev before udev-postmount?

there's a program called autounmask, if you want to unmask it easily
in /etc/conf.d/rc there is an option to save the devices on shutdown and restore them during boot

ok, thanks maul

noice, i'll check that out. thanks

it's RC_DEVICE_TARBALL

cool. alright, i'm going to reboot.

gl

thanks

i tried to follow the "expat" instructions in /topic but the ebuild fails

no luck. I got this error: http://pastebin.com/d7447c11

oh I know
udev-postmount pulls in your module
put them in /etc/modules.autoload.d/kernel-2.6
and they will be loaded sooner

paste all output from emerging a package (like say sys-apps/portage?) somewhere.

how should I apply TARPIT to gentoo-sources to make it work?

http://pastebin.com/d2fdbc8f6

have you read it ;-/

zlin its pretty simple it stops right before it normally says "unpacking"

please give us that config.log too

what do I need to emerge in order for an icon to pop up on the xfce desktop when I plug in a usb memory?

it's not modules that I need though, it's the devices that don't show up in dev. I thought udev did the /dev management?

Does anybody use Dual monitors, if so what kind of card do you guys have? i bought 2 card both missing drivers for linux could somebody let me know which card is compatible with linux kernel 2.6.x

depends on how the patch looks, give me a link to the patch page and i'll have a look.

yeah, but not udev-postmount

integreated intels have good opensource drivers

what card did you buy?

I understand how to do the patching, but actually patching disables me from compiling the kernel

and basic devices show up without udev too. like hard disks

teach MicroMole about error

To inform us about a problem, you need to give us the /fault/, not the error. Error lines that begin with "!!!" or "make" are of no use to us. Fault lines, like "foobar.o: no such file or directory" or "undefined variable cabli" are. If it is over a few lines, please use a pastebin such as http://rafb.net/paste/ (details: gentoo-detailed/error)

oh, I see. Is there anything I could do?

are you sure the module is loaded at startup?

zlin http://pastebin.com/pastebin.php?dl=d3207dbd9

a quick google would have given you that answer. http://lists.netfilter.org/pipermail/netfilter-devel/2007-August/029003.html

emerge –info ?

im using 2.6.22

I'm not sure which module you are talking about. Do you mean the ext3 driver? Because I have that built in.

sorry, I'm a bit thick
trying to get to the config.log
but the system is too hosed already

zlin http://pastebin.com/m551a006e

usb maybe.

is it an usb drive?
the one you can't reach at bootup

yes, both of them are. I will try that. thanks again!

np

so… that patch is of no use. skbuff indeed seems to be the problem. what did you google for, btw?

Adekoba, i did few stuff since i had one intel on the board .. so i went and bought ATI 9250 .. pci tried to get both working .. that didn't go well

oh man, the ATI 9250 is a bitch.

is there a seperate package for nvidia legacy drivers?

Adekoba, can you have onboard card + PCI card for dual monitor ?

I have that exact card and it literally took me a year to get it working

tarpit kernel patch

and emerge –debug portage ?

I'm not sure… I would think so.

Adekoba, thats what i thought but seems like you can not

zlin thats a lot of lines :-) one sec

k, ill take some more looks in that search. any further ideas?

Anyone have a tip on a command line client that can do ripping like grip can?

not right now.. sorry

cdparanoia, tagging, encoding (flac)

Ehnvis, i tried onboard + ATI 9250(pci) and tried ATIL 2400 Pro

and whats not working exactly?

hmm.. your {C,CXX,LD}FLAGS are weird.

zlin http://pastebin.com/m73f6cf73

ola ?

zlin yeah i know

alguem do brasil ?

Ehnvis, if i have onboard + pci card dual monitors does not work
Ehnvis, if i have pci express card + pci card dual monitors does not work

was that all of the –debug ? and why do you use weird {C,CXX,LD}FLAGS ?

can i have FEATURES sanbox and usersandbox togheter ?
i got this problem http://rafb.net/p/A9NDiu55.html with x11-libs/vte-0.16.8

zlin the flags are for icc settings and that was all of debug

nad your sure you have setup your xorg.conf correctly? and your window manager has support for dual monitors?

usbcore was already being loaded anyways, so it didn't help. I'm thinking it does have to do with udev though because it is what populates /dev

but i already turned off all icc stuff (thats why it looks strange) -

Ehnvis, i think i did .. but when i switch between video card only one worked .. for example .. i modified the xorg to reflect one card either time ..it didn't work .. only one worked ..others get disabled .. look like i need both pci card or dual supported one card

If I have a file in utf-8, how can I convert it to iso-8859-1 ?

ok. you can go back to #-portage and show them those pastes and hope they can help. it helps to say that your question really is about sys-apps/portage and not some random package.. should probably tell them a bit about what you've been doing (with icc or whatever)…

do an lsmod

zlin hehe thanks i will try that

and put all your currently loaded modules into autoload

OK, here's the config.log http://pastebin.com/d61121919

sheldonh, emerge –oneshot sys-libs/libstdc++-v3 had finished and then I'd restarted cupsd, I still can't print

ok.

can anyone help me with that vte problem ?

hope I'm not too annoying with my newbieism

also, some usb devices need 10-20 seconds to be fully usable

never mind i used FEATURES="-usersandbox" emerge -av –resume and it compiled

i cant see the problem that it shouldnt work as both onboard and pci works alone. i think i've seen such a solution before but not sure where i saw it.

That may be the case, but it was working before. Would not having an initrd have anything to do with this?

but im still interested why it did this ?

I think I've found a solution here: http://enterprise.bidmc.harvard.edu/pub/tarpit-updates/

Ehnvis, actually on the normal setup .. when i have the pci card installed onboard card automatically get disabled

ah. hopefully it works

hi, i install ltsp java server hosting and when i start client i got an error PXE-E97: NBP is too big to fit in free base memory, can anyone help me ?

compiling now, keeping my fingers crossed

Ehnvis, so i bought 2 card one with pci express and another with pci .. seems like the same story one get disable d

you should be able to set that in bios, if it should get disabled or not

Ehnvis, my bios does not have any option for video card

well id you have an initrd the modules inside it initialize a lot sooner
if*
why do you want to mount an usb drive at bootup anyway?

that sounds wierd as machines with onboard stuff usually has the possibility to turn em on and off in bios.

if nothing else, you can also put "mount /dev/sdb1" into /etc/conf.d/local.start and it should work

on the other hand, that pci express card does it have dual outputs, like one DVI and one VGA?

yes, that's true

Ehnvis, i know ..but my cards not dual cards .. i tried with 2 cards one pci express and one with pci card .. it didn't like it

you mean that the pci-e card (which i guess is a ATI 2400 according to what you said above) hasnt got dual outputs on it?

hi everyone, i'm trying to compile postgre with kerberos support. I ran emerge with the kerberos USE flag but when I run pg_config it says '–without-krb5'
does anyone have an idea why?

can i remove /usr/portage?

why?

maul, cause its on a partittion that i need for storage some important files
maul, so im thinking in remove this partition to do this

I think you can, if you don't want to use portage anymore

maul, but i cant get it again /
?

for some reason, my desktop aint limited by the monitor limits, so the right side is acting like my monitor is twice the size, any ideas?

you probably could

I thought someone answered that at 14:20ish?

maul, how ?

^^

its weird

i can umount this partition booted on gentoo?

should be a wiki on how to move /usr/portage on http://gentoo-wiki.org

by downloading a portage snapshot, like when you installed gentoo the first time. but it would be easier to just tbz /usr/portage

do stage4?

did you check your xorg.conf?

I have a wierd problem, I compiled postgresql with kerberos support but pg_config shows it compiled with –without-krb5

i dont have cdr to burn =/

what can be wrong?

well dont really know what to check for, but i tried setting the resolutions down, didnt work. and i tried looking for a VirtualSize or something but there were none

Deifyed, order the resolutions from high to low in xorg.conf

there might be an easier way, but without any free space, I think redownloading it later could work

kk ill try that, thx
didnt help

also, if you need free space, delte /usr/portage/distfiles

module is compiled succesfully

maul, i need this partition

NOOO

hi! I have a friend using gentoo and he can'

"ERROR: "secure_tcp_sequence_number" [net/ipv4/netfilter/ipt_TARPIT.ko] undefined!"

Deifyed, x restarted?

maul, think in umount this partition , mount like /testing and cp the files of portage for / partition

omfg

can't you just mount another partition and move /usr/portage there?

no i just launched gdm ^.^

maul, the other things its swap and /home =/

ouch

what about the root partition?

maul, i will cp file for there

can somebody point me to the xdm script resources ? i want to load my GUI on startup. I have already changed my inittab default runlevel to 5.

ok
i need buy a cdr

hi! I have a friend using gentoo and he can't change the root password using the init=/bin/bash argument to the bootloader … it gets the "Give root password for maintenance (Type Control-D for normal …)" … isn't it possible to reset the root password using specific arguments to the bootloader?

cya

(I am referring to single or init=/bin/bash)

boot from the livecd, chroot in, and change the password then

if you have free space on your root partition, you can just copy it there and it will work

well, I'll leave tarpits for what they are for now… thx for help, got really close

hello ?

or copy it into /home and make a symlink

anybody here to help me with xdm ?

relax, if someone knows they will help ya

what do you mean?

I know about that … but can't it be done from the bootloader?

do you want to run kde/gnome on startup?

Debian, Fedora (to name only a few) allow this

maul, yes exactly. I want to have that login window where i can choose session and all.

rc-config add xdm default

I'm not interested in password recovery, I know hot that is done … I'm only wondering why some distributions allow password recovery using bootloader arguments and others (such as gentoo) don't

and then set your dm in /etc/conf.d/xdm

maul, I already have done that. but on startup it expects some /bin/…/xdm which it does not find

which dm do you want?

edit /etc/conf.d/xdm
like, set it to gdm and then make sure you have gdm emerged

maul, i have set DISPLAYMANAGER="gdm" in my /etc/conf.d/xdm

x server restart command ?

emerge gdm

2.16.4 when tried emerge –search gdm now.

Hi. How do I disable netplug in the install cd?

and it still doesn't work? what's the exact error you get?

got compiz-fusion and avant-window-navigator running lastnight.. woo, great stuff

maul, Where can i find system bootup logs ? also one thing is i can start GUI session with startx command.

SidGod, dmesg & /var/log/messages

SidGod, do you have xdm emerged?

for X /var/log/X.*.log

SidGod, if so then /usr/bin/xdm will exist

just close your X session and run "rc" in a terminal
he wants to use gdm not xdm

is he trying to boot directly to a splash screen instead of CLI?

maul xdm starts gdm

maul, Ooops i am really making fool of myself. xdm is not emerged. sorry for all mess!

hmm where are the configure scripts getting the information about which/what linker (ld) to use?

no

Is there a way to remove a partition for raid1 with one mdadm commands? currently, I have to use 'mdadm -f /dev/md3 /dev/sdb3' and then 'mdadm –remove /dev/md3 /dev/sdb3'

gentoo's xdm is not xdm

maul, it always has been for me

xdm is like gdm, but gentoo has a service called xdm which is not xdm actually
if you can follow me

anyways thanks for help all!

maul, to start xdm you need /usr/bin/xdm which belongs to x11-apps/xdm

yes, but if you want to use gdm you don't need xdm, because /etc/init.d/xdm is not the same xdm you're talking about

wiki SidGod gdm setup

http://gentoo-wiki.com/HOWTO_Gdm_setup

for some reason, my desktop aint limited by the monitor limits, so the right side is acting like my monitor is twice the size, any ideas?

Hi, I just set up drivers for my rt2500usb wireless device using a custom ebuild, and the module loads on startup, but there doesn't seem to be a device for it when i do iwconfig… what am I doing wrong?

why use a custom ebuild?

does dmesg say anything?

/usr/lib/gcc/i686-pc-linux-gnu/4.1.2/../../../../i686-pc-linux-gnu/bin/ld: : No such file: No such file or directory - any help possible??

gcc-config -l

gcc-config -l
[1] NATIVE
[2] i686-pc-linux-gnu-4.1.2 *
[3] i686-pc-linux-gnu-4.2.0

try gcc-config 3

(sorry i havent been using irc for quiete a long time)

to use a CVS rt2×00, the builds in portage don't seem to work… i'm using this: http://forums.gentoo.org/viewtopic-t-577202.html

maul same error message

sounds like a binutils problem to me

binutils-config -l

don't suppose you changed your CHOST recently or anything?

TFKyle yeah well if i use "ld" at command line, it works
TFKyle hmm might be a good hint in the correct direction lemme check
might be, yeah, cant say for sure now
binutils-config -l
[1] i686-pc-linux-gnu-2.17.50.0.18 *

also try a binutils-config, could be that somehow that symlinks got killed somehow

Deifyed, why not put you xorg.conf on a pastebin?

that seems to be ok

(try changing it with it)

kk

did you read the last message of that thread? See also bug 185229

https://bugs.gentoo.org/185229 nor, P2, All, mike@ossmann.com-uberlord@gentoo.org, NEW, pending, net-wireless/rt2×00-9999 ebuild - updated to work with current CVS version

TFKyle yeah somehow it seems to be a symlink problem, i dont know how to regenerate them qnd since i cant emerge anything, it wont work that way

binutils-config 1 regenerates the symlink

maul ah thanks lets see

or at least it should

yeah, i thought that was for if the ebuild didn't work, for me it merged successfully

maul it doesnt say so, but lets see
at least it does not fix the error, no.

alternatively you could create the binutils-related symlinks manually, a bit ugly but I figure it isn't that bad as binutils-config should create them

is /usr/lib/gcc/i686-pc-linux-gnu a valid directory? does it have a valid bin directory?

maul its valid, but does not have a bin directory

here is some stuff from my dmesg: http://rafb.net/p/ekD3OT55.html

[–oldarch old-CHOST]

create the symlink manually, it's pretty ugly but should work
or try what tinycat just said

ok, i will first try what tinycat said.

ehm, cahoot, whats the best way of copying all the content in a file?
uehm, cahoot, whats the best web hosting way of copying all the content in a file?/u

meh

seems like the driver doesn't work. I assume it's still in development

yes

nopaste, wgetpaste

it works for other folks… what should i try?

kk thx hkBst

I dunno, I also had the same problem with iwlwifi, but the next release worked for me then
maybe you can try passing custom options to the module, check the documentation for them

http://selene.tfkyle.dyndns.org/~kyle/Temp/foo.txt are the symlinks I have in /usr/i686-pc-linux-gnu/bin (just want to link the ones related to binutils, and obviously to the right version you have if it's different than mine)

fix_libtool did not help

okay, i'll look into that. also there was another ebuild for it that i could try, i'm not sure if that would help

TFKyle thanks alot, i will try that

you can never know

okay

TFKyle but i have the symlinks there already?! i'm feeling a bit dump actually.
dumb of course

actually, I don't have those directories at all, but my system works..

well, just the ones you don't have (I imagine ld and atleast a couple other binutils related ones)

What is the other way to sync? without emerge –sync?

ls -la /usr/i686-pc-linux-gnu/binutils-bin/2.17.50.0.18/ld
9

haha, it works. It was an initrd issue, so I just compiled the usb drivers directly into the kernel.

in i686-pc-linux-gnu/bin, sorry

good

http://rafb.net/p/zCUTsB15.html

thanks again for your help

np

is there a way to recover a kernel config from a bzImage or a system.map ?

yes

that was a ld of the files in that dir, the part to the right of the arrow is the symlinks target

emerge –sync keeps starting over

boot into that bzimage and it's in /proc/config.gz if you're lucky

maul’ how to do ?

it's the 8th time I think

hmm, you don't have /usr/*-linux-gnu/bin?

use emerge-webrsync, use another –sync mirror, or get a portage tarball

(depends on your CHOST of course)

yeah I checked and it's there

/proc/config.gz I think

it's ld binary links to binutils-bin's ld

yeah

kojiro, k thinks. I think my connection is too unreliable or somethin

i have that, too.

maul@frankie ~ $ ls /usr/x86_64-pc-linux-gnu/bin/ld -l

/usr/x86_64-pc-linux-gnu/binutils-bin/2.17.50.0.18/ld

it should say something like this

yeah, it does.

and it still doesn't work? is /usr/x86_64-pc-linux-gnu/binutils-bin/ld a valid binary?
or wherever your final ld binary is

maul yeah
/usr/i686-pc-linux-gnu/binutils-bin/2.17.50.0.18/ld

yeah

/usr/i686-pc-linux-gnu/binutils-bin/2.17.50.0.18/ld: no input files

you have /usr/i686-pc-linux-gnu/bin/ld?

and it still gives that error?
typo
I missed the version

TFKyle yeah i have

really? portage doesn't seem to agree

and yeah its still the error

hey

u still with me? :p

my damn qt-compile still segfaults
in case someone cares

would i have needed to recompile my kernel after merging this kind of package?

hehe

when trying to make ur own minimal cd how do i emerge kernel sources when emerge isnt in the suqsh image ?

no

Deifyed, I see no obvious wrong - rerun ati-config?

*squashfs

you have to re-emerge the package after each kernel change tho…

just a reboot would be needed?

Hi, ALL! Have a problem. After update of Nvidia-drivers to 100.x my video is not working. In DMESG file are strings, that last version of GF4200Ti support is 96.x So the question is: what is the better - to use 100.x legasy or go back to 96.x?

is anyone having troubles with pypanel?
when i emerge it, no executable is installed

if it's a module, you should be able to load it with modprobe

i cant emerge :$

I'm literally talking about /usr/i686-pc-linux-gnu/bin/ld here btw, not /usr/i686-pc-linux-gnu/binutils-bin/…/ld

if i unmerge it, will it remove the modules as well?

kk. thanks for checking it out

probably not

is emerge supported on minimal cd ?

ls -la /usr/lib/gcc/i686-pc-linux-gnu/4.2.0/../../../../i686-pc-linux-gnu/bin/ld

/usr/i686-pc-linux-gnu/binutils-bin/2.17.50.0.18/ld

TFKyle i dont understand that really.

ah, ok, you do have it, sorry

can you run that ld command by hand? the one with ../../../ stuff?

when do you get the error?

maul sure
h0m3r bin # ld
no input files

(merging what)

how would i go about doing that?

but that exact command, not just ld
removing the module from memory?

