GNU bug report logs - #15392
xorg-server FTBFS

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 15392 in the body.
You can then email your comments to 15392 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Mark H Weaver <mhw <at> netris.org>
To: bug-guix <at> gnu.org
Subject: xorg-server FTBFS
Date: Mon, 16 Sep 2013 03:22:58 -0400

Hello all,

xorg-server fails to build from source.  See below for the tail of the
build log.  This is with aae4ead8142d4fd7c674a1e6e302f40469f878c6.

    Regards,
      Mark


--8<---------------cut here---------------start------------->8---
PASS: input
XKB: Failed to compile keymap
Keyboard initialization failed. This could be a missing or incorrect setup of xkeyboard-config.

Fatal server error:
Failed to activate core devices.

Please consult the The X.Org Foundation support
         at http://wiki.x.org
 for help.

FAIL: xtest
PASS: misc
PASS: fixes
PASS: xfree86
========================================================================
1 of 9 tests failed
Please report to https://bugs.freedesktop.org/enter_bug.cgi?product=xorg
========================================================================
make[3]: *** [check-TESTS] Error 1
make[3]: Leaving directory `/tmp/nix-build-xorg-server-1.12.2.drv-0/xorg-server-1.12.2/test'
make[2]: *** [check-am] Error 2
make[2]: Leaving directory `/tmp/nix-build-xorg-server-1.12.2.drv-0/xorg-server-1.12.2/test'
make[1]: *** [check-recursive] Error 1
make[1]: Leaving directory `/tmp/nix-build-xorg-server-1.12.2.drv-0/xorg-server-1.12.2/test'
make: *** [check-recursive] Error 1
phase `check' failed after 2 seconds
builder for `/nix/store/kgpvhgf4rin7syv7cxid78vmlckqqyfg-xorg-server-1.12.2.drv' failed with exit code 1
@ build-failed /nix/store/kgpvhgf4rin7syv7cxid78vmlckqqyfg-xorg-server-1.12.2.drv - 1 builder for `/nix/store/kgpvhgf4rin7syv7cxid78vmlckqqyfg-xorg-server-1.12.2.drv' failed with exit code 1
guix package: error: build failed: build of `/nix/store/kgpvhgf4rin7syv7cxid78vmlckqqyfg-xorg-server-1.12.2.drv' failed
--8<---------------cut here---------------end--------------->8---

Message #8 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: Mark H Weaver <mhw <at> netris.org>
To: 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Mon, 16 Sep 2013 03:38:01 -0400

Mark H Weaver <mhw <at> netris.org> writes:

> xorg-server fails to build from source.  See below for the tail of the
> build log.  This is with aae4ead8142d4fd7c674a1e6e302f40469f878c6.

Sorry, the build log tail I posted was actually from a slightly modified
xorg.scm, where I uncommented the build inputs for 'xkbutils' and
'xkeyboard-config' in the 'xorg-server' recipe.  Here's a corrected
version of the build log, using a pristine copy of
aae4ead8142d4fd7c674a1e6e302f40469f878c6.  The only differences are in
the nix hashes.

      Mark


--8<---------------cut here---------------start------------->8---
PASS: input
XKB: Failed to compile keymap
Keyboard initialization failed. This could be a missing or incorrect setup of xkeyboard-config.

Fatal server error:
Failed to activate core devices.

Please consult the The X.Org Foundation support
         at http://wiki.x.org
 for help.

FAIL: xtest
PASS: misc
PASS: fixes
PASS: xfree86
========================================================================
1 of 9 tests failed
Please report to https://bugs.freedesktop.org/enter_bug.cgi?product=xorg
========================================================================
make[3]: *** [check-TESTS] Error 1
make[3]: Leaving directory `/tmp/nix-build-xorg-server-1.12.2.drv-0/xorg-server-1.12.2/test'
make[2]: *** [check-am] Error 2
make[2]: Leaving directory `/tmp/nix-build-xorg-server-1.12.2.drv-0/xorg-server-1.12.2/test'
make[1]: *** [check-recursive] Error 1
make[1]: Leaving directory `/tmp/nix-build-xorg-server-1.12.2.drv-0/xorg-server-1.12.2/test'
make: *** [check-recursive] Error 1
phase `check' failed after 3 seconds
builder for `/nix/store/pjla94lqhahnrb0z2h05sb1zh8dlb9wp-xorg-server-1.12.2.drv' failed with exit code 1
@ build-failed /nix/store/pjla94lqhahnrb0z2h05sb1zh8dlb9wp-xorg-server-1.12.2.drv - 1 builder for `/nix/store/pjla94lqhahnrb0z2h05sb1zh8dlb9wp-xorg-server-1.12.2.drv' failed with exit code 1
guix package: error: build failed: build of `/nix/store/pjla94lqhahnrb0z2h05sb1zh8dlb9wp-xorg-server-1.12.2.drv' failed
--8<---------------cut here---------------end--------------->8---

Message #11 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: ludo <at> gnu.org (Ludovic Courtès)
To: Mark H Weaver <mhw <at> netris.org>
Cc: 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Mon, 16 Sep 2013 15:27:23 +0200

Mark H Weaver <mhw <at> netris.org> skribis:

> xorg-server fails to build from source.  See below for the tail of the
> build log.  This is with aae4ead8142d4fd7c674a1e6e302f40469f878c6.

