Re: [gentoo-user] Re: Masked package when upgrading world

2003-11-11 Thread Doug Weimer
On Tue, 2003-11-11 at 07:37, Eamon Caddigan wrote:

 Well, as a temporary hack, I went ahead and 'injected' the latest
 unmasked versions of Tcl/Tk (dev-lang/tcl-8.3.4 and
 dev-lang/tk-8.3.4-r1). Now, `emerge -pvuD world' doesn't try to
 downgrade Tcl/Tk, but `emerge -pvUD world' is still returning errors.
 This will serve me well until I install another masked package (which
 isn't likely to happen any time soon).

Have you considered copying the ebuilds for the ~x86 packages you want
to a portage overlay? Then change the KEYWORDS in the overlay ebuilds to
x86, so that they are no longer considered masked.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Volume full during world update - how to clean up now?

2003-11-06 Thread Doug Weimer
On Thu, 2003-11-06 at 06:47, Karl-Heinz Zimmer wrote:
 Hi,

 In /var/tmp/portage/ there are a lot of directories full of files (all
 together more than 2GB) - I asume this is because the previous emerge
 run was stopped due to lack of disk space, sp probably it did not clean
 up like usually...  it looks like these files are the sources where my
 packages get build from.
 
 Am I allowed to erase the contente of /var/tmp/portage after the emerge
 has finished running - or would that be less wise?  :-)

Removing everything is fine. Emerge will remove the tmp directory and
remake it each time you emerge a given package. So removing them
yourself wont make any difference.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] how to build only new packages

2003-11-04 Thread Doug Weimer
On Tue, 2003-11-04 at 05:38, Erwin Lang wrote:
 with this command I can install a prebuild package on a pc. but I want to know 
 how I can avoid to build the same package every week if there is no new 
 version available (as I said, emerge --buildpkg builds a package regardless 
 if there already exist a packaged version of this package in /usr/portage/
 packages or not.). Assume a script execute once a week: `emerge --buildpkg 
 apache tomcat samba ... (and other packages I need)`. So I don't have to 
 build packages manually if there is a new version available. And I also 
 need'n look for new versions of certain packages in portage.

You can use the --update option along with buildpkg. Assuming you don't
unmerge the packages after you build them, using 'emerge --update
--buildpkg packagename' will only build a new package if a new version
is available..

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] emerge grabbing wrong packages

2003-11-04 Thread Doug Weimer
 On Tuesday 04 November 2003 04:27, you wrote:
  Okay, a few days ago I emerged the 2.6 kernel to take a look at. I did
  NOT install it and have since un-emerged it.

  emerge --pretend nvidia-kernel
 
  These are the packages that I would merge, in order:
 
  Calculating dependencies ...done!
  [ebuild N ] sys-kernel/development-sources-2.6.0_beta8
  [ebuild N ] media-video/nvidia-kernel-1.0.4496-r3

The development-sources might still be in your virtuals file. Try: `grep
virtual/linux-sources /var/cache/edb/virtuals` to see if
sys-kernel/development-sources is still there. If so, just remove the
development-sources entry.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Making .ebuild

2003-11-03 Thread Doug Weimer
On Mon, 2003-11-03 at 02:53, raptor wrote:
 tring to make an .ebuild for tcng but in the make install stage I'm getting errors 
 i.e. tcng install script tries to write 
 outside the sandbox..  
 This is the problematic part of the Makefile :
 -
 install-tcc:tcc check-install-dir scripts/localize.sh
 ln -sf scripts/localize.sh .
 tar cfh - $(TCC_BINDIST) | \
   (cd $(INSTALL_DIR)  tar xf -)
 cd $(INSTALL_DIR)  ./localize.sh
 rm -f localize.sh $(INSTALL_DIR)/localize.sh
 -
 

The problem here is that INSTALL_DIR is set to /usr/bin. In order to
build inside the sandbox you need to install everything under the ${D}
image tree. From the list of files that are being installed it looks
like you probably want INSTALL_DIR to be ${D}/usr . You can try passing
INSTALL_DIR=${D}/usr to make in the ebuild.

Good Luck,

Doug


--
[EMAIL PROTECTED] mailing list



RE: [gentoo-user] Operating System not Found

2003-11-03 Thread Doug Weimer
On Mon, 2003-11-03 at 02:00, Hall Stevenson wrote:

 Gentoo suggests that you normally NOT have /boot mounted during normal
 use. If it's not, when you copy your new kernel image to /boot, it will
 fail. In my case, I don't use 'genkernel', but compile mine the
 old-fashioned way. Lastly, I run 'make install'. This does various
 things, one of which is to copy the appropriate files to /boot, make
 symlinks for System.map, and so on. I noticed one time when /boot was
 NOT mounted, that the 'make install' command did NOT complain or fail.

This is handled by genkernel as long as /boot is in your fstab. In the
compile function it runs: mount /boot  /dev/null 21 .

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] /etc/kernels/default-config

2003-10-30 Thread Doug Weimer
On Thu, 2003-10-30 at 02:16, Guy Van Sanden wrote:
 Hello everyone
 
 What does this file do?  is it copied to a new source-tree when
 downloaded?
 
 I'm asking because Im seeing differences in a 2.4.20-r7 kernel I
 compiled during install, and one I compiled later from the same sources
 without modifying anything.

The default-config file is used by the genkernel program. If you are
building the kernels by hand, this file wont be used without explicitly
copying it to your kernel source tree.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] xfree dependancy - why?

2003-10-26 Thread Doug Weimer
On Sun, 2003-10-26 at 16:27, Dhruba Bandopadhyay wrote:

 I get this too on a server machine.
 