TFKyle whenever ./configure is involved in a emerge

is emerge supported on minimal cd ?

rmmod

/usr/i686-pc-linux-gnu/binutils-bin/2.17.50.0.18/ld: no input files

but does /usr/lib/gcc/i686-pc-linux-gnu/4.2.0/../../../../i686-pc-linux-gnu/bin/ld
work?

maul yes

hmm, mind trying to merge something that has a ./configure and pasting config.log?

sometimes I get a symlink issue

/usr/lib/gcc/i686-pc-linux-gnu/4.2.0/../../../../i686-pc-linux-gnu/bin/ld: no input files

ok, sorry

er, pastebinning rather, pasting it would be horrible)

TFKyle no problem

somebody ?
is emerge supported on minimal cd ?

try it…

well it says emerge not found :$

then nope…

k how do i install emerge ?

it would be easier to use the desktop livecd

maul im on live cd

what are you trying to do?

the minimal one?

but trying to generate a minimal cd

TFKyle http://pastebin.com/m469cdfeb

hmm, that's cut off, try at rafb.net

emerge gentoo-sources so i can reconfigure the minimal boot cd kernel

er, ignore me

You can't use emerge from the minimal cd. To create new, customised minimal cd's you'll need to use catalyst.

ah
and i can do this from livecd ?
got a catalyst howto ?

there is a bigger livecd which has emerge installed I think

check http://gentoo-wiki.com

I have no /proc/config.gz

You can, however, install Gentoo from just about any live linux hosting distro, so I recommend you find one that has the features you want and use that.

well i wanne make it as small a possible

TFKyle thats the config.log created, its not cut off ?!

tried commenting out LDFLAGS in make.conf and seeing if it helps?

if you don't have config.gz then no

TFKyle ok

What exact features do you need that you believe the minimal install cd doesn't have?

the one earlier was 900mb :$

(probably not that, but who knows)

Hey all, I'm linux super beginner…. I just emerged Gnome and set that to my XSESSION in rc.conf, but it keeps restarting in this otherxsm interface… Any idea why?

TFKyle "heureka"

the minimal one is only 86MB or so and has everything you need.

support for my vt8251 via chipset (sata but ahci instead of via_sata

that was easy. hehehe.

so i rmmod'ed the three rt2×00 drivers, unmerged, merged with the different ubuild, now modprobed the drivers again, is that supposed to set up a wireless device for me? what do i do now?

it doesnt have emerge

I'm sure it could be a million things… But I haven't done much, I just barely installed gentoo
and nothing else really.

dmesg?

but I guess I'm "half way there" myself

it doesnt have that because you dont need it. after you unpacked the stage3 file and chrooted it will be available.

TFKyle i am using them for a long time, i have no idea whats wrong. but i have to agree that i played too much with my system toiday.

agh, same error, dur

TFKyle and maul, thanks alot for your help!

anyway, if the driver is not done yet, there's no guarantee it would work

care to search with me ? i didnt found one web hosting for online images with vt8251 that is still online :$

what was the solution? I missed it

and btw, doesn't -march=native only work in gcc 4.3+?

grrrrrrrrrrr

TFKyle i read it works in 4.2.x if i'm not completely wrong.

i just dide emerge –sync and its hosed my system.

maul outcommenting LDFLAGS.

k thkx tho

hehe ok

Unfortunately I have no experience with that exact chipset, so can't really help you.

all of a sudden glibc 2.3.6 is required for ALL versions of gcc, even ones that DIDNT require it before that i have installed
why is this?

ah, I'm probably wrong

it didnt need it before, it working fine, why suddenly require it?

they usually change things like that at gentoo

it makes NO sense

i'll keep trying though… i know other people have gotten this device to work

remember your current gcc is 4.1.2 btw :P

the package is already installed and working, why does it suddenly require something it didnt before, my only option now is to reinstall!

i'll try the legacy driver

TFKyle see http://gcc.gnu.org/gcc-4.2/changes.html native should work for 4.2.x

maybe they messed up, when did you sinc the last time?
good luck

i synched 3 weeks ago during all the expat craq
and again today

TFKyle yeah hehehehe

thanks for helping me, i appreciate it

try waiting, maybe they'll fix it, that's all I can think of
np
or… you could edit the ebuild manually

my guess is that 4.1 ignores a native march and thus you're probably not getting any march beyond i586 maybe
(or i686)

yeah
i wonder if there is a way to find out with what flags / fwhat host a binary was compiled for

there is

i686

im going to mask all of gcc 4.1

can see if it was compiled for x86 or amd64, but probably nothing better than that with just the ELF headers

Intel Corporation 82845G/GL[Brookdale-G]/GE Chipset Integrated Graphics Device (rev 01); what USE-flags should i set for building media-libs/mesa?

im not even using 4.1 on here…

(well, 32bit of 64bit actually)

maul will you reveal the name of that tool ?

Are you talking about "a binary" as in a "binary package" (ie. created with portage / quickpkg) or binary as in any executable binary file?

any executable file

as TFKyle said, you can see the arch, but not the exact processor type
with readelf -h

?

ah i remember that command yeah

how do i stop it keep trying to reinstall gnome-screensaver
i cant stand gnome-screensaver, it keeps hosing my xscreensaver setup

and if that says 80386 its bad, wh

ack damn up key sorry
it seems to be a required part of the gnome meta package for no real reason

list

using gnome or gnome-light?

gnome

it's required because it's part of gnomes moduleset

i dont want everything else removing, just gnome-screensaver

also try readelf -A -W

you can probably either change the gnome ebuild in a local overlay to not dep on it (and possibly other stuff, not sure) or use gnome-light possibly

the libraries, nice.

thats what i did :/
no other choice
gnome is becoming such a pain

WARNING: Library files for ".la" not found in paths. –something that I forgot to do?

just for people who don't like gnome stuff

maybe you use a virtual that pulls in all the gnome packages as dependency. doesn't it have a modularized version?

hmm

try revdep-rebuild

i might just reinstall this damn laptop
gentoo is nothing but a pain in the ass on it
its fine on my desktop, but on here the slightest upgrade takes years and it ALWAYS breaks

as root?

use distcc and it will only take half-years

Anyone know why I can't seem to boot up into Gnome? I have it set in my rc.conf, but it keeps defaulting to some XSM interface.

thats not the only problem

yeah, it's meant to fix library dependency problems

yep, distcc++

ok
theres something else i should do
that will fix most of the mess
see you in a few days, im gonna change my chost

on livecd i get thers no catalyst ebuild with emerge
what can i do to find this ebuild ?

emerge –sync ?

pentium 3…. i686 or i586?

i686

ta

686

(technically a superset of it, i686 doesn't include sse and stuff)
but i686 is the highest CHOST no need to have anything above that for x86 boxes (I think)

there's a table on the wiki with all the processor types and cflags for them

s//,/

nope

then I don't know
sorry

thanks anyways

ahhh
i see, its the crusade against linuxthreads.

please!

plus, my domain is banned from #kde…I would try there but I can't

how to record a cd from command line?
urgent!

the 'you WILL use nptl and you WONT get a choice in it' people are to blame for this.

copy the whole cd to an iso file?

maul, nop
create a cd from files on hd

mkisofs creates the iso
cdrdao can write it I think

you have to agree that there's hardly any reason to use linuxthreads anymore, and nptl is pretty superior to them

Audio cds have no filesystem.

i think he means k3b

he wants to create data, doesn't he?

k3b will burn

but that's not for cli

cli ?

hi akk

command line interface

(all

Hmm.. Maybe he does, the "record" part somehow made me believe s/hes working with audio cd's

thats not the point, i use linux for choice… speaking of which, when will there be an OSS4 ebuild, any news on it?

ahhn true

is .xsession-errors an important file or can I delete it? (it's taking up 1,5GB)

i cant wait to get rid of alsa and move to something that can do 5.1 properly with my soundcard :/

check bugzilla

you can delete it

why my Genuine Intel(R) CPU T2080 @ 1.73GHz is not working with speedstep-* cpufreq drivers but it works with p4clock-mod ?

Intel Corporation 82845G/GL[Brookdale-G]/GE Chipset Integrated Graphics Device (rev 01); what USE-flags should i set for building media-libs/mesa?

(or failing that when you make one :P )

i810

hi, does anyone know what package I have to emerge in order to type traditional chinese with scim? I have scim-pinyin, but it seems to me it's only for simplified chinese

VIDEO_CARDS="i810 i915" I think

thanks

why my Genuine Intel(R) CPU T2080 @ 1.73GHz is not working with speedstep-* cpufreq drivers but it works with p4clock-mod ?

what's wrong, if it works with p4clockmod?

choice to use something that's been killed upstream for a year or so and semi-unmaintained in the face that the alternative works perfectly fine for most stuff and is maintained upstream? (that's my interpretation of it, could just be me)

scim-m17n maybe

(course, that is probably exaggerated a bit)

I thought that it should work with speedstep-* driver

I currently have xfce4, kde and gnome installed (with all their respective packages/dependencies) How do I remove kde and xfce4 completely?

Because this CPU supports SpeedStep

guess it's only killed in the main glibc tree, there's another tree with linuxthreads in it in the same repo

equery l kde lists all your kde packages, you can unmerge them
or equery g kde

how do I find out why a package is hard masked?

maul, thanks

what, like xmms? :P

thanks, I'll try it

yeah, cause you know xmms was removed as well
though, I think with 2.6 you can still use linuxthreads if you really want to, there's the linuxthreads-tls USE flag
but no one would recommend it I imagine

where is the file that has all of the hard masked packages?

/usr/portage/profiles/package.mask

grrrrr
im going to have so many issues with this changing chost thing
it makes so many assumptions

thanks

anyone knows where ccache lives?

i can see me just reinstalling

Hi. I have various versions of python installed (2.3, 2.4 and 2.5). Is it safe to remove 2.3?

like, for some reason this laptop is on gcc 3.3, and the reinstall version of gcc set by emerge -av1 binutils gcc glibc is 3.4
:/
i didnt get a choice in that
so its not going to touch the 3.3 and glibc is going to be rebuild with the 3.3

Anyone knows? Would it be safe?

R-Gun-Rivale: probably yes

OK, I can always re-emerge if something breaks too much. THank you.

R-Gun-Rivale: yes, but it should work

can I merge all that that package want ?

pastebin the emerge output so we can see

ok

where is the best place to ask when specific packages in portage will be updated?

try to find it's maintainer

ok

and have a look at bugs.gento.org as well if there is someone working on that specific package.

Argh! So the live cd doesn't have a new enough version of the tg3 ethernet driver, it seems. How can I get a newer version of the module up? I can cross-compile on another x64 pc I have but with the installation cd having a ro fs I can't figure out how to load it.

where to find statd?

find any linux cd that has the driver, and use that for installing gentoo

statd: server localhost not responding, timed out seems a clue

whats wrong with the old tg3 driver?

Don't know exactly but the machine shipped with ubuntu 7.04 (2.6.20) and it worked but in 2.6.19 it's not fully recognized.

sounds strange, the driver doesnt load or?

for some reason, my desktop aint limited by the monitor limits, so the right side is acting like my monitor is twice the size, any ideas?

Driver loads and shows up as being 'Broadcom Ethernet Unrecognized Device' with an unrecognized subsys. Device shows up as eth0 but can't actually communicate. dmesg is sparse — just a bunch of messages about messing with duplex settings.

okies. I've got a broadcom myself and never had any problems with the tg3 driver. and i've been using it on every kernel release under 2.6 so far.

The hardware is a laptop (Dell) and the mainboard platform wasn't released until July 07 so I'm just staking my claim that it has something to do with that.

probably, it doesnt show up correctly when you boot a livecd either?

hey what's the relationship between source and binary when they are installed in a system? Example, a certain library compiled but an application which depends on it is installed as a binary..can they still relate?

basicly everything is compiled, you don't need sources at all
but if you want to compile something, you'll need specific headers of other packages

because sometimes i stumble upon a situation when a dependency also needs to be compiled in order for the said application to function/install.

It knows what the device is… BCM5906M… but it's the unrecognized dell subsys that seems to be the only thing that's off.

can openoffice integrate into firefox?

not under linux afaik

k

http://rafb.net/p/Oc6Zfq32.html how to solve this ?

WARNING: Library files for ".la" not found in paths. –something that I forgot to do?

morning

what if you try it without the cups use flag?
or without doc

Portage is having a problem overwriting the file '/usr/lib/python2.4/encodings/euc_jis_2004.pyc' - python is saying Permission denied, and even when I try to 'ls -l' the file it says Permission denied. I'm getting I/O errors in dmesg (http://pastebin.ca/671348). What is this, bad sectors? How can I get rid of it?

Updating from glibc 2.5 to glibc 2.6 breaks too much stuff?

libexpat.so seems to be missing after an upgrade… i hope revdep-rebuild saves my day

maul’ works :]

(

can someone help me unmerge avahi and mDNSRespinder?

scriptdevil, Don't forget to check the /topic. Oh, and welcome to expat hell.

yeah that libexpat.so caused me to be unable to login to my system as root becouse emerge update failed in the middle and hasn't come to 'shadow' which didn't work as I had to remove 'pam-login'… than I ended up exploiting some web interface script to fix it

bit`: sorry .. did not read it.. was blinded :P
/msg mzbot faq expat
missed the space :-|

ScriptBlue, no prob. It way help

well.. bit` i dont seem to get a reply

Uhm, updating to glibc 2.6 from glibc 2.5 is safe enough?

i got it
thanks

can someone help me unmerge avahi and mDNSRespinder?

how can I scan a disc for bad blocks and use it with reiserfs ?

I'm running nvidia closed source stuff, which changes should be done to use the free driver? (and, OT, where can I find a list of cards supported by the free driver?)

it made me unmask portage too… :-|

I'm trying to emerge Gnome on a new install of gentoo. I'm getting an error while it is downloading though. "Could not find a usable .config in the kernal source director"

bit`: am scared… the command seems to be running full memory for more than a few minutes

did you already compile a kernel?
check that "eselect kernel list" points to the right kernel. Or check directly "ls -l /usr/src/linux"

I'm using KDE, should I go with xdm or kdm?

kdm lets you auto login and looks better
but it doesn't really matter

kdm it is then.

kdm is nicer

short question, was there a version bump of gcc to 4.2.0 on stable i686 recently?

How do I make it so that when I –sync rsync doesn't start over if the connection is lost?

allright, how will I know its the right kernel?, eh, I'll just see what I get

for some reason, my desktop aint limited by the monitor limits, so the right side is acting like my monitor is twice the size, any ideas?

looks like its there…. linux-2.6.21-gentoo-r4

hi people

startx is the best

Deifyed, I know it'll do that if you Ctrl+Alt+'+' to lower the resolution

oO
i never wrote that tho
doesnt help zooming out nor in neither

hi

well… arent there buttons on ur monitor

check your running kernel with "uname -a"

doesnt help :p lol

I don't think so

is the left side of the screen blank?

what if you try to set the resolution in kde/gnome config?

2.6.16-ps3 #1 SMP ?
so… it's pointing to the wrong kernel?

hkBst, I don't think either, but then I wonder even more, why crossdev wants to emerge 4.2.0

Heh, I just saw a change in dispatch-conf whose only effect was to change the word "effect" to "affect" in a comment.

and then it comlains it's masked ~?

no its not black, left side works perfect

no it breaks somewhere

Deifyed, somewhere you must have set the resolution in your window manager to higher or maybe you selected a larger virtual resolution in xorg.conf

i've only tried using aticonfig and xorgconfig

yeah, you should probably upgrade your kernel

You have two monitors?

see if you can do something in kde's or gnome's config panel

definatly no virtual resolution in xorg.conf

I'm on x86_64 now, and I want to build a powperpc tool-chain and on my P4 I wanted to build a x86_64 tool chain

http://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=1&chap=7

kk ill check maul

Deifyed, do you have two manitors?

both tried to get a gcc 4.2.0

no only 1

what's wrong? crossdev compiled it perfectly on my x86 box

when I am merging xorg-x11 7.2, I noticed that it tries to pull in gcc-3.3.6-r1 for whatever random reason, and then, when gcc attempts to merge, it fails with an error saying: emake failed with bootstrap-lean. I also tried a diff version of gcc and no dice, help anyone?

Deifyed, try running just "X" from the console and see if it still that way. If it is, you know it's not your window manager

shit

what means "breaks somewhere"?

kk
its still like that

maul, here it does not resect the max stable version for the arch you want to build a tool-chain for

running X

oh, well yeah, it built 4.2.0 here too, but I use unstable anyway

hkBst, I didn't care because I can't use a working 4.2.0 for powerpc anyway

can you put your Xorg.conf on the web somewhere?

hkBst, I'll have to find a PPC kernel most likely right. I'm on a PS3

can i emerge noip-duc?

yea

Deifyed, then you know it has something to do with your xorg configuration. Try reading /var/log/Xorg.0.log look for anything unusual

it's noip-updater

After an emerge world, I rebooted and now all I get is "GRUB" on the screen. Any ideas?

maul thanks ;D

ask in #gentoo-ppc64

is your /boot/grub/menu.lst correct?

