could enable additional error locations
which can get handy if you start passing a lot of functions around..
What do you think?
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
http://www.haskell.org/pipermail/haskell-cafe/2007-July/028012.html
There have been some outdated statements that nix dosn't work on windows
etc. I've tried correcting them. But perhaps you can also have look at
this thread and give a more accurate response if you feel to do so ?
Marc Weber
How is this done?
Gentoo has /etc/env.d/ which looks like this
/etc/env.d/
/etc/env.d/gcc
/etc/env.d/gcc/avr-3.4.6-hardenednopie
/etc/env.d/gcc/avr-3.4.6-hardenednossp
/etc/env.d/gcc/avr-3.4.6
/etc/env.d/gcc/avr-3.4.6-hardened
/etc/env.d/gcc/i686-pc-linux-gnu-4.1.1
/etc/env.d/05gcc
for this implemented behaviour?
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
?
I've copied the relenvant lines from the builder now so it works.
Perhaps you can give your comment on what would be the best nix style to
do this..
Marc Weber
= ===
source genericBuild:
genericBuild() {
header building $out
as
installation description.. Then we no longer need the name attribute..
At least this would make the most sense to me..
But you can forget about binary installs then thingking of the
ghcWrapper ...
Is this all just nonsense or the way to go?
nix is crazy cool!
Marc Weber
Any comments?
It's not that good .. because you have to mention the flags for
configuring and dependencies even if it's just [].
{
flags = {
withX = { configure = --with-X; deps = [libX] };
mandatory = [ ... ]
would be better because you can sepecify configure flags and
dependencies at
We'd need either some kind of substring, beeing able to use __head on
There is __substr ..
M
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
note how useNixLibs implies jpeg, png and zlib and how
mandatory ensures that the not tested flags cygwin and quartz cannot
be used..
Drawback: I've used
__getAttr __hasAttr listToSet which are not yet stable or in svn
(which could be changed, niksnut ??)
I think its really
But maybe it simpler to just stick your postBuild value in a file that you can
then import in every Nix expression that needs it.
Hi Eelco..
What about adding a new phase purifyExecutables to setup-new2.sh
(which is not executed by default)
then all you would have to do is
Since this is on PowerPC, is it the case that the environment is not
totally pure, e.g., my Debian GCC (4.2.3) gets used instead of the one
from Nix or something like that?
Hi Ludovic:
I don't know why this trouble arises.
You are right, nix isn't totally pure.
Eg a lot of package do contain
Some time ago I've added php_unstable (configurable).
When using php = php_unstable instead of the existing php expression
you'll get MySQL support.
I think most people using PHP will need MySQL as well.
Does anyone mind me removing the existing php and replacing it by
php_configurable?
Marc
Current:
To create a new upstart-job (such as mysqld/ httpd) you have to change
upstart-jobs/default.nix
and write
upstart-jobs/yourjob.nix
example:
# DHCP client.
++ optional [networking useDHCP]
(import ../upstart-jobs/dhclient.nix {
inherit (pkgs) nettools
In case you want to tell us if you go to FOSDEM as well you can enter it
here:
http://mawercer.de/nix/nix.php
login and pwd is both nix
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
this?
I'm not sure how X window managers are implemented right now.
But I'd like to have something like this there as well. I'll
have some more spare time the next days.
Sincerly Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl
Sometimes nix only says:
unkown xy in all-packages.nix..
Oh my god where exactly is this happening?
This patch for nix might provide some more information in these cases.
http://mawercer.de/diff_nix_more_error_messages
Maybe it's useful for you as well?
maybe you want to set moreErrorLocations =
regex)
But I don't have need for it.
Thank you for any feedback converning my code/ changes whatsoever.
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
Personally, I can live with the current scheme.
That's because we have not that much services right now.
But this will change in the future because I hope that some more people
start using nixos in the future :)
May I create a services and sytem.nix branch to start working on this?
Then
I have the same feeling. This construct also makes it easy to implement
ambient authority [0], where you'd just pass anything to the Nix
expression, making it hard to determine what it really needs.
Agreed. But how to you test wether something is really needed?
I do be removing it and seeing
: On multiuser system there may be dozens of different
versions of the same lib because everyone has installed the application
at another point in time..