 $ emerge -Duvp world
 
 These are the packages that I would merge, in order:
 
 Calculating world dependencies ...done!
 [ebuild  N] x11-base/xfree-4.3.0-r3  -3dfx -sse -mmx -3dnow +xml2
 +truetype -nls -cjk -doc -ipv6 -debug -static +pam -sdk -gatos
 [ebuild  N] dev-perl/perl-tk-800.024-r2
 [ebuild  N] dev-perl/Tk-TableMatrix-1.01

perl-tk depends on virtual/x11. I'm not sure why Tk-TableMatrix is being
pulled in, but that's why your getting the xfree dependency. Maybe add
'-tcltk' to your USE flags and see if that helps.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Rogue dhcp server

2003-10-26 Thread Doug Weimer
On Sun, 2003-10-26 at 16:47, Chris I wrote:

 Short from trying harder to find him and kicking him in the shins, can  
 anybody think of a way to block dhcp servers, or to specify which is  
 allowed to be used. I've tried ip and mac filtering in iptables, but it  
 doesnt seem to be effective.

The dhcpcd client doesn't support this if I recall. If you're willing to
use a different client, dhclient from the net-misc/dhcp package supports
a reject keyword. The dhclient.conf man page has specifics, but in short
'reject ip-address;' should do what you want.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Emerge Telnet

2003-10-25 Thread Doug Weimer
On Fri, 2003-10-24 at 23:07, Alan Watson wrote:

 Can anyone tell me the emerge commnd to install telnet? Both server
 and client.

emerge net-misc/netkit-telnetd

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] change grub boot message

2003-10-24 Thread Doug Weimer
On Fri, 2003-10-24 at 16:40, Redeeman wrote:
 when grub starts, it shows a message somewhat similar to this: to edit
 press e, for blabla ...blabla
 how can i remove or change that message? thanks!

From what I can tell this is hard-coded into the grub source. If you
want to edit the message and rebuild grub, take a look at
stage2/stage2.c in the run_menu function.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Update question

2003-10-21 Thread Doug Weimer
On Tue, 2003-10-21 at 15:39, Rick [Kitty5] wrote:

 IMO etc-update should make a backup copy of all files replaced
 
 Rick

You can do this with two small changes. First, add the --backup option
to mv_opts in /etc/etc-update.conf. Note: The default backup option will
create a single backup named filename~. If you do want to maintain
previous backups, you can use something like --backup=numbered. Take a
look at the man info page for other backup options.

The second change is to the etc-update program itself. Unfortunately,
etc-update's current method using when using '-5' is to reset mv_opts to
. To add backups when using the '-5' option open /usr/sbin/etc-update
in an editor and locate the lines:

  if (( ${input} == -5 )); then
 input=-3
 export mv_opts=
  fi

Change the export line to: export mv_opts=--backup. Or
--backup=numbered etc.

Good Luck,

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Masking errors_Unmasking a pkg

2003-10-21 Thread Doug Weimer
On Tue, 2003-10-21 at 19:05, Rumen Yotov wrote:
 Hi all,
 
 Just used a 'dep-clean' command and here is the output:
 Cut-
 #dep-clean
 sort: open failed: uniq: No such file or directory

To fix this take a look at: http://bugs.gentoo.org/show_bug.cgi?id=30248

 Calculating dependencies   
 !!! all ebuilds that could satisfy =net-libs/libnet-1.0.2a-r3 have been masked.
 
 !!! Error calculating dependencies. Please correct.
 End-
 
 It seems that I need to unmask it or other. Also how to unmask something.

libnet-1.0.2.a-r3 is in ~x86. I would guess that you installed
something from ~x86, but aren't using ACCEPT_KEYWORDS=~x86 in
make.conf. For dep-clean to run properly, you'll have to set
ACCEPT_KEYWORDS in make.conf to ~x86 before you run it. If you don't
want to change ACCEPT_KEYWORDS each time you run dep-clean, you could
use the -I option to interactively remove the packages that are in
~x86.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] changing module search order for Perl and mod_perl

2003-10-21 Thread Doug Weimer
On Tue, 2003-10-21 at 10:01, Andrew Gaffney wrote:
 I believe the recent problems getting mod_perl to work correctly I've been having 
 are due 
 to the module search order. CGI.pm that comes with Perl 5.8.1 is in 
 /usr/lib/perl5/5.8.1 
 while the CGI.pm installed by CGI-3.0 that is required by mod_perl-1.99.09 is 
 installed in 
 /usr/lib/perl5/vendor_perl/5.8.1. I believe Perl is finding the original CGI.pm file 
 first 
 which is causing the Apache.pm not found errors that a few people have been 
 reporting. 
 Now, to my actual question: how do you change the module directory search order for 
 Perl 
 and mod_perl?

You can modify the @INC used by mod_perl through
/etc/apache2/conf/modules.d/apache2-mod_perl-startup.pl . Add a 'use
lib' line similar to the one already present in the file. I'm not sure
what version of perl you are using, but the CGI.pm module included in
dev-lang/perl-5.8.1-r1 appears to be identical to the one included in
dev-perl/CGI-3.00.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Re: animated gifs

2003-10-19 Thread Doug Weimer
On Sun, 2003-10-19 at 12:11, Andrew Gaffney wrote:

 So I don't have to go through the trouble of going over to my Windows computer and 
 then 
 transferring the files back, is there a small (smaller than the gimp) program that 
 can do 
 simple stuff like MS Paint can?