Hi folks, I'm hosed. I was rebuilding, and my system reset before I'd finished compiling pam. Now, I can't log. I can get to webmin, but no ssh or console. Any ideas of what I can do?

maul, http://rafb.net/p/mmEdeC52.html

It is, it remains untouched.

liveCD

etc-update?

just chroot and rebuild the broken stuff

drmikecrowe, boot in to single user mode. linux single or such I think

I can't get to a console.

Hey guys.. i have a aserver that is crashed and i'm trying to boot it remotely off the livecd to use mdadm to fix the array..

kk i will thanks

reboot to a livecd and you will have a console…

is there a command to boot a faster smaller boot then the normal boot?

what's your resolution?

"by using this ebuild's config option" - how would be that?

bobbylee, what do you mean?

etc-update says everything is up to date.

like.. i just want to be able to use mdadm but it takes forever to boot off the livecd (because im redirecting it from my computer here over a java app)

maybe it is a hardware problem?

bobbylee, i have no idea what you are trying to do

1280×1024 but i got told to try 1280×768
neither fixes the problem

ok, it was just strange

ok have you ever booted the gentoo live cd?

bobbylee, not in a long long long time

that corrupted some files? Dunno, I'm able to boot from the livecd and everything seems to work when I chroot into my install.

bobbylee, does it open vnc or something?

no but it just loads a bunch of crap i dont need

join #gentoo-nl

use the "nox" boot option?

and my console is f'd so it keeps disconnecting me and i have to start over

bobbylee, do a gentoo install from knoppix

why does it have 2 screen sections?

not sure

I guess you could try reinstalling grub. ANother thing is to check your emerge log for anything suspicious.

tried removing one of them but fucked something up

make a backup, then delete the first screen section entirely, and all the viewport lines

?

such as a grub update
?

Yeah… I guess I'll try that next.

after the kernel "gentoo nox"

I just wanted to know what the "GRUB" is supposed to stand for first.

kk i will, just gotta walk the dog first :p thanks for the help

it sucks when you have a 16-disk raid 1+0 and a/b both fail at the same time

you mean what grub is? It is the bootloader

did you install from the liveCD, originally?

No, I think I built from a stage1. Long time ago.

why will qemu not compile? app-emulation/qemu-softmmu-0.9.0 keeps telling me qemu requires gcc-3 in order to build and work correctly, but I have gcc4. what am I doing wrong?

I'm downloading a minimal cd and will do a chroot and rebuilt

that's fine, anyway you remember how to chroot and stuff. A lot of the new users who use the GUI installer don't know that stuff.

it needs gcc3 not 4

makes me cry

so how can I install it if I have gcc4?

The "GRUB" message that shows up.

I *think* I remember. I need to make sure I get the environment set up

either install gcc-3 or not use it at all

grub usually gives an error code at least. It doesn't do this "GRUB" thing.

This was originally a 2005.1 install, I think

http://www.gentoo.org/doc/en/grub-error-guide.xml it does lots of crazy things

why is it in portage if it does not work? What other emulation tools can I use?

virtualbox is pretty good. qemu does work, it just needs an older gcc

after almost nearly destroyed the partition where I had / , (and tryed to recover it) I have an error during the boot after the INIT part, the error saids "/sbin/rc: cannot execute binary file", could somebody help me?

I guess I don't understand why for any other reason I would stay with gcc3.

problem 11 it seems

xaque, why arent you using virtualbox ?

it's not a question of staying with gcc-3, it simply won't compile hosting on gcc-4. ask the qemu developers why they don't fix it

because I never knew of it?

try it.

hi

will do.
can I use my qemu images?

i have a problem with the newer kernel, anyone of assistance here?

it works with gcc4, is faster than qemu and opensource, too.

newest*

woo it booted!

xaque, i dont know, but look @ virtualbox.org

go go crappy console card

No, that's not it. It's more like the 7th, but there's a single "GRUB".

support for solaris? redhat? bsd? and vista?

please, colud somebody helpme to recover my gentoo?

ah, I will check it out
thanks

The problem is that i can't get gentoo to recognize my Intel SATA drive
And i can't find the module SCSI - Lower Devices
anyone?

you can't find the setting?

Comments

Can you help me A friend has somehow managed to destroy his ext3 it keeps saying bad blocks and it wont boot

Cache-Limit ?

whats that -f comand doo

hello, i was just wondering if editing /etc/mailcap by hand is safe, or if there's some external program i should be using.

it is if the box is on the internet

a user box? and what services would your iptables rules be blocking?

the vpn to my remotely available user box.

for one torents a fire f wall blocks ssh and vnc

(ok, so it's more like "sshd", not a real vpn)

even still, if its properly configured, whats the use of iptables except maybe rate limiting spamming connections?

that's it.

right, and so my original advice stands. a user box doesnt need iptables

i trust liable hes tought me that right stuff he tought me how to fix
lots
were is my note book gone too?

time to sleep for me
laters

how are you today?

Salut à tous
hi all

turion x2 with 1gb of ram and 256mb ati card
do i buy that for 150… ??
:P

I have no idea. 150 what? how large a display? Do the graphic and wireless cards work with free drivers (if you care about that)? I hear laptops with AMD CPUs can get hot.

15wxga
glossy
1gb of ram
i am surfing ebay for possible candidates 150 bucks looks good

off to get some saop and do cloths today

15.4" wxga is what I have, is acceptable (96 dpi). Glossy is a minus, IMHO, depending on your surroundings you may like it or not.
I don't think you'll get it for 150 US$ final price.

okay, how do i configure sound in debian? i have an old vibra 128 card

sounds a bit too cheap

true, it is just a bid

stuart-: have you been through alsaconf?

look for "already completed" similar items to get a better price estimate

themill, not yet, will try
themill, command not found. what packages do i need to d/l? i'm clean on a net-inst

stuart-: alsa-utils

themill, thanks

!tell stuart- about search

themill, i config'd and alsa asked me to have fun. but xmms asks me to configure my soundcard and won't play mp'3s

Can I run a command that tells GRUB to boot a particular entry once, then boot the normal default upon the next reboot?

you might be able to hack something with setdefault

interesting idea — thanks.

stuart-: try something simple like aplay first.

see "grub-reboot"

themill, aplay just stalls

thanks

stuart-: odd. are you sending it an mp3?

can someone tell me what to do about an interface called wlan0_rename?

!z25

methinks z25 is as of version 0.090, udev has the ability to staticly rename ethernet cards based on MAC address. the addresses are configured in /etc/udev/rules.d/z25_persistent-net.rules . If you want the mappings to change, edit that file. if you don't want udev to touch your interface names, remove that file and z45_persistent-net-generator.rules in the same directory

z25 is the best thing since sliced bread

should I be able to use it with the name wlan0_rename? I am getting an erro when i try to set the essid on the card.

can you netinst through wireless?

my friend is having a problem beside me now, linux won't detect his wireless, and lists it as a firewire

I'm just assuming the name is set by udev - look into the file and see if that seems correct
you can chnge it to whatever you like there

stuart-: unlikely that it gets it that wrong. do you know what sort of wireless it is supposed to be?

stuart-: more likely he has a firewire port as well, and the wireless isn't listed

stuart-: look on both lspci and lsusb for the wireless card
stuart-: and some usb wireless cards have a kill switch that disconnects them from the bus not just turns off the radio

so do you guys recommend him instalilng the base first? because the network is slowing down the whole process
he says he has an 802.11b wireless, or something? haha

stuart-: sorry… I'm not with you. What have you done and what are you trying to achieve?

themill, trying to net-inst debian on his lappie with a wireless

stuart-: I don't believe that netinst will work with wireless. Can you stick a cable in the back of it to do the install?

hey.,. anybody know when the
KDE4-beta packages will be available to debian?

themill, yeah, that'll be okay. but i'll have to dc then. and this pc doesnt have wireless! oh no

hm. Is there a way to speed up massive deletion on an ext3 partition? It's idle, and I need to delete about a half of some 5M files.

d/l a full cd perhaps

themill, he has a router line at home though, how can i make sure his core installation recognizes his network card?

stuart-: you can get the base system working with the netinstall image… and if you've got a wireless card that is supported by the kernel without any extra firmware….. maybe.

yo

Hi

speaking of ext3 … when the hell is the 'superblock last write time in the future' going to be fixed

don't write from the future?

i don't, unfortunately… i'd write me the lotto numbers

stuart-: the netinstall image will get you to a working system that will see the network card (usually) and use dhcp or a static addr with it.
stuart-: I've not had netinstall fail to recognise a network card yet…

that's why i went jfs .. but i put ext3 on the laptop

investigating

you tell me

what kind of workload do you have on jfs?

it's all home based… simonrvn or peterS uses it, i forget which, they'd probably actually have a 'workload'
i went jfs a year or two ago, i forget, only one error since, and i can blame hardware i think

is it possible to pass some kind of "mtime" argument to rsync to only copy files of a certain age?

i read the manpage once, iirc yes

themill, okay i have the core system now. so there's no way i can set up wifi now?

there's no mention of "time" in the manpage or anything to do with dates

find, and rsync –files-from= ?

I wanted to specify the the mtime on a remote machine :|
i.e. pull files rather than push them

uh… find on the other machine :-)

I don't see how I can do that
I want to run rsync on the target

duh, copy over the file!
*file list

-f, –filter=RULE add a file-filtering RULE

I don't wnat to run anything on the source

not even rsync?

not interactively
or "actively"

can anyone help me get a dwl650 up and running, the machine has a broken internal wireless card and i think they are somehow competing….

whoops, what package do i need to install if it says lspci command not found?

I've forgot, what's the command to have info on a package ?
like the version …

stuart-: pciutils

apt-cache show package ?

installed: dpkg -s package-name. available for installation: aptitude show package-name

!search

to find factoids containing thing. see http://wiki.debian.org/WhereIsIt

Jean11, & jetscreamer thank you

ErNi_, pciutils command nto found

stuart-: install pciutils package

stuart-: that's not a command.

help! i cannot update my aptitude

more details required

Dynamic MMap ran out of room
Um erro ocorreu processando php4-dev (NovaVersão1)
Problem with MergeList /var/lib/apt/lists/security.debian.org_dists_testing_updates_main_binary-i386_Packages
As listas de pacotes ou os arquivos de status não puderam ser analisados ou abertos.
sorry

ErNi_, no installation candidate

lol

!paste

Please do not paste anything at all to this channel. Instead, use: http://channels.debian.net/paste/ or http://rafb.net/paste/ or http://picpaste.com for pics.

!tell SamuraiDio -about dynamic mmap
got that?

stuart-: http://packages.debian.org/pciutils — there is pciutils package in all branchs

I suppose I could do "ssh host find…" first

jelly, yeah, tnx
jelly, that values means bytes?

I think so, yes

!tell SamuraiDio -about dynamic mmap

hi there
i've configured an vpn network here, how do i set an interface as the default route for an ip range?
route add -net 150.162.0.0 netmask 255.255.0.0 dev ppp0
this command stops all my connections
and ppp0 goes down

jelly, tnx, it worked

How can i `run_init` to /dev/loop0 that is made of a cramfs file?

ok, file-rc seems to be nice, sysv-rc-conf is good, but I would like to have init more logical (like to see while restarting networking that privoxy restarts too …)
is there some init system that works well with debian too?

is your gateway also in the 150.162.0.0/16 range?

jelly, probably the vpn gateway is there, it's the first time i am using VPNs

some reliable initng replacement

jelly, so i don't understand much about this

pastebin the output of route -n while it works

privoxy doesn't restart when networking restarts

he wants it to

How can i `run_init` to /dev/loop0 that is made of a cramfs file?

sorry, tor should (should not it?)

it doesn't
i don't know of any instances in debian where running one init script in turn runs another

jelly, http://rafb.net/p/6O8ldn68.html
jelly, here

if you want, you can put commands in /etc/network/interfaces to be run when eg. eth0 goes up or down

I usually can read DVD without problem. Just inserting it and then I got ask what I want to do with it. I just got a new DVD and it KB3 tells me the DVD player is empty. mount /dev/hdb /media/cdrom answer no medium found. Where can the problem be ? (my gf can read the dvd on her Mac)

jelly, here is my ifconfig ppp0 http://rafb.net/p/qdotUu83.html

debian-newbie here looking for a little guidance

and your internet connection works with this route -n?

guys, I'm trying to do a netboot, but I can't get the network device runnig
what's the command to get it up and down?

ok, but could this be specified like: run this if any of wurking network interfaces restarts (wifi, eth ..)?

jelly, yes, this is how my routes look after i start the ppp0
jelly, my normal internet is routed by 10.1.1.1

eg. privoxy should restart after tor, no?

ifup or ifdown ?

that can be done via /etc/network/interfaces or /etc/network/if-up.d

ah nm

why do you have another default route via 192.168.0.1, iface eth1?

What happens when you add a group that could cause a delay? A script I use returns exit status 6 from useradd and I tried :groupadd -g gid groupname; useradd -g gid -u uid -s /bin/false username" on the CL and I get "useradd: unknown group groupname" but by the time I "cat /etc/group", it exists.

stew there is no if-up.d

jelly, i want to use the 150.162.0.0 ip class via ppp0

I'm getting "Your network is probably not using DHCP protocol. I'm loading debian onto on older (P2-233MHz-256RAM)

IN a console what is the proper pramiter to extract file.tar.gz into a folder the same name as the compressed file minus the tar.gz?, i just cant seem ot get it

jelly, this is my neightboor wifi network hehe

[Jasper-2]: yes there is

wouldn't it be nice to specify dependencies and there restart? (Like start net, load modules and start firewall before doing this?)
I ask only to know your opinion

no stew there aint
that's why I'm asking for different methods, I'm doing a netinstall, via a rarp server
debian web hosting for online images is only 5 mb big

jelly, i want to set my routes to access 150.162.*.* ips via ppp0, but i don't know how

How can i `run_init` to /dev/loop0 that is made of a cramfs file?

~ # ifup eth0
/bin/sh: ifup: not found
see?

[Jasper-2]: what version of ifupdown do you have installed?

stew NONE
it's a MINIMAL install
so there MUST be another way debian does it

[Jasper-2]: what version of debian are you running?

stew, I'm doing a netinstall
of a minimal version of 5 mb

stew my aslsa wont hold chy i try new card?

[Jasper-2]: why are you asking me about /etc/network/ifupdown?

I am not
I asked if there was a way to ifup or ifdown a network device
without the ifup command

well, your route line looks sane, but your initial config does not. ;-)
why do you always set up a default route to the free wifi?

hmm, well. i think that the current system of putting stuff in if-up.d and if-down.d is good only in that it is something that happens when you run ifup and ifdown which may happen independent of init scripts
i dont' see much wrong with the current system

i get a /var/lib/dbus/machine-id: no such file or directory problem when i start GAIM, any ideas?

jelly, isn't always, but i use it sometimes, it has more bandwidth than my ADSL

well, mazbe I am influenced bz gentoo

though I agree that once all the init scripts defintaely have lsb headers, some better system than sysv-rc can probably be implemented. but i haven't really looked at file-rc or initng
do you know which driver you are using?
and what do you mean by "wont hold"?

know i dont its deaflt one
have to re fun config each time
get pc sounds on boot

hm. and what's your vpn gateway, that .245.10 machine?

have to re run conf to get sound from pc like radio

jean_, probably, i'm not sure

i'm not familiar with gentoo's init scheme, though i'd be interested in reading about it if you had a link

can you pastebin the vpn config, without username / password?

ok, which vpn config? the one from /etc/ppp/peers ?

stew il look at drivers need rebbot hold on

why reboot?

"expires in less than 18446744073709551614 days". I think I am on the safe side :-)

I don't know, probably.

is there a simple way on the command line to determine if a remote machine is responding to ping which can be used in a bash script?

well, I only tried to use and it seems to be fast (own impression) I will try to search some links …

a lot of work was done recently into looking at how you could speed up the debian boot process, and its mostly cpu/disk bound, that largest improvement would be from switching /bin/sh from bash to dash

Hi All!!

Does anyone here knows a good debugger for multithreading apps such as PTHREAD?

Serial. Trouble is that if I insmod ftdi_sio, it attaches usb 2.7 to ttyUSB0 - usb 2.7 is (what I assume to be) the multimedia portion of my Logitech G15 keyboard. The brltty (braille display…) daemon ends up grabbing the 2.7 device away from ftdi_s

jelly, http://rafb.net/p/KWeP3Z75.html

I've used dash as /bin/sh for quite a while and the startup is still rather slow.

is there any site to get sound working on debian? i'm using a vibra 128 card

How can i `run_init` to /dev/loop0 that is made of a cramfs file?

do you remember who produced the pretty graphs? was it marga?

hmm, I haven't actually seen any of those for Debian

relevant /proc/bus/usb/devices and dmesg stuff: http://pb.marttila.de/147

stew how i paste u list of alsa drivers?

I use something like if ping -c 2 -i 1 some-host; then [ it works ]; else [it doesn't]; fi

jelly: yeah, http://www.marga.com.ar/~marga/debian/boottime/debconf-boot.pdf
/msg dpkg pastebin

right
how i get em stew what comand make em show up?

so, your vpn gateway is in that very same range too. that means that after adding that route, you also need to add a separate route for the vpn gateway back thru 10.1.1.1

cool! the return value of ping is sane!

how do i get my driver list for sound card

run alsaconf

srv1.vpn.ufsc.br ip is 150.162.0.82, so after adding the route for the whole 150.162.0.0/16 thru ppp0, add another one like route add -host 150.162.0.82 gw 10.1.1.1

done that stew

does it detect your card?

no

Is there a shadow file for group and if so, where is it?

!tell erphan about root

they try pinging 150.162.0.82, it ought to work

[GNU]Order: gshadow

what did you mean by "my aslsa wont hold"

i have to run alsa conf each time i want to get sound
alsa store dont hold

but alsaconf doesn't detect your card?

think its sound blaster 16 bit

thanks.