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
Hi
today I've noticed that you can pass NIX_CONFIG and NIXOS when running
nixos-rebuild.. However the code compiling nix first has the hardcoded
path
/etc/nixos/pkgs .. does this make sense?
Maybe we should introduce
NIXOS_PKGS here as well ?
Marc Weber
On Thu, Mar 13, 2008 at 01:39:33PM +0100, Marc Weber wrote:
We're lacking a feature:
Every single person can choose it's version.
However the system administrator has no chance of defining a config
default which may be overridden by the user.
So what about
/etc/nixpkgs_conifg.nix ?
which
Is anyone using services without upstart-jobs?
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
for it,
right?
Feel free to exchange remove an add topics
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
Hi Andres,
Building on impure systems:
ptrace/ strace ? Is a sandbox approach no longer regarded?
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
On Thu, Mar 13, 2008 at 04:35:20PM +0100, Eelco Dolstra wrote:
Marc Weber wrote:
Is anyone using services without upstart-jobs?
Yes. The Subversion server on svn.cs.uu.nl uses the services tree but
doesn't run NixOS. And until a few weeks ago there were a bunch of other
machines
Hi.
I've finshed my proposal. it should integrate within the existing system
so you can choose which one to use.
The only thing which is left to be done is option checking but the
implementation is already there.
Example given here:
http://rafb.net/p/g0oRw476.html
note the attr set args at
What do you think of having multiple X-Servers with different window
managers?
Or what about choosing the window manager when logging in?
Does someone like to work with me on this?
I'd like to introduce something similar i've done with services. Just a
simple file
{ pkgs, ... } : {
// kind of
I get your idea, but with really runtime dependencies, you do not need to
feed them to the component expression at all
I disagree here. It can be seen as documentation.
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
Were you ever able to run vncserver? Because I run it with no problems. I
hope your commit at least won't break it..
No. Then I think I should revert the commit.
In which way are you running it?
Using vncserver ... ?
I guess you've seen my error message within the svn commit.
Sorry
Marc
I've trouble downloading the kdepim and kde4libs packages using
nix-prefetch (wget works fine).
It's not saving the tar.bz2 but a xml file containing mirrors.
Do you have this problem as well?
eg source contained in pks/desktops/kde-4/libs/default.nix
Marc
Configure says either lib is missing
May I add it ?
Which one is preferable?
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
We've talked about running testing features recently..
Best qualty assurance can only be given if the test runs on your machine
because CPU's might differ.. But don't know wether it's worth creating a
wrapper wrapping the package and running tests which is not included
within the binary build
recover?
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
collector options:
--print-roots: print GC roots and exit
--print-live: print live paths and exit
--print-dead: print dead paths and exit
default : delete dead paths
= ===
then ?
Marc Weber
[EMAIL PROTECTED]:~]$ sudo nix-collect-garbage
error: path `/nix/store/3qcjl6qv6mz2kr42z8fy2l83ghy0q225-env-manifest' is not
valid
0 bytes (0.00 MiB) freed
The path doesn't exist and the nix-store --verify output doesn't contain
that hash. The whole /nix directory (including store) doesn't
/RpathIssue, it looks like libtool adds it by
default but it is switched off with --disable-rpath. How does Nix
deal with packages that don't use libtool?
Have a look at all binary distributions (eg acroread)
There is the tool patch-elf --rpath mypath1:mypath2 executable
HTH
Marc Weber
do you agree that kind of option description for nixpkgs would be useful
as well? (each time getConf is used) ?
Where is the right place to put it?
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
/mercurial and cvs repositories I've written the
somewhat unofficial nix-repository manager. It's a small haskell tool
fetching the repository once making it easy to update and publish the
new versions to other nix users. Don't think this is waht you need here
though.
Marc Weber
I've read a thread about why kde has chosen cmake once..
And they struggled using scons. Once switching to cmake they had less
trouble. So I guess cmake is the way to go.
http://lwn.net/Articles/188693/
But this is dated 2006.
Marc Weber
___
nix-dev
expression takes 1m20s.