There is xpaint, which has an ebuild.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Fw: emerge command returns errors

2003-10-19 Thread Doug Weimer
On Sun, 2003-10-19 at 19:14, Alan Watson wrote:

 I've just finished installing from a stage 1 tar ball. I chose to run
 /dev/hda3 as ReiserFS. The problem is I made a mistake and omitted to
 install ReiserFS tools. After rebooting I ran the command:
 
 emerge -k sys-apps/reiserfsprogs
 

The reiserfsprogs are in sys-fs rather than sys-apps. Try:
emerge -k sys-fs/reiserfsprogs

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] python configured for tk?

2003-10-18 Thread Doug Weimer
On Sat, 2003-10-18 at 10:14, Claus Ladekjær Wilson wrote:
 On starting fetchmailconf I had this error message:
 
 bash-2.05b$ fetchmailconf
 Traceback (most recent call last):
   File /usr/bin/fetchmailconf, line 9, in ?
 from Tkinter import *
   File /usr/lib/python2.2/lib-tk/Tkinter.py, line 35, in ?
 import _tkinter # If this fails your Python may not be configured for Tk
 ImportError: No module named _tkinter

This is usually due to python being emerged without the tcltk USE flag.
Run 'emerge -pv python' and look for the tcltk flag. If it is '-tcltk,'
you need to add tcltk to your USE flags and emerge python again.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] changing reiserfs partitions to ext3

2003-10-18 Thread Doug Weimer
On Thu, 2003-10-16 at 17:25, Wes Gray wrote:

 If you end up doing this please post how it went. The question seems to
 be, just how complete is tar?  And what options does it need to catch
 everything.

I recently did a quick backup/restore with tar and ssh to convert two
partitions (/home,/usr) on this machine to a new filesystem. My basic
procedure was:

From host1:
tar zcv /home | ssh [EMAIL PROTECTED] cat  /backupdir/home.tar.gz
(covert /home to new fs and remount

From backuphost:
cat /backupdir/home.tar.gz | ssh [EMAIL PROTECTED] tar -C / -zx

I didn't have any issues with permissions/symlinks although it's
possible I just got lucky. If you want me to check for specific problems
that I might have missed, just ask.

Doug


--
[EMAIL PROTECTED] mailing list



RE: [gentoo-user] login problem

2003-10-13 Thread Doug Weimer
On Mon, 2003-10-13 at 10:14, Davide Fanciola wrote:
 The problem is that I cannot login in text-mode console!
 But the X-based login (entrance) is working fine.
 
 When I say anything I type, it means : any user account present on my
 system...

Are you prompted for a password? If not, you probably have a bad
/etc/pam.d/login. Information about this bug can be found here:
http://bugs.gentoo.org/show_bug.cgi?id=29563 . Basically, just move
/etc/pam.d/login out of the way and reemerge sys-apps/shadow.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Find ebuild from file?

2003-10-12 Thread Doug Weimer
On Sat, 2003-10-11 at 23:55, Bryan Whitehead wrote:
 How do I find what ebuild is responsible for a file on my system?
 

You can use the qpkg program from the app-portage/gentoolkit package.

qpkg -f /path/to/file

 simple example... I want to find the owner of /usr/X11R6/bin/xcmsdb
 
 With an rpm based distro I'd just run: rpm -qf /usr/X11R6/bin/xcmsdb
 [EMAIL PROTECTED] ~]$ rpm -qf /usr/X11R6/bin/xcmsdb
 XFree86-4.3-5mdk
 [EMAIL PROTECTED] ~]$

You can get similar functionality from the app-portage/epm package:

epm -qf /path/to/file

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] compiling new kernels

2003-10-12 Thread Doug Weimer
On Sat, 2003-10-11 at 22:51, a park wrote:
 when i emerge a new kernel to compile, does the /etc/kernel/default-config 
 get replaced?
 

No, emerging kernel sources basically does the following:

1. unpacks the source tarball
2. applies any patches
3. runs 'make mrproper' 
4. runs 'make include/linux/version.h'
5. moves the resulting tree to /usr/src/linux-(kernel-version)

 when i genkernel --config does the /etc/kernel/default-config get 
 overwritten?

The updated configuration is stored in:
/etc/kernels/config-(kernel-version)

 does genkernel make use of make.conf when compiling kernels?

The only option in make.conf that is used is MAKEOPTS.

 when does the /usr/src/linux/.config get written to?

The .config file is written before the 'make oldconfig' step and then if
your using '--config' will be updated with your changes.

 i'm currently using the gentoo gaming-source, because my nforce2 board was 
 having problems with gentoo-source.  i just found out that the kernel did 
 not have usb printing enabled, so i'm interested in compiling a different 
 kernel possibly 2.4.22 (any thoughts on this and nforce2 mboard?)  i want to 
 preserve my config files for each of the kernels that i compile, just in 
 case i have to go back and recompile it...  any suggestions...

  I'm running the 2.4.22 aa-sources with an nforce2 chipset, and it has
been working well. Although, the machine isn't used for more than basic
desktop functions (email, web, xawtv, etc). The only issue I've ran into
is that the nforce-audio drivers didn't work, I had to use alsa.

If you use genkernel with the --config option the kernel configurations
will be saved as config-(kernel-version) so it will keep each
configuration separate. For example, the kernel for
gentoo-sources-2.4.20-r7 will be stored as config-2.4.20-gentoo-r7. If
you want to save more than one configuration per kernel version, the
easiest way is to use the 'Save Configuration to an Alternate File'
option during the menu configuration.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Find ebuild from file?

2003-10-12 Thread Doug Weimer
On Sun, 2003-10-12 at 00:27, Bryan Whitehead wrote:
 Thanks for the quick reply...
 