YW

it returns false even if just one packet is lost, my real condition is a bit more complicated but you get the drift

Can 500-1000 TIME_WAITs for mysql port cause a load between 2-3? RAM is fine (no swapping), CPU utilization is 24%… I'm trying to narrow this down, only thing that looks strange is TIME_WAITs

thanks for the extra info, much appreciated

how come i can't tar -jxvf a pidgin.tar.bz2 file?
cannot exec: no such file or directory

what happens when you try?

stew that list pci device comand

aptitude install bzip2
stuart-: ^^
stuart-: not you

it comes up with to outer sound cards

lspci

and then it runs fine til reboot

stew, oo thanks!

does alsaconf detect your card? or does it return an error?

it dects minee

what driver does it say it is going to use for your card?

Do I only want the first cd if I need a barebones etch system?

yes
or possibly a netinst which is only ~150 mb

Ensoniq ES1371 [AudioPCI-97] (rev 08)
says my card is that
let me look at what is in case hole one
it doest tell me it does a scan
picks card

Hi

gtk+-2.0 = 1.3.7 gdk-x11-2.0 glib-2.0 gobject-2.0 but I can't find the correct package names…what are they?
(I also need any -dev lib of those)

use apt-cache search
apt-cache search gtk+-2.0
or just gtk+

cteaative labs ct 4810 card

when the card is working, run "lsmod | grep snd" and show us the output; use a pastebin: /msg dpkg pastebin

i'm trying to run "sudo apt-get build-dep gaim" to get dependencies for a pidgin package i'm trying to set up. but it says unable to find source package for gaim. any way to remedy that?

stuart-: apt-get build-dep pidgin

use pidgin o_O

stuart-: but why are you trying to build pidgin instead of just installing it?

stew, no pidgin source in my apt-get

stuart-: it's in lenny and sid

stew, do i need a special sources.list line?

Is a user with shell set to /sbin/nologin or a shell that exits right away, able to use scp?
remotely

hey.,. anybody know when the KDE4-beta packages will be available to debian?

stuart-: you can install it from backports.org: /msg dpkg bpo

this script from #alsa might help you get some info…. http://nopaste.com/p/aFihqMSkX/txt

ask #debian-kde on irc.debian.org

hmm okay..

proftpd only listens on ipv6 socket, how to get it working on ipv4 ?

what does "netstat -plant | grep :21" return?

how can I map the usb device number from dmesg (e.g. [88179.644000] usb 2-7: FTDI USB Serial Device converter now attached to ttyUSB0) to something in /proc/bus/usb/devices?

no, i don't think so, perhaps look at scponly

steve, tcp6 0 0

Comments

I know a friend of mine has to clear the cache about once a week on a site unfortunately he doesnt have any cron

Bobson - Did you say you use Form Builder?

yes - I have it running at the moment on 1.1
thanks kino
will learn how to make tags

Whats the cms_module code you use to display the form?

will go look

this is simple script you can fetch from db how many you have them, random which one you want and then fetch from db only the one you want

kino - thanks need to work out how to fetch with php from db… I am a total php newb, but its good to learn

http://dev.cmsmadesimple.org/search/?type_of_search=soft&words=quote&Search=Search

CMSMS Forge: Search (at dev.cmsmadesimple.org)

there are some project already running

cool
{cms_module module='FormBuilder' form='contact'}
if contact is the name of the form
thats what I use to call it Al3x_forum
brb - coughing up a lung

hmm
thats weird

I think that your new location of web's not met minimal requiments…

The gallery now works
but the forms dont

one question did you clear cache?
cmsinfo moce

i'm not following you…

cmsinfo move

move is clear cache, backup db, backup files, move files, restore db, edit config.php to reflect changes, clear cache

have you done all steps above?

cache cleared
same issue
I'll wait till this arvo
will be fine :0
Thanks anyway kino

how do you clear the cache?

cmsinfo clear cache

Global Settings, and press clear cache button.

cool didn't know about this way…

why do you need to clear rhe cache?
I mean I know you have to do it
and I know the reason is valid
but what is the reason
just so I know

statistics - / (rev: 63) silmarillion:

I know a friend of mine has to clear the cache about once a week on a site, unfortunately he doesnt have any cron access, know if there is anyway within cmsms to clear the cache periodically?

hmmm.. javascript
I'll have to use javascript if I want a style sheet switcher wont I

if you have only static pages you dont have to clear cache… I believe… but becase of modules you should…
if you have only static pages you dont have to clear cache… I believe… but becase of modules you should…
if you have only http://www.oxxus.net">static pages you dont have to clear cache… I believe… but becase of modules you should…

I have a question

so ask and dont tell us you have a question

You should clear the cache no matter what. Because smarty takes the templates (and content end up being smarty templates — it's a long story) and compiles them with paths and stuff in them.
@later tell calguy1000* Did you even look into the module manager script?

The operation succeeded.

cool
ok
my question is:
I have a two column layout - the right hand side I wish to be populated with links to articles.
the articles are to load in the left column
I'd like it so that as soon as a new article page is created - this right column is automatically updated with a link
is that possible?

Ted - sorry if my post on the calendar caused any grief, apologize for the inconvenience.

Is an article a page or a news article?

a page

No worries. The guy is acting weird.
I just didn't want to say that in the forum.

the guy is weird, and his product sucks

basically they have a load of case studies that they'd like to have pages for

but the basic framework is what i need built out from…

So what you do is use a menu…. but make it a menu of all child items from a particular parent. When you add a new Case Study, add it under the Case Studies parent and it'll automatically show up.
Does that make any sense?

yes.

I even took a dig at him when I wrote the reply…. Dude, it's free advertising.

thanks

Sure think
thing

Im sure I will mess it up - but at least I have an idea what I'm doing now. If I have any problems no doubt I'll pester you all again
sorry

No worries. Here all day.

thankyou

what time over there?

07.51

^is

re-working the template at the moment so its cleaner. Just hope it still works in ie 6 :P
its 12.51pm here

I know what time is in UK

you in uk?

but still messed up with transcontinental time
nope you I belive

I am yes

I have +1 from you

Yeah, 6 hours to the Eastern US. 9 to the west coast.
calguy is 8

I need another monitor.. may buy a Mac. I'm allowed to now

I belive arcansas is 7

Yup
Mac! Mac! Mac!

I find mac users a bore.. but, the new imacs are lovely
hehe

Ok, not really. We are a snide bunch of bastards.

lol

But the computers are still cool.

My mate Adam bought a mac, because he wanted a new machine, with an Amiga like flavour, that wasnt linux or windows.
He's turned into a beast
He was asked to leave PC world for being really rude about PC's

Started wearing hood sweatshirts and putting his nose just a wee bit higher up than normal?

calling their tech guys a bunch of M$ hores
yep
thats the one

my brother is about to start his another school day in arcansas…

Oh

*whores.. cant spell
sorry

M$ …(rude word) lol

Never been to Arkansas…

hehe M$ = kick ban

Ted. never been to eastern US
but loved LV, NV and LA, CA

It's… meh
You're not missing anything
Yeah, LA is something else. That's _SjG_'s town.

hopfully I will travel to US again next summer to pick up my brother and I\ll travel to eastern US
maybe… or I just stay in texas

Comments

A friend asked me to translate japanese text and comments from C source code I know a friend that knows japanese

types like uint_8 uint_16.. are defined somewhere ?

- you might want to peer into it.

and under windows ?

The popular Windows compilers do not support C99 yet.

when in doubt, have the person compiling it make the decision

so i have to do my own include ?

just use the c99 define, and make them supply it manually if they don't have c99

in C99, and on older POSIX systems uint8_t and friends are defined there.

inttypes.h doesnt exist under windows

just downloa dit

Mornin'

hiya

just provide your own inttypes.h, optionally

i need make something like progress bar i get with wget

keep move in the same line

Why, then, do you seem to be wasting valuble productive time on IRC, the trash of Internet garbage?

i can not do this with printf

you can

why don't you download wget and check the source?

you could use \r i.e.

I wonder if that is portable.

works forme

*NIX and win, yes

Existing characters printed might also not be cleared.

if you give such a shit about 'portable' don't bother programming

thanks readme

yup, i had some fun debug sessions with \r :p
"wtf is there 123f instead of 123?"

"portable" is another subjective issue. More is better, but at some point you go insane.

I still say progress bars are made with curses.

no
progress bars are made with OS specific GUI coding

or, relying on terminal-voodoo - which is the realm of curses and the terminal-drivers

Well, that's kind of curses's job.

progress bars are made from whatever you have available
just abstract it and be done with it

well, that was true back even in teletype-days. Which is what I was trying to tell 'em

and it will remain true in the days of squid oriented polyp interfaces

hmm, never saw those progress-bars ;-)

well, the stars are not yet right

oh Zhivago this is where you went after efnet

ah, a visitor from the punctuated country …

i guess

when i exit a program where i create/close a socket, then i try to re run it too fast it has a bind error
i seem to remember this can be fixed, but i dont recall how

you can set the tcp reuse option if you want

ah ok

SO_REUSEADDR

any reason why i shouldnt?

do what you want

there are some security issues.
if you use SO_REUSEADDR without a specific interface it may be possible for someone to open the same socket on a specific interface and use a man-in-the-middle attach
er, attack

ah

it's probably useful to allow the user to decide to use that feature or not

for some reason, it sounds like it'd work usefully with inetd and restarting server/services

that's one place that it is perhaps most dangerous
but those services can probably bind specific services
it became fameous when some people used it to attack nfs

infamous? ;-)

strtok(NULL, " \n" );

either will do

will that look for " " and \n
or either one

or

ah

Stop it

stop what?

hrm i cant find any thing in the man page about what i have to do with what strtok returns
is it allocating memory?
it returns a brand new string it seems

it is slicing up your string

oh
so its deliminting it with ?

it modifies your own string and returns pieces of it to you

does that means it ruins the copy its working on?

the manpage explains it quite well
assuredly ruined

yeah i misunderstood how it worked

strtok's design leaves much to be desired

makes it easy to parse the strings into a struct of char pointers tho

no
it just makes it easy to step on yer nutsack

strtok makes it easy to shoot into your foot

i assume all the strings will still exsist aslong as i dont free the buffer they came from?

ok, PoppaVics explanation is better

so is this ok
http://rafb.net/p/0RqrVx19.html
assuming request is a buffer of chars on the heap
my logic is aslong as the original buffer isnt freed the split strings will be fine like that, right?

if anyone accesses those strings w/o being aware of that use, you will want to cry. Otherwise, go mad baby
for internal uses you don't expose to users accidentally, you'd be ni worse off than normal
ni/no

im probably better off copying the buffer into new memory
and adding a pointer to it in the struct
i reckon
that way the original strings intact

I'm having a problem with feof not returning EOF if I have requested and read the exact file size from a file. It only returns EOF if I have tried reading *past* input

hrm valgrind isnt so good when youre working with many procs

is there a workaround?

I suppose you could read-up on your input-func.

that's how it works iirc
what's wrong in using the return value of the reading function as indiciation for EOF?

hello all..

hello to you too tikmo

i'm starting to learn c.. how can i compare 3 digits here to determine which is the largest? am i going to use the if statement here?
hello azi`..

three numbers or digits?

if I read a file in 100 byte chunks and the file is 200 bytes, it will return 100, 100, 0. But the problem is that I need to know if it's the last chunk *before* the file read is taking place

Hi

azi`: three numbers

so yes the best way is to use if

hmm maybe I could use fseek

maybe you might also change your loop and stop screwing with realloc or whatever

azi`: that means, i'll compare first the first two numbers then compare the output between the two to the third number?

I'm trying to use opendir(), im looking at the 'errors' in man opendir
one of them is [ENOTDIR], how do I check the error which is returned from a function?

i think i need to store somewhere the result of comparisons..

didnt know that Ahmadinejad was a C programmer

tombee, look into errno.

ok cheers

Now you know

ie. if (errno == ENOTDIR)

Ahmadinejad, you program the applications that manages the atomic bombs? :P

Hi, is this a string.h function to put each word of a string into an array of pointer

No, _i_ dont

no

We have enought on that project, now its time to new projects.

i would like to have a word array char**

write the code then

tombee, but errno's value is not valid if the previous function call did not indicate that an error happened.

fax should i use strchr strstr..

no
just set every space chr to a '', and save a pointer to the start of each word as you go

oh nice idea thanks

http://rafb.net/p/onB81l56.html
I've got that code which makes a linked list of a directory's recursive contents

fax last question, if i made IRC service software, the service need to send a string with second since 1970 how to strcpy(buf, "NICK ChanServ 1 0 TIME in second escription); and then send(…)

is it a common technique to store many strings in the same concurrent buffer and have them logicly seperated by 's ?

when I print it out, I want to be able to make a 'tree' out of the contents, how could I do that?

probably you wanna use sprintf

can i send to socket using sprintf?

no
format your data into a buffer, then send that buffer

it's not "socket print formatted"

Wulf_ i know
fax can you write it i really dont figure it out

hmm.. Well, except for the fact I can't /*coment*/ smack-dab inside a word being parsed, this is looking ok.

The time is %d", time());
send(sock, buf, strlen(buf), o);

can you use snprintf instead?

fax thanks a lot

No, they can't

may need a \r\n or \n in that.

ah yeah, thats true
don't forget \r\n like I did :p

rn for irc
yep

time() returns time_t not int, perhaps "%lu" and (unsigned long)time() would be safer?

I think using the format specifier %d for the return value of time() is not safe or portable.

heh

you should just hardcode the time for safety.

I don't care :p

efg

whats egg

evil-effin'-grin

pizza_ so i use %lu?

jhehe

come on, at some point the querant needs to open the books, manpages, etc

more simple with you
you all robots lol
fax thanks again

strftime(). :-)

or asctime
or N

system("cal")

or a substring and %s
or 3 dancing polar-bears

…. if you wanted seconds since epoch.

or, I can just fold The Chair - and bop some folks ;-
(try the veal, I'll be here all week)

how can I include a file as a symbol inside an object file?
that looks like a char array in C?

oh, please

What

as in, turn an arbitrary file into an elf file that I can link with C code

write a program tht outputs C code with an array that contains the file.

I assume you're wanting to know how people embed things like PNG images into their programs …

Well, not images, anything
I think there was a program called bin2elf somewhere

you'd have to write a program to do that for you, there's no 'generic' one I can think of, but it should be pretty easy to do
just realize if you're trying to be portable, using something like bin2elf probably isn't the best host way to go

should I open my sqlite database at the beginning of the program and close it at the end
or keep it closed if I am not doing something in it

[Yarcanox]: it depends of how many connections your program will have in a range of time

not many avenger_
I am asking because of data loss
what happens if the program crashes while the database is opened?
(but no transaction running)
is there any difference to the situation as if the database would be closed?

[Yarcanox]: a corruption will happen only if a data manipulation is running on the moment de server crash

that is not related to C.

ok thank you avenger_
I am very sorry auris-, I asked here because I thought someone here might know it
and it takes hours till someone answers in #sqlite …

Auris-: it'd relate if they needed to sync, flush, whatever - as it is: it's just mumbling.

As you said, if something breaks, the transaction is rolled-back.

ah rolled back?
hm

life sucks, this ain't #sqlite-alt

that sounds good..
hi poppavic
nice to see you

Yarcanox, that is what transactions are for. Unless you commit it, every failure leads to a roll-back

hopefully nothing should happen if a program crashes
transactions beat locks

I like sql

Yarcanox, so, corruption basically cannot happen.
(in theory)

sql is pretty terrible
but transactions are quite good

why is it terrible zhivago

because it doesn't do relations properly

I guess sqlite is still better than just dumb textfile-saving without any indexing or sorting

and the syntax is insane

but what happens if the crash happen at the moment of a "commit"

I heard sqlite was more robust than mysql…

then the commit either succeeds or fails.

Avenger_, commits are atomic

why do you care?

SamB sqlite is perfect for an IRCbot with much data but not incredibly much

because he knows it will happen sooner or later ;-)

no, I mean — why does he care if it succeeds or fails?

in a real implementation, it either happens before the commit is finished or after

if it failed, then he does it next time, if it succeeds, then he doesn't

oh, that

just set your stuff up to be idempotent where possible

yes
exactly

The more I hear of transactions, the more I think that they end up being like a serial database/log of who-connected/when, and whom did what - chrono-ordered.

a bad way of thinking about it.

PoppaVic, it is. It's a log filled with actions to perform. At "commit" they are performed. Failing means rolling backwards through the log undoing changes up to the last commit-point.

a transaction is essentially a predicate and a modification
the predicate contains your expectations about the universe which are relevant to the actions being correct.
if the universe changes such that the predicate fails, then the transaction is aborted.
think of it as being "do this, but only if my expectations of reality are true"

well, the issue I'd have with it is if A asks for a', and Z asks for a' - and both are planning to change a' but neither know the other are.
ahhh

I've covered that with the above.