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
On Wed, Apr 16, 2008 at 09:51:48AM +0200, Ludovic Courtès wrote:
Marc Weber [EMAIL PROTECTED] writes:
Only reinstalling the already installed expression takes 1m20s.
Using nix-env -i, nix-env -Ai, or any of these?
always -Ai / -iA
I htink the trouble is that the libraries it evaluates
thoughts only
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
.
And I'd really like to see a nixpkgs manual to put all this stuff in.
Eelco told me to wait because he wants to start it. That's fine.
Maybe I should start podding him occasionally I guess he will tell me.
Thanks for your reply - that's the only way to keep improving
Marc Weber
need is a substitute for Eelco so that he has time
to work on this :) But finding such a person will be terribly hard
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
(cygwin default)..
Does someone mind me moving the files to cygwin default locations (cyg
dist only of course)
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
On Fri, May 09, 2008 at 01:38:08AM +0200, Marc Weber wrote:
When adding a cygwin port do you think we should move the configuration
file to /etc/nix.conf? Nobody (not knowing nix) will find the file in
/nix/etc/nix.conf without reading the manual...
niksnut has told me on irc that it's ok
think intsalling libraries should be as easy as
either using all or [ pkg1 pkg2 ... ] to your personal .nixpkgs/config.nix
this will be added to (note 1) in the near future..
Do you have another suggestion?
Marc Weber
note 1:
You define your wrapper this way
instead of lib.flattenAttrs
like Nix so far! Thanks Nix team.
You're welcome.
Sorry. I wasn't sure that I can give you good answers on your other
questions. So I hope someone else will do.
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo
] ( pkgs: {} ) ) pkgs;
to all-packages.nix
I'm already using this for ghc68extraLibs an ghc68executables.
I'd prefer having a nixpks manual with tips etc first before splitting
all packages.
Do you think this one svn repo slows down developement?
Sincerly
Marc Weber
familiar with unixODBC internals and has a
nice idea?
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
* subaterm[MAX_ARITY+3];
MachineWord word[MAX_ARITY+3];
} *ATerm;
I'm not that familiar with C so is there a special compiler option to
make this code pass?
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo
wonder what will happen if you enable unicodeSupport and do not
disable cplusplusSupport.. doesn't this result in
--enable-unicode-properties --enable-shared
--disable-static--disable-cpp ?
^^ no
space
Marc Weber
The git expression contains:
+ ''# Install man pages and Info manual
make PERL_PATH=${perl}/bin/perl cmd-list.made install install-info \
-C Documentation '' # 2
+ (if guiSupport then ''
# Wrap Tcl/Tk programs
for prog in gitk git-gui git-citool
about dependencies. Maybe you want to try it
and give me some feedback?
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
of everything after changing it ..)
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
time to be spend on maintainance of
the packages..
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
What do you think about adding
for dir in $out/{man,share/man,info,share/info}/; do
[ -d $dir ] find $dir -type f -print0 | xargs -0 gzip
done
at the end of the fixupPhase in the setup.sh?
This will cause rebuild of almost everything..
Marc Weber
$ gitk
/nix/var/nix/profiles/per-user/ludo/profile/bin/gitk: line 3: exec: wish:
not found
Can we either remove the option altogether (my preference), or make sure
Tcl/Tk programs don't get installed when the option is disabled?
* git: always install the Emacs mode but don't
I've implemented echoing the note and returning exit status 1 now.
See my last commit.
and returning nonzero exit code so automatic tests won't as likely think
it's working
How is automatic testing of gitk and git-gui done? :)
(I know there are tools.. but It's cumbersome)
Anyway thanks for
/laptop-mode
and
/etc/sysconfig/laptop-mode
(Debian and others)
Which location is best for nixos? Just /etc/lapotp-mode?
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
- Ideas about improving nix (such as spearating doc, runtime, dev files
of packages, ideas on decentralizing nix (distributed vcs etc))
- A lot of howtos (ndiswrapper nixos) etc even if there is only the note
doesn't work yet
Of course they must reference the manual if appropriate.
-
Doing some uml experiments I got the same error (no upstart)
Still don't know yet why this happens.
As workaround you can try replacing /var/current/system/sbin/upstart
by a existing one from the store.. (either from debug2 console or by
editing the stage2.sh file within the store directly)
You
what about using this line in stage-1 prior to any mounting?