 What about finding files that are not yet installed? For example, if I'm
 looking for what package (ebuild) would provide a file called xrdb.
 
 In Mandrake this is simple:
 [EMAIL PROTECTED] ~]$ urpmf xrdb
 XFree86:/usr/X11R6/bin/xrdb

Unfortunately, I don't know of anything similar for gentoo. You can use
the emerge -s and -S options to search for matching package names and
descriptions, but there's nothing to locate which package installs a
particular file. 

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Dia (again)

2003-10-12 Thread Doug Weimer
On Sun, 2003-10-12 at 00:45, Patrick Marquetecken wrote:

 I have installed Dia 0.9.2- pre4, and i have now my text back, but the
 Cisco drawings are gone  :-(
 
 Can i copy the Cisco drawings from Dia 0.9.1 to my current version?

You might want to file a bug about this. I installed 0.9.2-pre4 earlier
today and the Cisco shapes were installed. You should be able to copy
the 0.9.1 files over though as the /usr/share/dia/shapes/Cisco
directories are the same in .9.1 and .9.2. Don't forget to copy the .9.1
cisco sheets into /usr/share/dia/sheets if they too were not installed
in .9.2-pre4.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] nvidia problem

2003-10-12 Thread Doug Weimer
On Sun, 2003-10-12 at 11:43, Meka[ni] wrote:
 
   Already done that. Could you please check if you have that file on you source 
 tree? I
 really don't what to do anymore.

Did you run mrproper on the kernel tree after compiling? This will
remove the version.h file. To generate it, you can run 'make
include/linux/version.h' from the /usr/src/linux-(kernel-version)
directory. As a example mine contains:

#define UTS_RELEASE 2.6.0-test7-bk3
#define LINUX_VERSION_CODE 132608
#define KERNEL_VERSION(a,b,c) (((a)  16) + ((b)  8) + (c))

This is from development-sources-2.6.0_beta7-r3.ebuild .

Doug



--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] halt run as normal user

2003-10-12 Thread Doug Weimer
On Sun, 2003-10-12 at 13:19, Jan Meier wrote:
 Hi,
 how can i configure halt, that it can be run with a normal user account?
 Thanx in advance 
 Jan

There are several options, some of which are mentioned here:
http://forums.gentoo.org/viewtopic.php?t=67151highlight=halt+user . I
personally use sudo and allow a group of users to use shutdown. Sudo can
be found in the app-admin/sudo package and you'll want to look at
/etc/sudoers and the sudoers man page for configuration examples.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] checksum scanning

2003-10-11 Thread Doug Weimer
On Fri, 2003-10-10 at 14:09, Andrew Gaffney wrote:

 I was looking for something that was a bit more verbose. If nothing 
 exists that does what I want it to do, I'll just write a Perl script 
 that grabs the MD5 from /var/db/pkg/*/*/CONTENTS and compares it to the 
 actual file. I want to be able to see the name, mtime, old size, new 
 size, etc. of the file that doesn't match. I'd run it as a cron job and 
 have it email me. I run 'emerge sync  /dev/null; emerge -upDv world; 
 emerge -uDf world' as a cron job once a day and have it email me the 
 results.

You can take a look at /usr/lib/portage/bin/chkcontents . Given a
CONTENTS file it checks md5sums and symlinks.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Blocked Deps.. how do I get rid of this?

2003-10-09 Thread Doug Weimer
On Thu, 2003-10-09 at 09:32, Paul K. Dickson wrote:
 Yes, I have to unmerge both, then I type 'emerge -u xfreewhatever, and
 it emerges XFT along with it, then the conflict is back again.  Quite
 irritating:)

'emerge -u xfree86' for all ebuilds 4.3.0-r1 are now blocking xft and
they wont bring it in as a dependency. Your virtual file might be
bringing it in. Take a look at /var/cache/edb/virtuals and find the line
with virtual/xft in it, do you have x11-libs/xft in that line? If so,
remove it. Then 'emerge -u xfree86' and it shouldn't be building xft any
longer.

Good Luck,

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Blocked Deps.. how do I get rid of this?

2003-10-09 Thread Doug Weimer
On Thu, 2003-10-09 at 14:59, Doug Weimer wrote:

 'emerge -u xfree86' for all ebuilds 4.3.0-r1 are now blocking xft and
 they wont bring it in as a dependency. Your virtual file might be
 bringing it in. Take a look at /var/cache/edb/virtuals and find the line
 with virtual/xft in it, do you have x11-libs/xft in that line? If so,
 remove it. Then 'emerge -u xfree86' and it shouldn't be building xft any
 longer.

Sorry, xfree86 should be xfree in all instances above. I always mess
that up.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] nforce2 audio

2003-10-07 Thread Doug Weimer
On Mon, 2003-10-06 at 21:34, a park wrote:
 i'm trying to set up my gigabyte nforce2 with sound.  i'm using the 
 2.4.20-gaming kernel and have emerge nforce-audio and have placed the name 
 of the driver in the /etc/modules.autoload.d/kernel2.4 file.  when i run 
 update-modules command i get the following error message:
 
 depmod: *** Unresolved symbols in 
 /lib/modules/2.4.20-gaming-r3/kernel/drivers/net/starfire.o
 depmod: *** Unresolved symbols in 
 /lib/modules/2.4.20-gaming-r3/kernel/drivers/sound/nvaudio.o

This isn't an exact answer to your question but have you tried the alsa
drivers? I had this same problem earlier and the alsa drivers were the
easiest solution. If you want to give them a try, take a look at
http://www.gentoo.org/doc/en/alsa-guide.xml . The specific alsa module
you want is snd-intel8x0 .