I see it, thanks

there's a model for inter-thread communication called transactional memory based on the same concept as database transactions…

A reads V as 10, then say "if V is 10, change V to 11"

got it

STM
also works across networks, etc.
at an extreme, it becomes a mobile agent

I can see that now.. So, it's an expansion of what I said. I like that.

the best thing is that transactions compose, but locks don't
yeah

Transactional Memory in 15-20 minutes: http://blip.tv/file/317758/

Well, it offloads some (appropriate) responsibility to the server, and forces each client to commo properly.

zhivago I guess it's something like a FS journaling function

he really does tell you almost everything about transactional memory in that short talk

with these atomic transactions

exactly

I'd bet it would really be frustrating to many clients that wanted to make their own change to V@10.

except that database transactions can be a bit bigger ;-)

PoppaVic, that's no problem. If it failes, it's retried. And eventually will succeed.

yes, I realize that - but his poor ol' client is gonna' be loopin for awhile ;-)

but it doesn't help much to have transactions in your database mysql or postgresql hosting server if it's stored on a FAT or ext2 (or other unsafe filesystem) volume

or, allow some folks direct-access to the files, anyway

not that you need a server

it will run on a freebsd machine I think
have ext3 here on my local machine, but that won't be the final platform…

as long as it's not the kind that you need to run scandisk/fsck on reboot for

heh

why can't you build transactional storage on top of unsafe filesystem?

well, ext3 is really lousy for writes

freebsd is intended to be secure, I don't think that it will install an insecure filesystem per default

and a journalling file-system actually tends to hurt safe transactions

hrnt because you can lose data then
although you got that atomic transactions

well, if your filesystem is unrecoverably corrupted, your database is surely gone

because you can't know when the underlying file-system will roll back to

that can happen with safe filesystems too

well, you have synchronization, and direct opens, and so on

SamB, transactions on that high-level prevent manipulation corruptions. The prevention of physical corruption is the duty of such a filesystem. That hasn't a thing to do with each other.

FreeBSD will only use the ufs family by default

but, they do.
let's imagine that database A says to client B that its transaction has gone through.

all I mean is that a safe high level won't help you much when you combine an unsafe low-level and a power outage ;-)

now imagine that the machine that A is on reboots unexpectedly, rolling the file-system back to before C's commit.
B is now completely stuffed.

you can implement the safety in that transactional storage.

Zhivago, then the timestamps are out of order which would trigger a roll-back

and this is why databases can't take advantage of journaling file-systems
you didn't read what I wrote.

Then why did machine A say the transaction went through?
when it didn't

it did go through.

Then it's written
so if it reboots, fine

the machine rebooted immediately after it went through

Then it's written

and the journaling file-system discarded it …
think …

Why did it discard it?

Zhivago, ok, read.

Transaction complete = written to disk

Agreed.

so what?

Journaling file system has now forgotten it entirely

do you know how a journaling file-system works?
wrong, you idiot.

So it can reboot until it's blue in the face

shouldn't the database have waited until the transaction had made it into the joural?

data is still there
and accounted for

the journaling file-system wrote it into the goddamned journal.

flux

the goddamned journal was discarded back to the last checkpoint on reboot

and checkpointed?

so your goddamned transaction data got thrown away

"reflux"

Heh

are you bloody well reading this?

No, it wrote it to DISK

where do you think the goddamned journal is?

Heh, pointless

writing something to disk doesn't give you consistency, which is what a check-point requires.

:P

zhiv doesn't have tab completion

I was under the impression that the point of the journal was to play it back after a power failure or other unclean shutdown?

only to the last point of consistency

Journals don't discard anything, and journaling file systems also vary wildly in their methods
So your sweeping generalizations are well.. retarded

you're obviously wrong.
what happens to data that was partially written into the journal?

UFS2 + softupdates = nothing like ext3

at a minimum that partially written data is discarded.

It gets lost, and the transaction was never confirmed
that's what happened

so, you admit that you lied.

No, you said the transaction went through
Do you speak english?
or are you checking a book as you go?

what does the transaction going through have to do with the journal being checkpointed?
think about it.
why would I have to read a book for this trivial nonsense?
who is suggesting that?

it's not necessarily an improvement, but at least you know something is wrong if you lose the entire database.

well, if you use FAT you stand a good chance of losing the whole thing…

the fun part is.. All this still reminds me that N folks can write diff-files for Z files and patch can still get screwed merging all the patches back in ;-)

if you lose a single transaction, there is a risk that nobody notices anything

data is silently corrupted

You're the one bringing up transactions, not me

true

You're takling circles around yourself

I think you're projecting here.

Yeah, definitely not a native speaker

but anyway… why the heck would the DBMS say the transaction had completed when it hadn't?

why are you confusing database transactions with journaling?
well, it did complete.

I'm not

then the machine went back in time a bit

now imagine that the machine that A is on reboots unexpectedly, rolling the file-system back to before C's commit.

okay.

YOU brought up transactions

SamB, dbms can't know what the filesystem is doing.

so, why are you confusing transactions with journaling?

Auris-: it can't?

well, no — it's more that the dbms _must_ know what the filesystem is doing.

I'm not
you're just confused

otherwise it can't make a claim like that, which is what makes journaling a useless overhead.

well I don't see why it can't find out with a journaling filesystem…

SamB, how do you ask a filesystem if it is all synced up? well there is fsync() which may or may not work.

please try to speak coherently
well, it can, but in order to do that, it has to force a checkpoint to happen after each transaction.

I suppose so
but there's always the other issue

at which point … what advantage is your journal giving you, other than making the system twice as slow?

the transaction completes, but the notification never makes it to the client

Zhivago, please learn to understand english?

hmm.
not as bad, though

that's less problematic, since the client knows that it didn't get a confirmation

the client could find out about it after the connection goes back up…

I understand english — please try to speak coherently.

Obviously not, your responses have nothing to do with what I said

please stop babbling.

couldn't the DBMS alternatively just wait until the next checkpoint to confirm the transaction?

Quit spouting nonsense?

The easiest thing for you to do would be shut up

well, that might well deadlock the system
yes, please

1) It's safer to work on a dup than the master; 2) You almost need a special FS and partition for the db transactions.

well, the db needs to be aware of the underlying system in order to manage its transactions effectively

yeah, I can see that

filesystem "begin work" (checkpoint) - database "begin work" - database write - database hosting commit and notify - filesystem commit (journal checkpoint)

ok — you can't nest transactions — you need to compose them.

the problem is that between database commit and notify and filesystem commit we might have power failure, and so filesystem rolls back

so, whatever happens the db needs to make sure that its changes get down to persistent media.
and at that point it means that you're duplicating and/or bypassing whatever a journalling file-system does, so …

….but the database does not, and other clients might have seen the change

well, I'm starting on unboxed variables now.
does not what?

good luck

hmm.

sounds like you arel making major improvements to ecl

roll back, because it already committed (and for example notified the client that the transaction was ok)

looks to be pretty straight-forward — I'm looking forward to unboxed closures

well, it depends how often these "points of consistency" come

well, if it nailed the file data down first …
using fsync or O_DIRECT or whatever
then it can know
I'd generally recommend O_DIRECT

hrnt, db commit shouldn't return until the os (fs) has confirmed that write went through.

well, I guess you don't need the journaling for the DB stuff

that's my basic thesis

yep
the whole filesystem is already a form of database.

I'm tired of writing C code in C

but it's nice to be sure that you don't lose the whole directory tree
;-)

Auris-: yup

well, if you really cared, you'd be using a partition

amen

I don't blame you Zhivago, It's not ideal

and prolly on a duplicate

ecl plus gcc look like they might do a reasonably half-arsed job

It's interesting how lisp can morph like this
e.g into HTML, javascript, C .. etc
forth

could work, but a journaled fs which you bypass for the database files sounds good enough to me

sure.
as long as you're not creating or deleting files

what I still find to be the biggest clue is that most lisps I see mentioned bypass the std compiler/linker stuff

I don't imagine you're allowed to bypass the journaling then ;-)

is it guaranteed that with O_DIRECT or fsync(), the data is nailed down well enough, so for example a journaling filesystem won't discard it?

in this case, we use C linkage

bypass or extend?

nice

not sure what oyu mean exactly

well, O_DIRECT should be, but a given file-system doesn't need to
bypass as in write to the physical media underlying a file rather than to a journal
well, convenient, and also a pain in the arse

he means some lisps, I think

I think he means journalling and O_DIRECT

oh ok

ah, sorry.

sure. The more complicated a tool, the more irksome the interface/maint.

the idea with O_DIRECT is "just write the damn data to the disk already!"

it's all part of my plot to subvert CL.

A hammer is just a hammer; a nailgun hides a hammer in a monster API and you have to maintain all that mess - plus the hammer

Subvert how? :p

hey
does this archivist make archives?
^^
I mean, logs?

by writing a compiler on top of it
what archivist?

nickname archivist

Yarcanox, no.

Yarcanox, yes

he/it seems to be in all popular supportchannels

there are two compilers that I'd like to see — one is a CL-CL compiler.

heh

ok not a bot thn
archivist really?

CL compiler

hell no

you're doing logs?
just watned to know

ah, lag.

"CL" being "common lisp"?

yeah

ok

BUGS
"The thing that has always disturbed me about O_DIRECT is that the
whole interface is just stupid, and was probably designed by a deranged
monkey on some serious mind-controlling substances." — Linus

would it compile it down to lower-level Common Lisp or something?

im sorry what does a CL-CL compile do? (other than absolutely nothing)

please discuss ;-)

o.O

pass - I never much care ;-)

it compiles good CL into bad CL.

it told me I would be spamming :?
what the hell

Yarcanox, I work as a volunteer in an archive

You may speak again. = very nice :/

e.g., type inference to insert declarations, reducing it down to a small number of primitives, and so on.

whatever for?
oh

ahh

because there are a bunch of CL implementations around, but their compilers tend to suck a bit.

You've already worked in that direction, hHaven't you

opti-foo ;-)

can it compile bad CL into worse CL?

that's racist, just because I type fast

but, given a portable CL compiler, you could do the clever stuff portably.

depends how bad it starts out, probably ;-P

well, if by worse you mean 'unreadable'

except for the clever stuff that isn't portable

I'm hacking some game engine to add some functionalitly, however, i'm not used to reading C-Gcc errors, so what does this mean please ? "snd_ras3d.h:55: error: expected =, ,, ;, asm or __attribute__ before * token"

well, you have a high level and a low level.
I figure that if a CL implementation can handle the low level, then portable code can handle the high.

typedef const snd_buffer_t* (*snd_fetcher_getsb_t) (void *sfxfetcher, void **chfetcherpointer, unsigned int *start, unsigned int nbsampleframes); = Line 55

ripping off SBCL's compiler for the high level, might be suitable

qantourisc this means it expects something different before your * char

but then I think about the debugger, and don't bother.

(*snd_fetcher_getsb_t

ok thanks

there seems to be the value type missing :?

what is wrong with that?

aslong as i know that the error is "some type you are using here is not know"

hrnt as I said, it is just a guess

somehow, your CL:CL tool sounds a lot like I remember in the early days of forths. ;-)

look at the previous line.

in particular - when they began "metacompiling" New forths from forths in forth ;-)

what z said, and also check that "snd_buffer_t" is visible.
(the typedef for it, that is)

aa yes searching for snd_buffer_t turns up a struct i ferot to copy
thx

hm bad guess… I should leave the error searching to you :P

good.

that was intent, yes?

something similar

I mean - as opposed to merely changing the current environment in the running CL ?

well, there isn't an enormous amount of difference
but probably

hmm.. well, except - yeah.. lisps tend to be interactive already, so the latter is usually alread a given.
already
ok, Sounds like you've a plan with a lot of work ahead.

hrm my webserver is working fine, spawning procs and not leaking mem
but thwen when i increase the server host load some procs start randomly not freeing
or something weird like that
which shouldnt happen since theyre all identical with identical requests

chances are the lock up and become zombies - use more salt.
theY

salt?

Yes. Always salt a zombie to kill it dead-dead

==16549== at 0×40217B8: malloc (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so)
==16549== by 0×8048AD5: header_alloc (in /home/daniel/work/sshttpd/a.out)
==16549== by 0×8048B89: httpparse (in /home/daniel/work/sshttpd/a.out)
==16549== by 0×80487AE: handleclient (in /home/daniel/work/sshttpd/a.out)
thats weird, the end point is malloc
maybe malloc is failing?

race conditions :-(

hehehe

not threading, forking

hmm, malloc can fail yes

oink-oink

you're probably overwriting some buffer and screwing up malloc's internal stuff

that too

pizza but it only happens randomly in random procs

but usually valgrind mentions that doesn't it?

all with identical requests
so same thing is happening every connection, yet some leak

last time I saw his code, he was throwing long func-calls in if/while and didn't bother to test malloc anyway

what does valgrind say killed it?

are you showing a valgrind leak, or is valgrind catching a bad read/write ?

one sec ill re run

the line above would probably be helpful

if/while _conditions_ (sorry)

oh wait some are stil reachable
http://rafb.net/p/mCRUlu16.html
theres a random example
its like the proc is closing before it can free
i mean i guess its not a big deal, they just become unreachable they dont leak

i'm pretty sure you should not concern yourself with 'still reachable'
think it just means that brk() wasn't called to release the memory back to the OS … but that's just internal to your OS's malloc()/free() calls afaik

hello

i see

these days mmap() tends to be more popular than (s)brk.
reachability of memory doesn't have to do with that, though.
what it means is that there were pointers to those allocated regions at the point of exit.
the good news is that this means that they weren't leaks - the bad new is that you were too lazy to clean up

it only happens when my app is under heavy load
and theres no chance of it being race conditions since i fork

eh, valgrind's faq words the 'still reachable' differently, but I guess they're talking about C++ and STL mainly

if the forked processes are hung, there is something they expect to see/do that they can't

don't remember seeing him say anything was hung

well, he said they were laying around and wouldn't finish, cleanup, exit clean - but only some

http://mateos.ath.cx/~daniel/sshttpd/
if anyone cares to look at the code

but thwen when i increase the server load some procs start randomly not freeing

my test is i just hit refresh on my browser

think you read 'freeing' as 'freezing'

and it works fine till i increase my refresh speed

must have a race condition

are you freezing somewhere, or was your leak your only concern?

"freeing" - and "randomly". Yeah, brad_mssw could be - "freeing" what? The mem they use? forks have their own mem.

only the leak

eh you here?

It sounds to me like he isn't testing something, or he's mulching the stack or some other silly thing.

lubos,
yeah

if valgrind is only showing 'still reachable' messages, don't worry about it at all, nothing wrong with your code, everything is being freed, suppress 'still reachable' messages in valgrind since you don't need to see them, it's confusing you

one weird thing is either all frees work, or they all fail
but all the mallocs work
it says malloc 3/ free 0

it seems that glDrawPixels() is even slower that glPOINTS heh

i have 3 destinct points where i malloc
its weird all three diffrent free requests fail

implement an xmalloc and xfree, to write to a log. Check the log after valgrind.

not in http://rafb.net/p/mCRUlu16.html that you pasted before
malloc/free: 3 allocs, 3 frees, 1,466 bytes allocated.

==16627== malloc/free: 3 allocs, 0 frees, 1,385 bytes allocated.
above it

but both are slow when drawing 1024×768

oh, you're showing multiple children

yeah

yeppers

it's showing nothing lost in your leak summary though
for that same process

isnt that weird in it self?
if theres 3 mallocs and no frees, there should be leak
oh wait no, thats the same summary

dman_, why doesn't your child process actually exit?

add -g -Wall to your command-line

oh, hmm. maybe it does.

then valgrind would tell you where the "leackage" begins

lubos, what about using textures?

will show you line numbers

-g doesnt work

dman_, I'd rather do exit() in the child if-branch instead of break.

what compiler are you using?

Auris-: ah yeah i will

Auris-: xexit() (and the logfile again ;-)

its stupid but for drawing all i want to use is drawPixel().. and make everything from scratch..i found a pixelToaster libraly but its c++

apparantly the most popular NT example driver to copy off of is a toaster driver…

NT? oh.. m$ NT?

lubos, well, why can't you create a texture with drawPixel then use it on a QUAD covering the whole viewport?
I don't know which is faster… try to benchmark it

http://www.electric-escape.net/node/1475 This has to be the funniest IRC chat i've seen……lol

Stop it
No one cares, just stop spamming

faster probably involves either (a) not manipulating pixels individually or (b) knowing the format of the image you are manipulating…
oh, btw, do you have shaders on your card?

thx i will try it but i read that glDrawPixels() is overally slow bacause its in processor memory

dman_, that problem() function of yours seems fishy. you use it in other paces too but it is defined only in main.c and there are no prototypes anywhere for it.
s/paces/places/

not at this one, but have it on next pc why?

compiling with warnings should have caught that.

lubos, checkout if you can use FBO then

Auris-: is in main.h

well I thought you could maybe consider writing a shader program

dman_, parser.c does not include main.h.

oh
still
that wouldnt cause the prob

why not?

shaders are applied over on already drawen scene?

dman_, problem takes two arguments but in parse.c you call it with just one.

Auris-: true, fixed it
still leaks

hehehe

I believe they can be, but there's certainly no reason they'd need to be…

if you ever run into a problem, random stuff happens.

that was never ran or it prob would have crashed

nope.

Auris-: the prob isn't random, his args are ;-)

dman_, it would have taken random values from stack.
well yeah.

oh

what are you trying to do, just draw 2D points on the screen with opengl?

dman_, why is there no makefile?

well fixed and it still is leaking

ok thank you, adding it to list

Auris-: havent made one yet, i just use cc *.c atm

yes, drawing pixels quickly

you're not compiling with any warnings, are you?

dman_, that's just moronic.

nah

and you have tried glBegin(GL_POINTS) and it was too slow?

well, if you intend to shoot yourself in the foot, have fun.

this is why I got tired of trying to help him days ago ;-)

I mean the prototype for problem() is in main.h, but not included in parser.c … but you're still calling it … that's rather bad

yeah that was my bad, but i fixed it

i'm sure you're doing that all over the place

it was just somthing i forgot to do when i added it

not that slow but in every glBegin was only 1 call to draw() eh..yes i was stupid ..

dman_, if you weren't being a lazy moron, and used a makefile which compiles the project with warnings turned on, gcc would have told you about your problems.

so it will be faster when i will draw 10000 points at one glBegin() ?