It's unlikely that you'll mount a partition with user option specified
by configuration.nix but not impossible..
Do I have missed a point?
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
to recover from problems.
Or can nix alraedy do this in some way?
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
. thus make -j1 might differ from make -j8
TODO: How to limit access to this feature? Only adminstrators who are
aware of this risk should be able to use it (?)
Sincerly Marc Weber
How to see that it works?
copy paste shell start
mkdir -p /tmp/num_cores_test
cd /tmp
in the past..
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
or such.
Just curious wether we should separate the os nixos from hardware
configuration by adding a new repo trace/computer-configs or such
Another config would be qemu-with-sound-and-network or such ?
I'd say go for it.
Let's see what others are saying
Marc Weber
= [ * * * * * ${pkgs.myecho}/bin/echo
helloworld ]; };
or would you prefer
pkgs : {
# the system configuration
}
making configuration.nix a function rather than an attr set?
The latter would mean less tweaking and less remembering
Marc Weber
___
nix-dev
Thanks for this. Unfortunately, it looks like while nixpkgs and
services works, the clone of plain old nix is failing. Does it work
for you?
Not yet.. I've been uploading it.. done.
Should work now.
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
Which is the best way to catch this case:
You forget to enable perl bindings in subversion,
you enable svnSupport in git and wonder why git-svn can't find all
libraries..
I think the best way is using in subversion
{ inherit perlBindings (1); ... }
and in git expression:
assert svnSupport -
On Sun, Aug 03, 2008 at 12:05:48PM +0200, Pjotr Prins wrote:
Quite a few packages have duplications of version numbers hard-coded.
One way to prevent that is defining a variable 'version'. It would
even be easier if we could get the version from the fetchurl (with the
name of the package) as
Sorry, I should have used a more descriptive commit message.
It adds gitGit with a glob patch so that you can view all commits in
git modifying */udev/default.nix or such..
It also adds another check ensuring that perlbindings are enabled when
installing git with svnSupport = true (another
On Sun, Aug 03, 2008 at 09:55:25PM +0200, Andres Loeh wrote:
Extracting the version from the URL is in general even more difficult than
extracting the version from the name-version pair.
Of course this should only work for some obvious patterns such as
http://.*/([^/]+-[^/]+)\.[^/]*
for this change?
Woulda small script creating one symlink for each name-hash work for you
as well? (maybe in /nix/erots/ ?)
When implementing kind of hooks (after installing/ before removing a
path) this could be done automatically easily..
cu
Marc Weber
___
nix
This is joe's window manager's update.
Thanks, I've updated the url
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
Has anyone thought about how to add bash completion features to nixos
yet?
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
On Thu, Aug 07, 2008 at 11:13:18AM +0200, Pjotr Prins wrote:
coreutils does not compile, for one.
My kernel version has been 2.6.9 on my virtual server
Sorry, do you mean you can't compile corutils with that coreutils patch?
Why did it help you then?
Marc
On Thu, Aug 07, 2008 at 02:11:11PM +0200, Peter Simons wrote:
Marc Weber writes:
Has anyone thought about how to add bash completion features to nixos
yet?
Would installing those files to $out/etc/bash_completion.d/$name suffice?
And then use for i in ~/.nix-profile/etc/bash_completion.d
still works?
Sincerly Marc Weber
diff --git a/upstart-jobs/postfix.nix b/upstart-jobs/postfix.nix
index ebc6f79..a956d86 100644
--- a/upstart-jobs/postfix.nix
+++ b/upstart-jobs/postfix.nix
@@ -133,6 +133,8 @@ in
start on ${startingDependency}/started
stop on never
+env PATH=/var/run
enhancements as well.
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
fatal and put the at the beginning of
activate-configuration.sh? I'd prefer this last solution
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
What does composedArgsAndFun do? (And it's almost certainly not a good
name ;-) Also, I really dislike all those weird nulls in all-packages.nix.
I haven't look at the source, but I guess it's similar to the question
you've asked once: It merges the given attr sets until null is passed.