Good Luck,

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] genkernel problem

2003-10-07 Thread Doug Weimer
On Tue, 2003-10-07 at 08:51, SMS WebMaster wrote:

 I installed vanilla-source and now when I want to compile it (after 
 coping .config to /etc/kernels/default-config)

 My question is why genkernel said :  config type: non-gentoo (default)  

The non-gentoo (default) message is simply the message that goes 
along with using /etc/kernels/default-config. Genkernel currently looks 
for three config files in the following order:

/etc/kernels/config-(kernel-version) 
reported as gentoo (customized)
/usr/src/linux-(kernel-version)/kernel-config
reported as gentoo (default)
/etc/kernels/default-config
reported as non-gentoo (default)

Genkernel will use the first config it finds, so if 
config-(kernel-verison) exists then kernel-config and default-config
will not be read. Each of these configs is used in the exact same way, 
so the 'non-gentoo' message only means that genkernel didn't find one 
of the other configs. The only thing to watch out for is if you run 
'genkernel --config' and customize your settings, then genkernel will 
automatically copy the new config to 
/etc/kernels/config-(kernel-version), 
so from then on default-config will have no effect.

Hope this helps,

Doug

PS: (kernel-version) in the above refers to everything after linux- in 
the directory your /usr/src/linux symlink points to. So if 
/usr/src/linux links to /usr/src/linux-2.4.20-gentoo-r7, then 
(kernel-version) will be 2.4.20-gentoo-r7.


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] file manager suggestion

2003-10-03 Thread Doug Weimer
On Fri, 2003-10-03 at 05:16, Mark Knecht wrote:
 On Thu, 2003-10-02 at 13:57, Klaus Neumann wrote:
 
  
  FileRunner is my favorit.
 
 Is thee an ebuild for this?

It doesn't appear to be in portage yet, but an ebuild for the latest
version is here: http://bugs.gentoo.org/show_bug.cgi?id=15329

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] NFS vs. samba

2003-10-03 Thread Doug Weimer
On Sun, 2003-09-21 at 08:33, Marshal Newrock wrote:
 On Sun, 21 Sep 2003, Janne Johansson wrote:
 
   If the laptop's not plugged in, after about 10-20 seconds rpc gives up
   with destination unreachable and the filesystem just doesn't mount.
 
  Yup. If you are worried about the delay, you can always take it out of
  the fstab and mount it manually.
 
 It seems to me that a speedup would be to first attempt to ping the host,
 and if that succeeds, then try to mount the NFS share.  This would have to
 be optional, of course, since some hosts may be configured not to reply to
 pings.

A bit late on this, but you can setup dhcpcd to execute a script when a
machines ip is updated. Just add -c /etc/dhcpc/dhcpcd-eth0.exe to your
dhcpcd_eth0= variable in /etc/conf.d/networks. Then every time dhcpcd
receives an update it will run the script dhcpcd-eth0.exe with the path
to the new info file and some other parameters (they're in the man
page).

The info files are setup as VAR=value so you can just source $1 in the
script to get the new info, then you can do things like:

if [ ${DHCPSHADDR} = XX:XX:XX:XX:XX:XX ]; then
#mount nfs shares/samba shares here
fi

Just put your dhcp servers mac address in for XX:XX... and it will run
the code whenever eth0's ip is modified by dhcp. There are a few
caveats, but that is the basic idea. If anyone has questions, let me
know and I'll write up something with more in depth examples.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] emerge -Uvp hosed

2003-09-28 Thread Doug Weimer
On Sun, 2003-09-28 at 12:46, Shawn wrote:

  
 Calculating world dependencies
 Traceback (most recent call last):
   File /usr/bin/emerge, line 2096, in ?
 if not mydepgraph.xcreate(myaction):
   File /usr/bin/emerge, line 996, in xcreate
 if portage.db[/][vartree].dbapi.match(x):
   File /usr/lib/python2.2/site-packages/portage.py, line 3193, in match
 mydep=dep_expand(origdep,self)
   File /usr/lib/python2.2/site-packages/portage.py, line 2629, in dep_expand
 return prefix+cpv_expand(mydep,mydb)+postfix
   File /usr/lib/python2.2/site-packages/portage.py, line 2562, in cpv_expand
 if (not mydb.cp_list(mykey)) and virts and virts.has_key(mykey):
   File /usr/lib/python2.2/site-packages/portage.py, line 3145, in cp_list
 mystat=os.stat(self.root+var/db/pkg/+mysplit[0])[ST_MTIME]
 TypeError: stat() argument 1 must be (encoded string without NULL bytes), not str
 

Do either of the solutions here:
http://bugs.gentoo.org/show_bug.cgi?id=22701 fix your problem? Since
regenworld didn't work for you, try running 'rm -rf /var/cache/edb/dep/*
 emerge sync'.

Good Luck,

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] nvidia use flag?

2003-09-25 Thread Doug Weimer
On Wed, 2003-09-24 at 17:18, James Hanna Jr wrote:
  -BEGIN PGP SIGNED MESSAGE-
 Hash: SHA1
 
 Is there anyway to get Gentoo to NOT install nvidia-kernel and
 nvidia-glx when I emerge -upD world?
 
 I uninstalled them and it wants to reinstall them, which always
 results in an error since I'm running 2.6-test5

This is probably due to nvidia-glx being used by virtual/opengl. For a
quick check run: `grep nvidia /var/cache/edb/virtuals` . If you have an
nvidia-glx entry there, remove it. That should get rid of the dependency
that's pulling the nvidia ebuilds into an emerge world.

Good Luck,

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] gcc-problem?