-pedantic only gives comment warnings

it should be, at least

ok thanks you

there ought to be a list of too lazy people available.

how do these flags work for you : -std=c99 -pedantic -W -Wall -Waggregate-return -Wbad-function-cast -Wcast-align -Wcast-qual -Wdisabled-optimization -Wendif-labels -Wfloat-equal -Winline -Wmissing-declarations -Wmissing-prototypes -Wnested-externs -Wpointer-arith -Wredundant-decls -Wshadow -Wsign-compare -Wstrict-prototypes -Wundef -Wwrite-strings

hrm few errors
well warnings
i should use a makefile, i need to find an example tho

Auris-: !warnings (I think)
My makefile has an OORT var of them, but I usually live with a lot fewer

PoppaVic, what, candice warns us about lazy people? :-)

Auris-: oh, sorry - misread: meant the warnings-list
Yeah, the "lazy" thing is why I peeve folks when I stop answering ;-)

PoppaVic, !warnings gives you a link to the wiki now, where the warning option list is.

Auris-: yes, better that way - a lot of the original warnings were specific to (x86?)

!warnings

warnings is http://www.iso-9899.info/wiki/WarningFlags

All the last months additions I've made to candide are all urls, now.. Easier on the screen.

!reading

reading is READING HURTS

it can

ah, much faster, but GL_POINTS arent pixels, when i will use 1024×768 they are too big

(I hate books with tiny print in bad fonts with horrendous sentences and paragraphs)
otoh, I can't play video-games shrug

can't
why not
trouble distinguishing relaity from fantasy?

his fingers always press two buttons at once
only good at mortal combat shrug

the funky controls kids get today are just flakey-weird, and I don't care enough to learn to train my hands to play a game.
uthe funky controls kids get today are just flakey-weird, and I don't care enough to learn to train my hands to play a game./u
I'd rather read, write, type, code, reload, shoot - or build something with other tools.

hi

whats the best way to convert an float to a char?
atof?

a float is not a char, it can't be.

a float to a string….

what's a "string"?

to a char array..

print it.

atof is not working the best for me right now.

use *printf
geezus

SORRY
a char array to a float

you are one seriously whacked lil' dude

sscanf?

strtod.

I'd bet you are also still sans book and reading

atof should do the job and on the man pages it says it returns an double

Auris-: strtod or strtof?

ascii-to-float()

sorry guys…i have a little problem..

'd' for "double"

i get this error in my little program

and recommendation for small and open source project i can join

../client terminated
Aborted (core dumped)

does strtof: used the same as atof?

what does it meaning?

/bad/ llama?

include/parse.h:44: warning: function declaration isn't a prototype is one warnign im getting
http_t *header_alloc();
looks like a prototype to me

zabin, read the documentation.

(void)
otherwise, specify (args..whatever)

hi pop

() tells the compiler "shut up"
hi Cin ;-)

matt__, I'd use doubles, you can do what you want.

this is the code http://rafb.net/p/7HKlrl31.html

Hello

i got it to work but its give me a core dumb after after words..?

what about main.c:68: warning: passing argument 1 of 'event' discards qualifiers from pointer target type
its event(char*) and im calling it like event("test");

http://rafb.net/p/8b40WW44.html

welcome to learning what the errors/warnings mean. Examine the actual function and discover.

A friend asked me to translate japanese text and comments from C source code. I know a friend that knows japanese but I have no clue about C, any possibility to analyze the text with some script or program and that says what line of code in each file has japanese text?

/* comments */
"text literals"

grep for utf-8.
unless they aren't in utf-8.

Auris-: human-translations ;-)

do you know what i get a seg fault here?
http://rafb.net/p/8b40WW44.html

honestly, I don't care. I'm sorry, but I've got other issues

zabin, unless you start compiling with warnings on, don't bother.

what lib is strtof in?

negative-feedback education has been useful for millenia.

!book zabin

There are text on menues too, not just comments
Auris-: It seems is some japanese encoding

im getting strtof.c:10: warning: implicit declaration of function strtof

#include string.h

Auris-: http://www.dcemu.co.uk/vbulletin/showthread.php?t=21136

i did that

Auris-: pc88quasidc.rar

stdlib.h?

i put that in to

#define _XOPEN_SOURCE 600 /* or #define _ISOC99_SOURCE */
just rtfm

i have string.h , stdlib.h, and stdio.h
and ctype.h

hi, i am studing the C language. can i see runtime my memory usage to track memory leaks?

Auris-: He did put a lot of japanese text, some text was translated using amikai :P

use some memory debugger

Wulf_, i need only numbers, is there way to print current memory usage?

apastinen, use i.e. 'valgrind' to search for memory leaks

c itself can't do it

tont, it means that you did not respect the bounds of your arrays and overwrote stuff you are not allowed to touch.

Does anyone know what this means? : strtof.c:10: warning: implicit declaration of function strtof"

Wulf_, ok thanks
triton, thanks

#define _XOPEN_SOURCE 600 /* or #define _ISOC99_SOURCE */

Auris-: Anyways, thanks for your response

some malloc.h implementations will support reporting/checking - other folks use valgrind. Others use gdb, and others check it themselves.

Thanks to you too

that didnt work

PoppaVic, Ok, thank you. i think i will try valgrind

put #define _ISOC99_SOURCE as the very first line of your program

It means you did not include the header file in which strtof is prototyped.

whatever order, whatever platform.
it means he didn't read the manpage ;-)

i include stdilb.h

timofonic, just use grep to look for non-ASCII characters. grep -P '[\200-\377]' file

strtof is not on my system, so it must require some special tender loving care.

just did that it still didnt fix it

which OS / c environment?

prolly wants atof

value = strtof(test)
linux ubnutu

btw. is ansi C reference somewhere to free to read?

!books apastinen

books is http://www.iso-9899.info/wiki/Books

i mean specification
candide, thanks

Think nothing of it, apastinen

apastinen, read ttopic, lookup url.

Auris, next time

http://rafb.net/p/COnNel46.html

thanks, and bye

there is a free pdf of the old c99 spec shrug

rtfm

I guess I suggested strtod because it is C89, and strtof isn't.

float strtof(const char *nptr, char **endptr);
i.e. this function requires TWO parameters

Auris-: well, strtod is a better choice anyway

what would i use for my second parameter?

Let them return a float if the range is right, or burn in hell.

rtfm

im not using points or anything fro taht
Wulf_:whats rtfm?

read the f***ing manual

heh

Floats are really only handy, imho, when they are mucking about with int-and-under clusterfucks.

Auris-: Uhm
Auris-: But it can say the line number and what file? Or is more difficult to do?

timofonic, I suggest you read grep's manpage.
or try it.

Auris-: Ok, thanks

!rtfm zabin

Auris-: Anyway, it seems BSD grep is different
Auris-: Thanks for all

timofonic, probably. but some tool will be able to do that. write a perl script.

I'm abivalent if parsers/interpreters should allow long and double input or not.. It sometimes seems safer to use long and double for internal use and scale back into int/float sizes - or chuck errors/warnings around.

input ought to be robust.

!rtfm timofonic

I agree. otoh, I find "scream and die" a better form of "robust" than "well, we'll just assume and change shit w/o telling you"
I like it ;-)

PoppaVic, I like robust to mean that the program won't crash and is able to complain smartly.

Auris-: I believe "validation" and "lexing" are a form of "robust", too.
OK, THAT - well… Shit can still happen that suggests death
ok/oh
Auris-: I tend to agree

….

Auris-: basically, what you mean is that abort() is likely a bad aolution.

PoppaVic, is that a liquid?

it's fluid, anyway ;-)
"solution" /homer

PoppaVic, sometimes it is appropriate but you still need to complain smartly before aborting. :-)

Auris-: I sure as hell agree
Unless yer dicking with silly lil' testbeds
brb

Auris-: thanks for all
Bye

np

Auris-: I'll try to learn more about grep
bye

Auris-: and, with a silly testbed, you can check returns or gdb, etc - so folks can either: 1) learn something or 2) get sloppy-stupid

hi, in this statement char* temp = "hello" is there an implicit null terminator?

yep

okay good
thanks

"string literals always add a "

PoppaVic, smart complaints don't hurt in testing either. I'm quite annoyed with asserts like assert(a == 1); which won't tell you what the value actually was without looking with a debugger.

Auris-: I never really liked assert
I always felt it just sorta' didn't perform well.
Auris-: and, I _really_ didn't like the idea that it was DEBUG conditional.. Either it matters, or it didn't.
I'd rather write code that paranoid-tests, then someone-trusts.

I like __LINE__ and __FILE__ and gcc's __FUNCTION__

if you make your own version, you can have debug asserts and product asserts.

yeppers
I'd like to see a nice lil' header-plus for CYA that didn't just disappear if DEBUG ain't defined.

CYA?

or, at least - some book we could recommend.
cover-yer-ass

lol

"I am getting serious here, these conditions must all be true or I scream and die on your ass"

hehe, yeah, screw graceful handling

What folks often do not realize is that some funcs can live with cascading failures. Some can't

ive been messing around, it seems the frees() only dont happen when i use send()
on the client socket
even if they dont even use the data that the free unallocs

WOW!

One of the reasons I always liked forth and cooking for comparing to "learning to program" was how they could cross-relate and remain "graceful"
well, "scream and die" simply means you get the hell out of wherever you were that was about to explode.. If you want to do a message, prompt, cleanup, sync - whatever - you sure can.
at some point someone is supposed to listen, learn and resolve.

hell, still with that?

When you have a branch with two options, one of which must occur, do you use if(cond1) {..} else if (cond2) {…} else { abort();} or if(cond1) { …} else { …} ?

if(state) …; everyone;
only you know how flow is supposed to proceed. if() is just a test for a resolution there or "else"where

Well, right. It's a style question.

whatever floats your boat - we can laugh later.

Generally, is it better to add that extra else to catch some odd bugs, or to just use a catch-all else and put an assert on top?
Ah.

Cin likes folks really handling-an-error. How you do that is moot.
if(false) {…}\n/*we passed - rock-on */
Other times I'd cascade if()..else if()…else if()…else…Failure.
You want readability, simplicity, and modularity.

PoppaVic, in C? no chance

oh, please. Folks need to work at it, but it is doable.
and, if they are in school anyway - let them work it.
this is /not/ to say that can't write cleaner/faster in other languages.

my solution for now is food.

Auris-: eat well ;-)

oh certainly, proper design being more important than implementation language, perl excluded.

well..
perl just bugs me.. It's mutated and starts looking like C plus sh on psychedelics.
Folks use it for strange things.

how can I pass a function as a parameter of a function, please?
I mean… I want to write a function that has got a function as one of its parameters

you specify the parameter as a function pointer of appropriate type

I think the huge issues began when folks left the simplicities of BIOS and .com for .exe, and then up further into drivers, then daemons, then further up into processes, and threads, etc

i believe processes existed before BIOS :/

The terms started to get messy, and access started to diverge.
only in ugly ways.

int foo ( int(*func)(int par) ); int blah(int bar); foo(blah)

thanks. Thanks for specifying Wulf_
and then, how can I call the function func() in foo()?

I'd debate the whole mess over a pitcher, but that isn't much use to us - except as lube with chatting and chow ;-)

func(123)

oh yes, of course! (I did a stupid question). Thank you

Do you think it's funny that we're re-inventing the kernel/process paradigm with hypervisors and virtual machines?
Only with much lower performance.

I find many things both funny and sad - and with no end in sight.
If I can't chuckle a little, I just want to cry.

np. and happy coding

I guess you just laugh when it's somebody else's problem.

thank you

not really.. I do often laugh at what they see is a "problem"

Oh! I didn't mean that — that was the general "you".
But yeah.

well, more often than not - folks create "problems" because they won't/can't think, read or test.
But, yeah - in some cases you get to that "this always blows" point.
I believe freebsd and *nix all started out to fix such issues, but got confuzzled along the way. I can't see a better way to descibe it.
describe

hi
audio/libavcodec/dsputil.h:485: error: static declaration of lrintf follows non-static declaration

lo

mean
follows non-static declaration.. but where?
I can't find it

lrintf? maybe you mean fprintf or something?

static inline long int lrintf(float x)
nope
hm.. long int
i'm on 64bit

what's the full error message?

I'm not on drugs.

In file included from audio/acodec/acodec.c:31:
audio/libavcodec/dsputil.h:485: error: static declaration of lrintf follows non-static declaration

I'm on osx ;-)

it doesn't show you the previous declaration?!

PoppaVic, ;-)
Wulf_, nope.. :-/

use grep to find it yourself (:
or.. remove the function. my system has a manpage for this function ^^

hm!
oh
then thats the problem
Wulf_, nice.. you are the man
/tmp/ccXmQwM9.s: Assembler messages:
/tmp/ccXmQwM9.s:24: Error: suffix or operands invalid for `pop'
ouch

I need a constant pointer to function that returns a constant pointer to an int

that looks fubar

cut and paste will kill you ;-

but i can't seem to find the syntax

dang.

you could just declare your int with global/file scope? :-)

it's inside a struct … (wel ok, it's in C++, but that only removes the const from the function pointer, so the syntax is very near)

I don't understand your problem, sorry

ok i'll try to write it
const int* ( *fptr) (const int, const int)
but that syntax is wrong

you can NOT equate C++ and C - I wish you silly fuckers would get that straight and stop bothering ##C
C++ stole some WORDS and then proceeded to change the universe and meanings.

anyone here can give me some idea about concepts of performance and scalability ?

already tried wikipedia on this?

performance is overrated, aim for correctness first and then, if it works perfectly but is too slow, worry about performance

require a lot more reading than IRC will give you

And, frankly, asking in ##C for those issues is stupid.

PoppaVic, I don't think so
C is known for its performance isn't it ?

C is known for its potential for performance, i'd say heh.

and for it's potential for foot-shooting

C does nothing for scalability though.

well, it's like asking ##C about "asm"

"C combines all the power of assembly language with all the ease of use of assembly language."

o rly?

build/temp.linux-x86_64-2.5/video/libavcodec/dsputil.o: relocation R_X86_64_PC32 against `put_pixels4_x2_c' can not be used when making a shared object; recompile with -fPIC

so says someone

tut, no luck , no one is willing to talk to me ..

hm.. thats weird.. I have compiled it with -fPIC

yeah, some folks believe it.

I'd s/all/most of/

performance is a matter of the right algorithm in first place

pizza, asm is not of my interest

you miss the point

doesn't matter if you use asm, c or java for this

All it means to me is "C is more portable than that 'asm' thing"

Wulf_, I get that…I need something concrete….how do I know if I have concepts of performance ?

you learn and test and practice and goto SOL.

ever heard of complexity theory?

i shaving the knowledge of time complexity enough ?

algorithms are much more important than implementation language

ya.. so can I say that if I know O(log n) is faster than O(n) then, I have the concept ?

that is a start

or are there some particular algorithms that set the benchmark for *knowing* the concept of performance

except, C doesn't speak O(killmenow)

besides sorting algorithms,BFS,DFS,

what are you interested in doing/learning, specifically?

ohh..sorry PoppaVic
pizza_, I just want to find out if I have the concept of performance and scalability

not me, I never speak that - but the /language/ doesn't know wtf YOU mean.

scalability is very platform-dependant. you haven't even mentioned your platform yet.

btw, what do you mean with this?

platform independent, mostly linux though

or wtf "scalable" means

architectural software, duh. blueprints and stuff.

ya what does scalable means ?
performance I can understand..scalability, growing in size ?

an integer isn't a fp-num, so… how many bits is my integer?

what is "the concept of performane and scalability"?

you begin to see why I get heated?

ya what is it exaclty, thats what I am asking

depends on the situation, usually means that a system can handle a large number of simultaneous transactions

hence why i usually ignore people who can't go look things up on wikipedia.

however that is defined

omg.. transactions again ;-)

i mean that very generically, couldn't think of a better word. "work"?

I think he means requests

"requests"? "input"?

rubbings.

well, that's one resource - but, not the best and just a titillator ;-)

zonkings

so'k, though - it's important - like folks grumble "code-points"

I had to recently build a daemon in C, that would load thousands of indexed lexicon so that it can perform a map operation of O(logn) in memory. Instead of loading the lexicon form the file…does it sound like scalable or improved performance
guess it was C/C++ mixed :P

lol

how much RAM did the lexicons take?

honest-to-bog… Where is cthulu (sp?) when his ass is needed?

maybe. if memory was at a premium, then maybe not. depends on the system requirements.

it was a language processing server..so the system was robust

Programs often have less mem than files; files are slower than programs. None of this matters if the basic concepts are already screwed.

SamB, more than the lexicon, for each operation , the loading and mapping time took longer, so instead of doing that the lexicion would load once and get mapped once in a daemon

then if high speed (work/time) was the goal, and loading them into memory at startup improved the speed of subsequent operations, then yes.

oh.. And (my usual): /everything/ hits limits.

oh, you were loading entire lexicons anyway?

Hi, i've got a question. I need to save a HUGE log array. I've tried textfiles but they end up in 100mb to Gb size. I think i have to go binary, but i dont know how to get it done right

yes SamB, I was translating english to another language..like google translation or babelfish

break them up

compression
zlib or something

1GB is too large?

so for a web page there are thousands of senetneces..can't load the whole lexicon for each sentence..that was a major problem

break-them-up

db ?