The
Similarly, how do I create a 32-bits package on a 64-bits machine
(cross build). Can anyone just give a quick example? All the stuff I
try does not work. It could be me ;-).
If you're lucky and you can still rebuild the system you'll have to add
nix = {
distributedBuilds =
buildInputs = [ (glibc.meta.function { kernelHeaders = args2.kernelHeaders; }
null) ];
} args ) null;
. o O ( What a mess )
Marc Weber
I'd cry for some syntactic suggar then :)
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman
How should we store nix-expressions in nixpkgs if there is no special order?
There is:
nixpkgs/maintainers/docs/classification.txt
But I'm not sure wether it contains the new SCI * categories
Compared to your idea using mailinglist categories we could also say
let's try the scheme gentoo is
chroot at startup to jump into /nixos)
The strange thing: I don't see mount doing anything which might cause
this failure. (see strace log
above) and it still fails within the uml environment.. Any ideas?
When symlinking the store ln -s /nixos/nix nix it works.
Marc Weber
I had need for the latest ghc head in a project.
So I've added it and all base and some additional libraries are already
availible. I haven't committed it yet because it only would cause a lot
of noise. I'm telling you this so that you don't have to duplicate my
work if you are interested.
Marc
the system
having build the failing one). But I havent comitted it yet.
This is totally independantn of grub though.
Sincerly
Marc Weber
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
Has some attendee written a very brief summary yet?
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
Glad this message has helped someone .. :-) So it was worth spending
time on it..
You need to know about your custom config:
~/.nixpkgs/config.nix
add
{
git = { svnSupport = true; };
}
there and try again
Sincerly
Marc
___
nix-dev mailing list
On Tue, Sep 23, 2008 at 10:03:39AM +0200, Pjotr Prins wrote:
Hi Everyone,
Thanks for the replies - it is heartening as I was getting a bit
desperate I would get any help through the mailing list. I am not much
of an IRC lover, but maybe I should change that.
There were too many variables
Eh. I think I removed that to try stuff. Only problem is - how do I fix it.
All Nix actions stop:
nix-store --verify :)
Marc
___
nix-dev mailing list
nix-dev@cs.uu.nl
https://mail.cs.uu.nl/mailman/listinfo/nix-dev
The attached patch makes opera compile again and it's not beeing
killed.. What does it do? It's using paatchelf-0.3 !
So what has changed?
Shall I commit this patch or should we fix patchelf?
Still unsure what is going on here exactly...
You'll need the last comitted patch making opera build
Hi Soren,
sorry, I'm very busy at them moment.
I don't see any chance adding it.
I gave up very fast because i didn't know which version to use.
So I stopped after adding kino which is no replacement.
But maybe you can just try and report where you encounterany trouble?
I'm sure you'll get soem
http://bugs.strategoxt.org/browse/NIXPKGS-85
In other words try using the default patchelf and setting dontStrip to true.
It
appears that strip breaks executables that have been patchelfed.
Thanks!
dontStrip=1 fixed it
Marc
___
nix-dev mailing
On Sun, Oct 05, 2008 at 07:19:41PM +0200, Pjotr Prins wrote:
Michael: a broken package is no reason to comment it out. Because of
late binding nothing will happen. I think the only acceptable
behaviour is to *fix* it or remove it if it has been pending for too
long.
Unfortunately it did go
Today I've tried to install nixos within qemu to start work on tuxonice
again..
However I forgot about fileSystems (so what about adding a placeholder ?)
and what about using the new syntax pkgs : config : ?
Or should we replace it by
args; where args contains config and pkgs once and forall?
On Wed, Oct 08, 2008 at 12:55:43PM +, Ludovic Courtès wrote:
Hello!
There are days when one had better stay in bed. Today was one of these
for me.
This morning, I decided I'd rm -rf some of the things under /tmp.
Among these things were `chroot-*' directories. After that rm -rf
On Thu, Oct 09, 2008 at 03:26:33PM -0500, Søren Renner wrote:
message: cannot load module pcidata
libpcidata.a and libpcidata.so do in fact exist
just did a nixos-rebuild test and nixos rebuild boot
xorgconfig
edited xorg.conf by uncommenting font path, module path
1 - 100 of 773 matches
Mail list logo