2003-09-22 Thread Doug Weimer
On Mon, 2003-09-22 at 13:54, Hemmann, Volker Armin wrote:
 On Monday 22 September 2003 17:36, Michael W. Holdeman wrote:
snip
  checking for C compiler default output... configure: error: C compiler
  cannot create executables

 yes, check your CFLAGS, bet you made a mistake. This is the 'standard' error 
 message, when gcc hits an 'unknown' option.
 
 Glück Auf
 Volker

Just for future reference, configure errors often show up in the
config.log file in the work directory for the package you are building.
For example, the gettex-0.11.5 work dir is in
/var/tmp/portage/gettext-0.11.5-r1/work/gettext-0.11.5 by default. This
directory wont be cleaned up if there is a build error, so you should
still have it. Take a look at the config.log file for the information
following the line: configure:1817: checking for C compiler default
output, hopefully there will be a bit of information about the error.

Here's what the error will look like if one of your CFLAGS are wrong:

configure:1817: checking for C compiler default output
configure:1820: gcc -ftracer -mfpmath=sse -mpni -march=pentium4 -O2
-pipe -fomit-frame-pointer -fVERYWRONG   conftest.c  5
cc1: error: unrecognized option `-fVERYWRONG'
configure:1823: $? = 1
configure: failed program was:
#line 1801 configure
#include confdefs.h

int
main ()
{

  ;
  return 0;
}
configure:1846: error: C compiler cannot create executables

Good Luck,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] A bit of scripting...

2003-09-21 Thread Doug Weimer
On Sun, 2003-09-21 at 00:48, bob bob wrote:
 Any of you guys any good as scripting?
 I'm trying to do a smarter version of this :
 
 ls -lR /lib/modules/*/kernel/  /etc/modules.autoload.d/kernel-2.4
...
 /lib/modules/2.4.20-gentoo-r6/kernel/arch/i386/kernel:
 total 28
 -rw-r--r--1 root root19016 Sep  8 06:57 apm.o
 -rw-r--r--1 root root 3044 Sep  8 06:57 cpuid.o
 -rw-r--r--1 root root 3848 Sep  8 06:57 msr.o
 
 
 Now out of that all I want is :
 apm
 cpuid
 msr

If you only wan the files you can use find with a -type of f:

find /lib/modules/*/kernel/* -type f -printf %f\n |sed 's/.o//g'

Although since your redirecting the output to
modules.autoload/kernel-2.4 I would think that you might want to
restrict the output to uniq items. Otherwise if you have more than one
tree in /lib/modules you'll probably end up with duplicate modules. So
something like:

find /lib/modules/*/kernel/* -type f -printf %f\n |sed 's/.o//g' |
sort | uniq

will prevent duplicate module entries.

Good Luck,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] A bit of scripting...

2003-09-21 Thread Doug Weimer
On Sun, 2003-09-21 at 01:22, Doug Weimer wrote:

 If you only wan the files you can use find with a -type of f:
 
 find /lib/modules/*/kernel/* -type f -printf %f\n |sed 's/.o//g'

I need to correct a small mistake here actually, s/.o//g isn't quite
right. Use s/\.o//g instead. With only .o it will return improper
results.

Sorry about that,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] rar file decompressor?

2003-09-21 Thread Doug Weimer
On Sun, 2003-09-21 at 11:00, Mark Knecht wrote:
 Hi,
What tools can I use to decompress a rar file under Gentoo?

$ emerge -upv unrar

These are the packages that I would merge, in order:

Calculating dependencies ...done!
[ebuild  N] app-arch/unrar-3.2.2  

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] Alt-Left/Right arrows on the console.

2003-09-21 Thread Doug Weimer
On Sun, 2003-09-21 at 11:58, Carlos wrote:
 Hey people,
   When I'm on the console and press Alt+Left or Right arrows it switches to the 
 previous/next virtual terminal. I find this pretty annoying since I like to use 
 those key bindings to switch irssi's windows. Is there any way to disable that 
 behaviour?

The Alt+{Left,Right} functionality is setup by your chosen keymap file.
The files you want to modify are in /usr/share/keymaps/i386/include .
You'll want to edit linux-keys-bare.inc and comment out the following
lines by placing a # in front of them:

alt keycode 105 = Decr_Console
alt keycode 106 = Incr_Console

After these changes, find the KEYMAP variable in /etc/rc.conf and then
run 'loadkeys KEYMAP'. For example, if you have KEYMAP='us', run
'loadkeys us' from the console.

Good Luck,

Doug

P.S: A few non-us keymaps use linux-with-modeshift-altgr.inc instead, so
if the above fix doesn't work for you, try changing the Decr_Console and
Incr_Console lines here as well.


signature.asc
Description: This is a digitally signed message part


RE: [gentoo-user] cron environment

2003-09-21 Thread Doug Weimer
On Sun, 2003-09-21 at 20:48, Gregory Symons wrote:

 I believe it gets the full environment the the user has; i.e. when it
 launches the shell, the shell does its normal initialization and sources
 the global environment and whatever the user has in his
 .profile/.bashrc/.insert init file here. The man page isn't too clear
 on this point (according to it, the only environment that gets set is
 SHELL, LOGNAME, HOME and MAILTO, along with any environment set in the
 user's crontab itself), so if someone knows better than me, feel free to
 correct me.

I just ran a quick test to see:

/etc/crontab entry:

*/1 * * * * root/bin/mycrontest

/bin/mycrontest:

#!/bin/bash
echo *Environment:
env

results:
*Environment:
SHELL=/bin/sh
PATH=/usr/bin:/bin
PWD=/root
HOME=/root
SHLVL=2
LOGNAME=root
_=/usr/bin/env

No settings were set from ~/.profile or /etc/profile. I ran a similar
test with the 'set' command and it did not have anything beyond the
typical shell/user information. I'm not sure where vcron is getting PATH
or SHELL though, defaults compiled in perhaps?