I guess the best thing is probably to mmap the whole things if they fit… otherwise, just the root area of each

i used STL's map for the lexicon part..

oh

I was just thinking. The textmode storage saves the numbers as a string, a number of lets say 10 digits would then take 10+2for "\n" bytes to store. In binary it would suffice with 6 bytes a number as 64 bits is 6 bytes

well, you definately don't want to be loading those from disk every sentence ;-)

is this log built up over time or generated all at once?

morten_, I suggest you use sqllite, it has a good C API too
bmorten_, I suggest you use sqllite, it has a good C API too/b

64 bits is 6 bytes on what architecture…

yes, SamB, thats why I loaded it in the daemon

and mmap incurs implicit blocking, definitely a no-no for scalability.

some arch where each byte is 10.5 bits

sorry, i mean 64bit is 8 bytes

some arch where each byte is 10.67 bits rather

C doesn't have "bytes" ;-)

RUNNING incurs implicit blocking, wise-ass!

humans do, and then they have contexts

Guys, Which library for using hash tables do you advise to me?

http://pastebin.com/m1c98371c

C has bytes

nope
it's got a baseline of 'char' and CHAR_BIT

a char is a byte in C
look it up

..and, as long as we want uint8_t, we are able to speak of octets and 'bytes'.

PoppaVic, you mean to say C has bits ? CHAR_BIT is 8 bit == 1 byte

quite right

dude, trust me - I used to fight the flow.. And there are many docs call it a "byte", but the purists get all bent.

PoppaVic, K&R sayas CHAR_BIT 8 (bits in a char) :P
s/sayas/says

It isn't worth another "char is a byte not a byte is a char" argument/discussion

so, now - I chuck out a "Beware: here there be dragons", and I go back to 8-bit bytes. Because most arguments against mean "smallest register" or some such

or does that depend upon system ?

Nope
8bit=1byte

even docs on 4-bit processors will do TWO fetch to get "a byte" or "a char" - I no longer care about the stupid arguments.

doesn't the no. of fetches from a processor depend upon the register size ?

primeArray"

what has it got to do with char is a byte or not
morten_, ask in C++

Well, ok

I think you can use fwrite though
fwrite(primeArray,1,primeArray.size(),fout)

no one care

I care PoppaVic

then write C

oops :p
that is messed up anyways
1 char = 8 bit :P

not everywhere

smurf(octets, nibbLes);

hmm..nibbles, aren't those smaller than bits

no

haha

void fun(); can be called with a variable number of arguments, right? fun(1, 2, 3) or fun(1, 2, 3, 4) works here, is there also a way to get the arguments inside the function?

half a bit

no
it'd just allow you to feed it crap

no, that declaration means the parameters are unspecified at that time

the (); proto just told the compiler "you don't care, don't test - stfu"

check out stdarg(3)

and I should make another prototype later?

see stdarg.h and the eillipsis (…) for variable numbers of parameters
ellipsis*
if that's your goal

k, thanks
was just wondering about that "feature"

"don't ask, don't test - just publish the place this lives"

's a declaration, not a prototype.

Auris-: foo bar(); // prototype

nope.

hm i can't really image a useful case for that

the definition can't be tested by the compiler, except the TU.

philipp_, useful case for what?

for such a declaration

it's another of our "yer allowed to shoot yer foot off" capabilities.

foo() is old C, used by those who are stuck in the past or don't know better.

k

Auris-: see above
It's not a /bad/ thing to allow, the prob is that folks think they can change shit at random (later), and it _might_ all work

philipp_, and you should read about type promotion when using such a declearation

yeppers
I'd be seriously paranoid passing such arguments.

k

morning!

philipp_, while it may be a cute exercise to have optional parameters in C, don't do it, it's a hack. varargs need a format and that makes them not very usable unless you actually need a format.

lo, Chris - it's late afternoon to me

3am here

Auris-: well, look at syscall(), too.

Hello gentlemen.

lo, Draichis
3pm here

Draichis, hi
0 here

i don't know the time i fucked up my clock

PoppaVic, well, glibc is not coded by beginners. open() has an optional parameter too (in linux at least).

I'll stick awhile, since this code is giving me a headache to rewrite.

Hi, where can I find a tutorials in system call programming using C?

Auris-: I understand, and the kicker is: he's a beginner (obviously), which means "use a fixed api or proper varidac"

alanhaggai, what is system call programming?
PoppaVic, which I think I told him to do. :-)

Who's maw
Hi Auris-

lo.

Where've you been recently.
I've almost nearly been worried.

Auris-: agreed. (although you degraded the latter)

Draichis, well, I've been here a few years now

Auris-, using the system calls provided by Linux kernel.

about 3-4 years I believe

Congratulations.

same place, every day.. Just busy with code issues, changes, plans and family-issues

You should certainly take a look at the switchtables availble on your os.
bYou should certainly take a look at the switchtables availble on your os./b
Or perhaps available on google.

how do i do a char[] as a part of a struct? I have tried struct datastruct { char text[48]; }, but if I do a datastruct test; test.text="something" I get an error that test.text can not be changed.

yer prob is yer ISP blows and you are never here long enough to notice diddly ;-)

I understand the latter.. apparently what we discussed was true.
It'll be sorted tomorrow.

nero, strcpy or char *

I promise.

PoppaVic, because it is ugly for anything else than scanf/printf. nobody who wants optinal parameters wants to do drawxy("axzd", 1, 2, 3, 4); — it beats the whole point of not supplying 0 for the "optional" parameter.

maw- gotcha.

yeah, some of the family-shit is peaking.. I worried myself sick for a day.. I can't care anymore - it's all events() to handle.

hi, im not sure this is the right place. I need to write a basic assembler that will read an assembley text file in a certain format and output in machine language. anyway according to the specs the "program counter" is 24 bits wide while the "instruction register" is 32 bits wide… that means each instruction is 24 bits length…

Comments

Hi all I have a fresh installation of apache 13 bundled with Debian I migrated a webpage from another server as

:

That may help you to test it right now, Then put the stuff like you want in the ~/.bashrc or the your zsh config
zsh rocks that much :-p

Is there a fully commented version of the Doxygen config file in the Debian Doxygen packages? (and if so, where is it?)

I tried setting the environment variables the program told me to, but it makes no difference

I can find all these simple examples, but ideally I'd like to start with a config file showing all options/defaults to work from

Use dpkg -L foo, where foo is the package name. And take a look

ok next problem I have is when I press 'down' in vi when I'm on edit mode, it writes a caps B

what xml-tool-related package is necessary for kernel building using make-kpkg?

I've looked around, haven't found anything yet - there are a myriad of smaller examples so far as I can see - but maybe I'm missing something or don't know which one to look at

after that, i may be sold on debian

vi sucks :-) use vim :-)

i mean vim

vim

Set nocompatible then

this is brand new etch install

Are you sure you're actually running vim?

TBH .. i'd ask the maintainer

You need a good vim config, see the vimtutor, on the end of it, it shows you how you can get that default config

vim

*nods* might drop them an email

ah, got it. xmlto

just ran vi and saw it say Vi Improved

or maybe ask the doxygen upstream ?

If you not have a .vimrc, it's set to compatible mode, that mean that it use the vi options
Nah, i told it a bit weirdy :-p

True

ahh. i dont have this problem on my other distros, its weird

I've been using the same .vimrc for a while - so I don't know what the defaults are

it's a symlink .. it's part of the alternatives system

what file manager do you use?

If you run it as "vim" do you get the same behaviour? Perhaps it's going compat based on the name..

]pablo[: don't take polls

i can try that…
rox is interesting!

ok, quick curiosity then

although I have this old one that looks like win98 explorer which is pretty sweet
forget the name…

sounds like x keyord of lang setting off\\

]pablo[: ask a better question

XFE?

XFE, yea

]pablo[: thunar, mc

nod, thinking of trying that shad0w1e

it looks crappy but pretty functional

kinda tired of nautilus :P

hah hah. also try rox. some people really eat that one up
and then again, konq is great, but if you dont have KDE install, its going to install a gizillion packages — not worth it
vulture, yes, when I run "vim" it works
how odd…

vim without a good custom config is very hard to use. I may suggest you to read the free vim book and go on #vim. Vim is a tool it take ages to use it rigth. Cause it have all that power. But believe it, you may learn the basic fast and learn things step by step when your confronted with it

"vi" probably runs vim.tiny ..

ahh

update-alternatives –list vi
you can of course change to point to vim

abrotman, and right u are!

!u

Damnit Jim! It's YOU. Y-O-U. Not *U*. U is a letter. YOU is a word. See ne1, or wud. Dutch for 'you' (formal singular). See http://ars.userfriendly.org/cartoons/?id=20041201

vim is much more nicer

is there any reason I'd want to be running vim.tiny?

i believe if you run `update-alternatives –config vi` it will give oyu the option
vim.tiny loads less plugins perhaps ?

or i can change the symlink..

I always install vim.full, i'm a bigg guy :-)

lol. yeas on my suse box, vi pointed to vim. here it points to vim.tiny
thank you!

nah, fat may much more appropriated :-p

hi all

if I want to install debian from the net, can I sue the buisness card install cd?
*sue = use

you can

yes selt

people, are there some utils for pdf2doc ? here I have pdftohtml only

selt, I did

!find pdf2 etch

(/usr/bin/pdf2dcm) in science/dcmtk ;; (/usr/bin/{pdf2dsc|pdf2ps}) in text/gs-common ;; (/usr/share/gs-esp/8.15/lib/pdf2dsc.ps) in text/gs-esp ;; (/usr/share/gs-gpl/8.54/lib/pdf2dsc.ps) in text/gs-gpl ;; (/var/spool/hylafax/bin/pdf2fax.gs) in comm/hylafax-server ;; (/usr/share/gs-afpl/8.53/lib/pdf2dsc.ps) in non-free/text/gs-afpl.

what the big difference between the buisness card one and the bigger netinstaller one?

just make sure you get dhcp on your Ethernet adapter and can get right online

i think you have to do a few iterations

unable to install initramfs-tools

you must have network connectivity for the buscard .. netinst has enough for a base install on the CD

how ?
ahh ok

ok, thanks

It's meant to be put on a little CD

you go from like pdf to html and then html to doc or something

I'm actually installing debian on a virtual machine

I try that, but don't have html do docor ps to doc :-(

huh ?
search around .. it's possible

ok

bonjour, I'm looking for some advice about how to enable my gateway to connect to a ftp+ssl server. I configured lftp the same way my client behind the NAT are and added some INPUT/OUTPUT ACCEPT iptables lines for 990/989 and 21/20. ftp works nicely but I can't make the ftps-data connection. Any hints ?

hey, thanks guys, i think im good!

doubtful!

hi , this offer unlimited space for free , share all together in acuatoria.com , now games, porn , mp3 and others uploaded

please don't spam us

ok
but that isnt sapm
spam

yes it is

it's offtopic, if nothing else.

isnt

see .. spam!
thanks staff

Does anybody know what i have to do make the "Last login:"-line (that you see when you login) dissappear ?

would have thought /etc/login.defs .. but i don't see it

http://www.debian-administration.org/articles/546

well .. i saw that .. but thought it was for the other stuff

h3sp4wn, ?

how do i get archive-manager to support rar?
and where do i get 7zip?

aptitude install rar (i guess)

man apt-cache

unrar-nonfree probablty

kk

apt-cache search 7zip

there's a catch. it's unrar now. in non-free

which he said

didn't i say that ?

no

the package itself isn't "unrar-nonfree" anymore. now there's "unrar" in non-free. and "unrar-free" in main

unrar-nonfree probablty

it's "unrar" in non-free repo

oh .. that's just stupid
it's still unrar-nonfree in etch i think

wow and i thought i was anal retentive

aaaaaaaaah ok

let the puck go!

mmmm guess not

damn warez kids

works now
thanks

http://packages.debian.org/stable/utils/unrar
a href="http://packages.debian.org/stable/utils/unrar"http://packages.debian.org/stable/utils/unrar/a
a href="a href="http://packages.debian.org/stable/utils/unrar"http://packages.debian.org/stable/utils/unrar/a"http://packages.debian.org/stable/utils/unrar"http://packages.debian.org/stable/utils/unrar/a/a

google? :P

oh .. unrar-free
how freaking stupid ..

it's to make warez kiddos feel less guilty about it

they're stealing

I can't change the owned of my external hard drive is that normal I wanna me able to read and write as normal user but when I try chown nuxis:nuxis /media/usb I just get chown: cannot access `/dev/usbdev': No such file or directory
the user now is root…

Comments

I have void free_mystructmystruct *m { if m = NULL { freem m = NULL } } and I call it for example with free_mystructb-m

She told me over the phone that she thinks C should be shot an buried.
why?
because it easily allows buffer overflows.

that's lame

allows a lot more than that, god damn

(for the average deskjob 9-5 programmer)

Ok. So why did you say she thinks it's a good language for beginners?

a good woman.

She didnt say "good" she said its not that bad. I asked her if it would be good for me to learn adn she said absoutley… Most unix/linux is written in mostley C thus making it essential to big programming and OS programming although it will soon be outdated.

ccameron, outdated?

are you familiar with her?

"should be shot and buried" .vs. "good for beginners". There seems to be a disconnect here.

Well, not outdated persay… but less common.

Xgc:

no, but she has the basic idea right about C.

You misinterpereted me.

I quoted you.

a lot of major operating systems are c

well, shooting and burying isn't necessarily bad for beginners, either

eh

Right but you played my words… She doesnt like the language, but she thinks it is necessary to learn and is a good language for a beginner to learn to open up doors to other languages.
ponder that.

Hello All!
My Name Is Italy

if I first tried to learn C, I'd be shooting anything I coudl get my hands on

because C is sufficiently high-level enough to write and read easily while still being low-level enough to poke around with the hardware

You started with?

ciao, italys

Hey There, GrooveBot.

parli italiano?

python

So why would you think she disagrees with me? I think you're putting words in her mouth that she probably wouldn't say.

what the matter if uint_value is not an uint

Good Joke. Those words went from her mouth to my ears.

because you're lying to printf about it

NickGarvey, I think I'd shoot myself if python were my first

why's that?

Well, C is somewhat my first, and I am diong fine… taking it slow.

Which words? Did she read my comments?

NickGarvey, I would assume all programming languages were like that, most likely

I dont want to start a flame. This discussion is over.

ahah, yeah, I went to learn java right after python and I was like "why do I freaking need to tell it that it is an int, isn't it obvious?"

NickGarvey, yeah, java hosting sucks

i just want to see the damn value in hex, i dont care if it is signed or unsigned

heh, I had an AP class in it (5 on the AB exam woo)

NickGarvey, unfortunately that's probably the language I'm most proficient in
NickGarvey, same here

oh? when'd you take the class?

NickGarvey, took the AB exam in May

printf cares

oh nice, me too

NickGarvey,

how do you think printf gets that value?

NickGarvey, let's not discuss it, though… the AP police might be listening ¬_¬

part 2's we can discuss

it uses it as unsigned value

no.
it decodes it as an unsigned value.
you don't pass it an unsigned int, it doesn't get an unsigned int to decode

NickGarvey, I knew QuickBasic, C, Pascal, and Lua before heading into the AP class, though… so it was a bit easy (and I suppose you knew at least python)

welcome to the wonderful world of va_args

heh yeah

it uses an unsigned int* decode?

my school only covered the A material though, had to learn the AB stuff mysefl
self*

NickGarvey, the one thing I give python, though, is its incredible ease of learning for programmers of other languages (I blasted through the tut in 2 days and was sufficient enough to write a CMS in it)

Don't be rude. I suspect you're taking a couple of (her) comments out of context. It's obvious C isn't a good place to start, for the reasons I state (and others). That doesn't mean it can't be done.

NickGarvey, same here… but I managed to get a whole period off for "independent study" for the AB exam

so, what happens when you decode an unsigned int * as an unsigned int?

hehe nice

I would have thought C is a better place to start then something higher, as it forces you to learn

eh, but it is such a pain in the ass that a lot of would be learners would give up pretty quick heh

something highly structured is the best place to start, because fundamental programming concepts are most obvious in them
s/them/it/

Could you point me in the right direction… here is the code that i have so far…http://rafb.net/p/aygp6h52.html how would one go about appending the th and nd's and rd's onto the dates in my program… ex: 13 August 1992 to become 13th august 1992…?
or anyone for that matter

It's better to focus on algorithms and system structure, than to worry about avoiding undefined behavior and managing resources that distract from the primary goal.

I'd probably.. just write a function.. to take a number and return the two characters..

lookup table?

split up the printf() so that you run a switch on the day, then a switch on the month

Pryon, what other kind of table is there?

hell it is really just a switch statement

i guess something like *(unsigned int*)&param[n]

Later, you can try your hand at building systems *and* also juggling the obstacles C throws at you.

ccameron, ask the user to enter the suffix B)

we were taught programming in terms of go get the computer to do some set of sums.

OMFG
i just pulled up my shell and in the last line it said "sly used here"
HOLY
wtf

u hav ben haxored lolol

lulz

the matrix has you

I do not see a question.

xgc is prolly sly b/c hes mad at me
?
wtf this is sketching me out
just chillin there "sly used here"
wf.
wtf

you're probably being haunted by the ghosts of dead mexicans

OMFG THIS IS WEIRD HES PROLLY SENDING PASSWORDZ And shit

don't shout

my comp is bein all slow wtf

You might want to ask your friend about "intrusion detection systems"

haha
or maybe nmap

xgc… touche…
is this u??? in my comp?

check yer logs, bro

how?
i already did who

….

A smart hacker would clear the logs of all traces…

and it came up with cameron:0
and cameron pts/0

Sounds like you found the problem.

which OS?

Ubuntu 7.04

xD

i guess something like switch(fmt_opt) { case 'x': printf_hexlow(*(unsigned int*)&param[n],length, fill, fillchar); break; }

ahahahahaha ubuntu

How chek log?
Problem?

ccameron, check netstat maybe

more like va_arg

OMFG
AMAZON.COM

heh
sly used here was probably a word wrap of "previously used here"

heheheh I love this new dell commercial
with devo

http://rafb.net/p/a4eUJP45.html
you should be able to work out how to add the st/nd/rd/th from that
And after I went to all the effort of preparing a hint for him…

He just wanted to someone to write it for him.

meh. It's not like it would be difficult from what I preped.

is chatzilla any good

fez, it's ok

leap year babies?
it sucks

fez, what are you using now?

myprefix_open if #define PREFIX myprefix_

sorry, _PREFIX(open)

_PREFIX would be a reserved identifier, and should not be used.