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] Enable spell checker in Evolution?

2003-09-19 Thread Doug Weimer
On Fri, 2003-09-19 at 08:29, Angel Gabriel wrote:
 How can i get the spell chcker working in evolution?

In evolution 1.4.4 the spell checking option is at Tools-Composer
Preferences-Spell Checking. The evolution ebuild checks for the USE
flag spell before installing the gnome-spell package. If you can't find
the spell checking option within evolution run `emerge -pv evolution`
and see if the spell flag is set (+spell will be shown in the list of
flags). If you don't have it, you'll need to re-emerge evolution with
the spell flag.

Good Luck,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] opengl

2003-09-12 Thread Doug Weimer
On Fri, 2003-09-12 at 07:08, Collins Richey wrote:
 On Thu, 11 Sep 2003 23:17:00 -0400
 donnie berkholz [EMAIL PROTECTED] wrote:
 
  On Thu, 2003-09-11 at 19:02, Svein Harald Soleim wrote:
snip
   dream but some other programs compline about 
   error while loading shared libraries: libGLU.so.1: cannot open
   shared object file: No such file or directory

Do you have the libGLU files? Do a ls -l /usr/lib/libGLU* and see if the
.so.1 symlink has been created. On my box:

$ ls -l /usr/lib/libGLU.so.1
lrwxrwxrwx1 root root   13 Aug 31 03:38
/usr/lib/libGLU.so.1 - libGLU.so.1.3

$ qpkg -f -v /usr/lib/libGLU.so.1
x11-base/xfree-4.3.0-r3 *

It may be possible that the libGLU.so.1 symlink wasn't setup right, even
though you have the actual libGLU.so.1.x library installed.

 I find this quite banal.  It seems that the most common recommendation
 (without any justification or reference to bugs) is to re-emerge
 pick-a-package.  If xfree didn't emerge properly the first time, why
 would it work now?
 
Well it does seem like an error in the xfree install process, so with a
re-emerge you should be able to end up with a proper install, a reported
error in making the symlink, or (if no error is reported) a bug for that
version of the ebuild. I think a re-emerge is often useful, but in your
case some investigation is probably worthwhile.

Good Luck,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] opengl

2003-09-12 Thread Doug Weimer
On Fri, 2003-09-12 at 09:50, Svein Harald Soleim wrote:

 Calculating dependencies -
 !!! all ebuilds that could satisfy virtual/opengl have been masked.
 !!!(dependency required by kde-base/kdebase-3.1.3 [ebuild])
 
 !!! Error calculating dependencies. Please correct.

What do you have in /var/cache/edb/virtuals for virtual/opengl? If the
only thing providing opengl is xfree you should have a the line:
virtual/opengl x11-base/xfree . If you have another package before
x11-base/xfree and it happens to be masked, then you'll get the above
error.

Good Luck,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] MySQL-4.0.15 -- when?

2003-09-11 Thread Doug Weimer
On Thu, 2003-09-11 at 10:20, Matthias F. Brandstetter wrote:
 Hi all,
 
 just wanted to ask when the new MySQL-4.0.15 ebuild will be out, does 
 somebody know? Just because of this password / buffer overflow 
 vulnerability of prior versions...
 

Not sure about the 4.0.15 ebuild, but the 4.0.14-r2 and 4.0.13-r4 have
been patched for the vulnerability. See:
http://bugs.gentoo.org/show_bug.cgi?id=28394 


Good Luck, 

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] How many packages in system not world?

2003-09-09 Thread Doug Weimer
On Tue, 2003-09-09 at 07:44, Dhruba Bandopadhyay wrote:
snip
 Thanks.  I simplified that to:
 
 emerge -Oevp system | grep -c ebuild
 
 It seems it cannot be done any faster e.g. from a file.

Take a look at /etc/make.profile/packages . From the intro
documentation:

# An initial * marks a package that is part of the
# official base system profile.  If there's a *, then emerge
system will
# use the line in its calculations of what should be installed for
this
# profile.

So a:

grep -c ^* /etc/make.profile/packages

should do what your looking for.

Good Luck,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] PERL and 'mail notification'

2003-09-07 Thread Doug Weimer
On Sun, 2003-09-07 at 08:01, Jason Stubbs wrote:
snip
 # Can I combine the following two lines in any way?
 # $mail_body = @$mail-body(); does not work.
 $mail_body_ref = $mail-body();
 @mail_body = @$mail_body_ref;

@{$mail-body()} works here. If you want, you can also use
foreach(@{$mail-body(}) below.


 @body_text = split(\n, join( , @body_text));
 
 @ntfy_mail = (Date:   $msg_date,
   From:   $msg_from,
   Subject:$msg_subj,
   \n);
 
 # Can I combine this into one line? Can I combine it with the above?
 for ($i = 0; $i  20; $i++) {
 @ntfy_mail = (@ntfy_mail, @body_text[$i]\n);
 }
 

Since you only want 20 lines you can use @body_text[0 .. 19] in the
join. If you want to reduce everything to a few lines, you can use a
second join to add the newlines:

$text = join(\n, split(\n, join( , @body_text[0 .. 19])));

or use map to keep the result as an array:

@body_text = map { $_ .\n } split( same as above );

Good luck,

Doug


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-user] new kernel - depmod problems

2003-09-04 Thread Doug Weimer
On Thu, 2003-09-04 at 20:14, Mark Knecht wrote:
snip
 QUESTION #1 - Is it a requirement that I need to rebuild xfree each time
 I build a new kernel? (Time consuming, but done.)

I've never had a problem using X11 after compiling a new kernel, but I
can't say for certain whether or not this is necessary for specific
configurations.

 The problems I'm seeing right now are:
 
 1) depmod -a produces a large number of errors:
snip
 I believe I did:
 
 make dep clean bzImage install modules modules_install

Is the new kernel the same version as your old one? I ask because if
both kernels are version 2.4.20-gentoo-r2 then there may be modules
lying around /lib/modules/2.4.20-gentoo-r2 that you did not configure
into the new kernel. If you are recompiling a kernel it is best to move
the old module tree out of the way before running 'make
modules_install'. 

 Probably related to question #2, the USB hotplug page talks about
 modprobing the USB drivers (modprobe usb-ohci.o) but when I try this it
 tells me it cannot find the file:
 
 Wizard usb # modprobe usb-ohci.o 
 modprobe: Can't locate module usb-ohci.o
 Wizard usb # 
 

This should be 'modprobe usb-ohci', don't provide the .o when using
modprobe. 

Good luck,

Doug



--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] new kernel - depmod problems

2003-09-04 Thread Doug Weimer
On Thu, 2003-09-04 at 20:51, Mark Knecht wrote:
 On Thu, 2003-09-04 at 20:42, Doug Weimer wrote:

  If you are recompiling a kernel it is best to move
  the old module tree out of the way before running 'make
  modules_install'. 
 
 It was, and this could be part of the problem. Thanks. I guess I can
 still just do that since it's the last step in the process, correct?
 

Yep, as long as you didn't run 'make clean' after the last install.

Good Luck,

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] Eroaster, problems with ebuild

2003-09-03 Thread Doug Weimer
On Wed, 2003-09-03 at 18:08, Ernie Schroder wrote:
 Is anyone else getting this?
 # emerge -Uv eroaster
  --upgradeonly implies --update... adding --update to options.
 Calculating world dependencies ...done!
  emerge (1 of 1) app-cdr/eroaster-2.1.0-r2 to /
 
 !!! File is corrupt or incomplete. (Digests do not match)
  our recorded digest: 217818c76031049ae9c7aece980d7e40
   your file's digest: 7792463490210872d9e736da80c5328d
 !!! File does not exist: /usr/portage/distfiles//eroaster-2.1.0.tar.gz

Take a look at http://bugs.gentoo.org/show_bug.cgi?id=27839 . Basically
different mirrors have different versions of eroaster-2.1.0.tar.gz
causing the digest to be wrong. The digest is based on the
sourceforge.net tarball so grab a copy of eroaster-2.1.0.tar.gz from one
of the sourceforge mirrors and put it in /usr/portage/distfiles.

Doug


--
[EMAIL PROTECTED] mailing list



Re: [gentoo-user] problem pinging

2003-09-02 Thread Doug Weimer
Tom,

I would guess that your problem is due to having both devices
initialized when only one is actually connected. Due to the change to a
static ip, eth0 will be initialized whenever the nic drivers are loaded,
regardless of whether or not eth0 is actually connected to the network.
To test this possibility take down eth0 when you are trying to use just
the wireless connection with 'ifconfig eth0 down' and then ping the
router. For a bit more in depth explanation see my responses below.
 
On Mon, 2003-09-01 at 17:08, Tom Hosiawa wrote:
snip
 Yeah I'm running both nic's simultaneously, but use the wireless
 connection most of the time.
 
If you are only using the wireless connection this is likely to cause
problems when communicating with local systems as the data may be sent
over eth0 or 192.168.1.3. For example your ping results: 

 or ping 192.168.1.2 (the other pc on the lan), I get:
 PING 192.168.1.2 (192.168.1.2) 56(84) bytes of data.
 From 192.168.1.3 icmp_seq=1 Destination Host Unreachable
 From 192.168.1.3 icmp_seq=2 Destination Host Unreachable
 From 192.168.1.3 icmp_seq=3 Destination Host Unreachable

^ 192.168.1.3 is referenced as eth0 based on your ifconfig results.

 ifconfig results:
 eth0  Link encap:Ethernet  HWaddr 00:08:0D:8A:6A:0A
   inet addr:192.168.1.3  Bcast:192.168.1.255  Mask:255.255.255.0
snip
 eth1  Link encap:Ethernet  HWaddr 00:06:25:2B:1A:54
   inet addr:192.168.1.4  Bcast:192.168.1.255  Mask:255.255.255.0

You seem to be able to access the outside world due to the specification
of eth1/192.168.1.1 as your gateway. Your routing table confirms this: 

 0.0.0.0 192.168.1.1 0.0.0.0 UG1  0   
0
 eth1

As an aside you are unable to ping tsn.ca because it doesn't appear to
respond to icmp echo requests. From my machine:

$ ping -c4 tsn.ca
PING tsn.ca (199.246.67.211) 56(84) bytes of data.

--- tsn.ca ping statistics ---
4 packets transmitted, 0 received, 100% packet loss, time 3013ms

but...

$ ping -c2 www.google.com
PING www.google.com (216.239.51.99) 56(84) bytes of data.
64 bytes from 216.239.51.99: icmp_seq=1 ttl=49 time=89.0 ms
64 bytes from 216.239.51.99: icmp_seq=2 ttl=49 time=88.0 ms

--- www.google.com ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1005ms
rtt min/avg/max/mdev = 88.007/88.505/89.003/0.498 ms

Hope this helps,

Doug


--
[EMAIL PROTECTED] mailing list