#define PREFIX(func) myprefix_##func

I want to be able to specify another prefix in a -D to gcc&

why wouldn't you be able to

-Dmyprefix_=otherprefix_ does not work

ah

Dr Z are you around?

re

hmm

what means int __init foo_init(void) ?
i know that it is function
what means __init /

yes

http://img337.imageshack.us/my.php?image=timestickhn6.jpg
Extremely easy, fun, and educational projects for children and adults of all ages.
not to mention "what do I do with dozens of little tiny clocks" ?

do you throw them really fast to test relativity?

which places in the world do I want to watch Time Fly ?
and being twisty-tied to the ruler, the batters are easily changed
^batteries

there's no __init in C, it's probably an implementation defined macro.
the linux kernel source defines __init in linux/init.h, for example

http://tuxgraphics.org/electronics/200612/article06121.shtml

ah ok than
x

sorry, redl. I need a better font/pair of eyes

hi all

2am on sunday.
someone here?

yes

4pm on monday here

I would like to know something about splines and MMSE

3pm monday here

I need a fast way to compress a curve
i.e. to get 1, 2, or more straight lines with least error based on a curve
that is a reverse spline

the term is to 'segment a curve'

yes
in fact, the word spline doesnt apply
because the points are deviated along some arbitrary curve

search google for "curve segmentation"

how on earth — these papers are new. Im surprised this wasnt solved 100 yrs ago

much like I'm surprised that you haven't yet learned how to spell "I'm" or "wasn't".

are you familiar with the code for this?

sure — look in ispell.

badabing!

Adam Roan

morning

afternoon

that's good

sometimes

not at breakfast though

perhaps

it really all depends

I find afternoon at breakfast quite good.

aku juga

which language is that supposed to be?

bahasa melayu

ah, I haven't used that for a long tme.
time, even.

indeed

mattuna matana?

Hello all. What is a good way to encode error return codes? Use enum perhaps ?

yes, perhaps.

enum, define, error class,…

alright, thanks people

i usually use simple defines eventhough that might not be the best hosting way

enums have the advantage of appropriate syntax support

alguien que vea Naruto que me hable en PV

English?

que??

;-)

no warez here. and if it isn't warez it still doesn't look like it would be about c.

bye to all users

Naruto must have been some japanese anime

hi all…I want to build a static library by ar, but I wnat the library I built including the library (such as glib) I called in my program, How to make it?

compile glib as a static library
unless you just want to link with it
you can statically link some and dynamically link others.

hello Zhivago

hi, I'm having a problem with iconv(). I'm pretty sure I'm misusing it, but not sure how: http://rafb.net/p/37m9gI37.html What am I doing wrong to make the cstr advance by 3 instead of 2?

good idea!!!

casting &ucs_consume is at least one problem.

what should I do instead? It's indeed a .long in assembly where it's stored. I thought I read in the POSIX spec that the input could be a series of ints or longs.

sigh.
staring at assembly.
ucs_consume is not a char *, as such having iconv treat it as such is likely to be bad juju.

so what should I be doing to convert these UCS-2 values: .long 0×68,0×65,0×6c,0×6c,0×6f,0×20,0×77,0×6f,0×72,0×6c,0×64,0×200,0×200,0×200,0×200,0×205 ?

how about you stop writing assembly here, m'kay?
which part of an mengine_cell is the pointer to the ucs-2 values?

mengine_cell is a union. I have in the paste: mengine_cell *ucs = ptr.pointer;
umengine_cell is a union. I have in the paste: mengine_cell *ucs = ptr.pointer;/u
..long 0×68, 0×65, …
so it's essentially a pointer to an array of longs

humm
i need to share a message queue between 2 threads
!_!

and?

and im failing

Windows I presume?

no

Linux!?

unix at any rate

kati^: paste your code in rafb.net/paste and be more elaborate on 'i am failing'

kati^: pthreads?
or fork?

meh code won't tell you much, just pthread A is mostly a message handler

or any other wierd thing?

kati^, don't make it guesswork.

and pthread B is a ui thread
and the data should be shared betweenthe threads but it doesn't seem to be

So create a 3rd thread

:

data such as???

create a testcase, show code,…

rafb.net/paste already

fine fine
one sec

kati^: how is your use of the mq_* functions failing?
(not that you need posix mq to have a queue inside a process, but it may simplify things)
b(not that you need posix mq to have a queue inside a process, but it may simplify things)/b
so why didn't you use the member name that is the char*?

I'll try adding that twkm to mengine_cell

http://rafb.net/p/jvHTKt25.html
no, its my own message queue
its just really simple.

kati^: you aren't using any synchronization.

true
but

well, wth did you learn to write threads that way?

since the message being read is unsigned long
isn't that atomic?

sigh.

I almost never use threads
mostly I keep things to single serial processing
but gtk doesn't really afford that option

is there a particular reason you need threads? I take it your UI is blocking, so perhaps you could make it more event-oriented and use Gtk+'s event and timer callbacks.

the main reason is that the backend is designed to work with various frontends, gtk+ being one of them
the backend and frontend need to work simultaneously

time to learn that synchronization isn't just for rappers.

kati^: the easiest way of doing this might be to create a pipe() or 2, fork() and have one path be for your UI, and another be for your backend. If the pipe isn't scalable enough you could use shared host memory for sending large amounts of data.

Hi.How do I inflate something encoded with "Content-Encoding: deflate" in C?

zlib

whats the best way to parse a buffer after \r\n\r\n separators?
strtok/strsep seem usefull only for 1 char separators

Thanks.

and the solution I'd adopt now is strstr

well it seems like mutexes would be good.. locks when pushing msgs and locks when popping them
right?

kati^: http://users.actcom.co.il/~choo/lupg/tutorials/multi-thread/multi-thread.html

locks are a terrible idea, generally speaking.
but, they're better than nothing.
you could write a simple state machine so that you can operate on streams.

a state machine within a state machine… hmm
but I don't see how a state machine could do the job here, can you explain a bit

what is a state machine?

I'm blocked

why would kati^ want to fork? pipe works just fine within a process as between two.

pending on different events the machine transcends through different states

do you speak english?

yes

please try again.

kati^ noted that the UI couldn't block, so that's why I suggested using the pipe() as a communication channel, and using fork() instead of pthreads.

you didn't understand my sentence or it was the wrong definition?
what I was trying to say is that a state machine is represented by states and the transitions between them that are caused by different incoming events

ok
so if you were to make a machine which had a state representing that "\r\n\r\n" had just been processed, then it could answer your question?

that would mean reading parsing through strstr

well, 'transcends' can't be used like that.
no, it wouldn't.

I found my problem. It seems that iconv() doesn't like data stored in a .long. It expected a short, as my compiler (written in Tcl) is only capable of support the BMP and UCS-2 I should do something else…

I think a state machine is a bit much for that

well, as you wish.

no. please, tell me
I want to know your solution

I already told you.

yes. theoreticly its clear. but the implementation would still require me to go through the string

stream

well, I don't have a stream, but a known sized buffer recv()'d via TCP/IP
I think I'm not thinking this clearly
missing the trick or something

well, what's a stream?

a continous flow of data

feh, locks don't seem to be helping anyway

kati^: do you understand where you need them?

well
i assumed, in my push/pop functions since thats really the only place shared between the threads so far
before doing anything and after doing anything in those routines

kati^: what kind of lock did you create and initialize?

Hi again people. For once more I need your expertise. Is this an approved methodology to handle malloc failures and to avoid leaks in between multiple malloc() calls? http://rafb.net/p/ioekNd35.html

initialized with PTHREAD_MUTEX_INITIALIZER, which I have just read means that items can be locked once and if attempted to be locked again can cause deadlock

"approved"!?!?

approved by the experts, I presume

pthread_mutex_t

if you're going to use goto, you might as well just jump to different labels depending on how far along the way you are.

yes yango, by the way I missed an mdepth++; anywway

this avoids your switch statement with magic case numbers and getting confused with mdepth

hi everyone

hi

mmm thanks Chris, that sounds a nice althernative

also, why do you prefer sizeof(struct matrix) to sizeof **mat?

nice trick! I haven't though that either
thought*

and why is goto uppsercase on line 18?

is the following code below a prototype of function pointer ?
void *(*request_input)(const char *title, const char *prime);

yes

great .thaks..

did you even try to compile this before you pasted?

opera-118: sry I need to ask stupid questions.. because I only know c++ ..hehe..

technically there's no such thing as a stupid question, only stupid people.

well, function pointers can be useful in c too

hahah..

c++*

no I didn't try to compile it. That was a typo error. Thanks again

opera-118: I'll read some more
opera-118: are you an engineer?

m.sc. yes

opera-118: you probably know alot about c… nice meeting you.

heh np

you also can't have a case label at the end of the switch block without anything

really? my compile host didn't output an error or a warning. I'll check the specs

which compiler?

what's the syscall used by gdb to control a program's operation during debugging in Linux?

ptrace

stathis, what warning level?

thank you twkm

Ah, my mistake _again_ thanks Chris / zacs7
Is that a warning ?

it's an error

Don't bother guys. I handled it all. Thanks for the hints
(it was an error indeed)

good night everyone

night

does this look good? http://rafb.net/p/usNOAl18.html
have I grasped your idea?

no
and you appear to have no idea of what a state machine is

yay

have you studied any CS?

yes I did, but I can't make that into a state machine
I already have a state machine that switches through different states on the account of the packages I recv

hmm, I suspect you should ask for your money back.

okay. well thanks anyway

7: error: asm/io.h: No such file or directoy, I'm trying to port a 32bit app to
I've checked and my system only has /usr/lib/klibc/include/asm-x86_64/io.h

but problem persists

hi
have a simple problem with some preprocessor statement
#define UINT64_C(v) (v##i64) gives me with gcc the errormessage
invalid suffix "i64" on integer constant
in the complete C reference it says that ## concatenates two toeks
tokens
in this case its probably more a variable definition. anyone an idea how to rewrite it
so that gcc will compile?

Is it a good tactic to set pointers to NULL after a free them ?

grabher, how do you use the macro?
stathis, if there any chance it could be reused it would make sense.

thank you wobster

it is used for a table initialisation
static const uint64_t CR[10][8] = {
{ UINT64_C(0×0080000080008080), UINT64_C(0×0080800080000080),
etc
and this piece of code was written with Visual C therefore the compile problems with gcc I assume

That would create strings like 0×0080000080008080i64.
Which is invalid.

ah so probably use L instead of i64

well, "names"
LL even

or LL?
okay great thanks wobster!

you'll need a c99 conforming implementation for that to work.
asm/io.h smells of linux.

is the directory even in your include dirs?
"/usr/lib/klibc/include/asm-x86_64/io.h"

good morning folks

twkm how do you mean c99 conform?
this is not c99 when I use the preprocessor ## with LL?

UINT64_C is a macro defined by c99.
using LL will not provide the same type, even ULL may not.

so when it is already definined I can throw away the macro?:

if you have an extended not-c99 compiler with a long long extension then ULL is perhaps good enough for your needs, but even then won't necessarily be exactly 64 bits.

ah yeah I always get the warning when trying to use 64 bit variables
but I need them for our tests

what warning?
you may not have the value you think if the initializer is truncated to a narrower type (than unsigned long long).

yeah the warning with 64 bit variables and it could be truncated

if you are using gcc (which does not conform to c99, but does have many extensions in that direction) you can try -std=c99.

thanks I am doing this!

spin your prayer wheel, and don't forget -pedantic -Wall -W

there is another problem, as I said the program I am trying to compile was written with Visual C
intrin.h: No such file or directory
is this a Visual C specific file, or do I need to set a path to it with gcc?

better figure out what it provided, and write the replacements for what you need.
compiling on windows — use an appropriate -I. porting to something else — you'll need to port the functionality provided by the header and any object code referenced by it.

sup RobbieAB|away

clearly not my internet connection.

cheers twkm have to see how to do it!

Hi
What book should I buy to learn Algorithms and Data Structures?
"Introduction to Algorithms"?

get hold of old Sedgewick book
"R. Sedgewick: Algorithms"
but the _old_ one
which has like 40-ish chapters in a single volume
it covers all needed basics, imho
new sedgewick books are crap
then there's Cormen's Algorithms

The "Operating Systems Concepts" book was fair.

and then there's Weiss book on algorithms
which author/

Sedgewick.

The old Sedgewick I can't find…

oh, he wrote even about OS-eS?

So I think I will get the Cormen's

It was also called the Dinosaur book.
Ugh, sorry.
It was silberschatz.
Hmm, I thought it was Sedgewick..

nah, that's silberschatz
:p
well, i haven't browsed it
(although a colleague has it on the bookshelf)
but it can
can't be worse than tannenbaum's :p

Fast Algorithms for Sorting and Searching Strings, was a Sedgewick book, if I'm not mistaken.

i don't know
but his old book is really good
compactly written, covers a broad range of topics, has code

Yeah, Jon Bentley and Robert Sedgewick.

Fast Algorithms for Sorting and Searching Strings, was a Sedgewick book, if I'm not mistaken.

ciao

Yeah, Jon Bentley and Robert Sedgewick."

I like http://www.nist.gov/dads/ and it's often useful to search for "algorithm animation" to see visually how some more exotic algorithms work.

I was able to compile a program with gcc without setting any flags

lol, thanks
thanks guys
I will buy the Cormen's

now when I use CFLAGS = -ansi -pedantic -Wall -O to see if it is ansi C compatible
I get one error which says
invalid application of `sizeof' to incomplete type `hash.h'
this is the line that is causing the problem
*resultContext = (HashContext *) malloc(sizeof(HashContext));

you did not deklare HashContext

also, don't cast malloc.

hello all
i am try to understand how secure linux web hosting tool work like
like ls cp ..echo.. etc

the source is freely available

is there something make reverse engineering more easy ?

reverse engineering would be harder

i have other problem i know can read C
but i don't know how Make and configure and this things work

google is your friend

Anyone know where I can find a dwarf2/3 example of try catch in C.

readme HashContext is declared in hash.h
as I said it worked without any CFLAGS but when using the ansi CFLAG I get this error

light, so u think this is good step

so something seems to be not in proper ANSI coding…

light, or there is something more sample i can read to have more understand of c and how it's work

your compiler doesn't seem to think so.

there's plenty of code on the web

Hmm, my net is crap

not with the ANSI C flag thats right

-ansi also disable some useful platform bits I think too from GNUC

without it finds the declaration

unlikely.
more likely it just ignores the issue and leaves you to pray you didn't make a mistake.

is it declared before or after the sizeof?

read the manual for gcc's -ansi flag

you might imagine that i hadn't done some.
so.
but you would be wrong.
glibc on the other hand might disable parts of itself, requiring you to actually read the documentation if you want to turn them on.

HashContext is declared as a typedef in the file hash.h
and included so it should be know to the compiler

what is HashContext?
is it a struct? a union? a function pointer?

morning

hash.h has the imcomplete type, not the entire thing.

a structure

well, by implication the entire thing does, but it is because of the h member.

struct HashContext {
/* the ID of the algorithm */
Algorithm algorithm;
/* the hash value length in bytes */
uint32_t hashValueLength;

dork.

that certainly is not a complete type :p
!topic grabher

I am sorry
my first time here I will ckeck it!

usually you do that before asking anything.

http://rafb.net/p/9HSrEj10.html
Here I have the struct to HashContent
and the compiler tells me it is incomplete when trying to do
*resultContext = (HashContext *) malloc(sizeof(HashContext));

Do you have Algorithm ReturnValue and Hash defined?

yes one in the same file above and the other one in an included header…

info libc "Feature test macros" you will need to enable some POSIX related macros because of the types you're using I think.

don't cast malloc.

twkm thanks I will consider that

and where is Hash defined?

imaginator thanks for your advice I will look into that, hope it helps

here's the part of the POSIX standard that lists the uint* types btw: http://www.opengroup.org/onlinepubs/009695399/basedefs/stdint.h.html

Don't consider not casting malloc, actually *don't* cast it

how about ReturnValue?

Hash is defined in the same file Hash.h, just above HashContext!

And ALgorithm.
there is nothing about HashContext in your paste.
so i tire of trying to help you figure this out as you dribble information. /me wanders off

!dontcastmalloc grabher

There is no need to cast the result of library functions that return void *; it makes your code hard to read, adds no value, and can hide a bug if you don't have a valid prototype in scope.; see http://www.cpax.org.uk/prg/writings/casting.php

^ worth reading this

except when you want C++ compatibility which is pretty ugly anyway

no
if you want "C++ compatability" write a header with whatever C functions you want, and extern "C" your includes of it
don't compile C++ code as C .. that would be idioitc

http://rafb.net/p/9HSrEj10.html here is the HashContext structure declared
twkm Algorithm is defined in the same include as ReturnValue

;

guys, what are the socket man pages called ? they aren't installed …
only man socket exists

man recv ?

but not send recv ..

hello; I have : void free_mystruct(mystruct *m) { if (m != NULL) { free(m); m = NULL; } } and I call it for example with free_mystruct(b-m); but before and after the function call, b-m is not NULL. Any idea why it is not null after the call ?

dunno ask #linux
free(NULL); is an OK thing to do

thanks man

sorry, I don't understand you

you're modifying a local variable value of the pointer
you'd have to pass the parent struct or a mystruct **m

I see, thanks
o right I should pass the pointer to the pointer
of the structure

Comments

« Previous entries · Next entries »