[...]

> PASS: input
> XKB: Failed to compile keymap
> Keyboard initialization failed. This could be a missing or incorrect setup of xkeyboard-config.
>
> Fatal server error:
> Failed to activate core devices.
>
> Please consult the The X.Org Foundation support
>          at http://wiki.x.org
>  for help.
>
> FAIL: xtest

This was discussed some time ago:

  http://lists.gnu.org/archive/html/bug-guix/2013-03/msg00140.html

It does seem like the xkeyboard-config data is not visible in its
default search path.

Apparently the X server can be started with
“-xkbdir /path/to/xkeyboard-config/etc/X11/xkb”, which should solve the
problem.

There were also discussions about introducing an ‘XKBPATH’ env. var.,
but I’m not sure if it was accepted upstream:
<http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=324768>.

Thanks,
Ludo’.

Message #14 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: Andreas Enge <andreas <at> enge.fr>
To: Ludovic Courtès <ludo <at> gnu.org>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Sat, 21 Sep 2013 19:23:52 +0200

I tried the following patch, setting some xkb-related options for configuring
xorg-server:

diff --git a/gnu/packages/xorg.scm b/gnu/packages/xorg.scm
index 9a0e3e2..096ad54 100644
--- a/gnu/packages/xorg.scm
+++ b/gnu/packages/xorg.scm
@@ -4279,10 +4279,23 @@ emulation to complete hardware acceleration for modern GPUs.")
         ("xf86dgaproto" ,xf86dgaproto)
         ("xf86driproto" ,xf86driproto)
         ("xf86vidmodeproto" ,xf86vidmodeproto)
-;;        ("xkbutils" ,xkbutils)
-;;        ("xkeyboard-config" ,xkeyboard-config)

+("setxkbmap" ,setxkbmap)
+("xkbcomp" ,xkbcomp)
+("xkbevd" ,xkbevd)
+("xkbutils" ,xkbutils)
+("xkeyboard-config" ,xkeyboard-config)
         ("xtrans" ,xtrans)
         ("zlib" ,zlib)))
+    (arguments
+     `(#:configure-flags
+       (list (string-append "--with-xkb-path="
+                            (assoc-ref %build-inputs "xkeyboard-config")
+                            "/share/X11/xkb")
+             (string-append "--with-xkb-output="
+                            "/tmp")
+             (string-append "--with-xkb-bin-directory="
+                            (assoc-ref %build-inputs "xkbcomp")
+                            "/bin"))))
     (home-page "http://www.x.org/wiki/")
     (synopsis "xorg implementation of the X Window System")
     (description "X.org provides an implementation of the X Window System")

The same error occurs.

Inspired by nixpkgs, I also tried the following patch to xkeyboard-config:

+    (arguments
+     `(#:configure-flags
+       (list (string-append "--with-xkb-base="
+                            (assoc-ref %outputs "out")
+                            "/etc/X11/xkb")
+             "--with-xkb-rules-symlink=xorg,xfree86")))

Always the same error.

Should nix not have the same problem?

Andreas

Message #17 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: ludo <at> gnu.org (Ludovic Courtès)
To: Andreas Enge <andreas <at> enge.fr>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Sat, 21 Sep 2013 22:02:03 +0200

Andreas Enge <andreas <at> enge.fr> skribis:

> Should nix not have the same problem?

They don’t run the tests.  :-)

However, the real server is passed:

  "-xkbdir" "${pkgs.xkeyboard_config}/etc/X11/xkb"

(See <https://github.com/NixOS/nixos/blob/master/modules/services/x11/xserver.nix>.)

But here, for tests/xtest.c there’s no server running apparently.

> +       (list (string-append "--with-xkb-path="
> +                            (assoc-ref %build-inputs "xkeyboard-config")
> +                            "/share/X11/xkb")

Looking at configure.ac, I think you have to do just that.  Then you can
check the value of XKB_BASE_DIRECTORY in config.log, to make sure
everything’s right.

Can you then run test/xtest in strace and check where it’s looking for
XKB files?

TIA,
Ludo’.

Message #20 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: Andreas Enge <andreas <at> enge.fr>
To: Ludovic Courtès <ludo <at> gnu.org>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Sun, 22 Sep 2013 11:01:12 +0200

[Message part 1 (text/plain, inline)]

On Sat, Sep 21, 2013 at 10:02:03PM +0200, Ludovic Courtès wrote:
> Looking at configure.ac, I think you have to do just that.  Then you can
> check the value of XKB_BASE_DIRECTORY in config.log, to make sure
> everything’s right.

I checked this already before, the attached header file looks good
(maybe not XKM_OUTPUT_DIR in the long run, but I thought that for the
tests that should be fine). Wait, could this be a typo in the build system?
One would rather expect XKB_OUTPUT_DIR with a B instead of an M.

Running xtest outside of the chroot of the guix build system, there is an
error message that xkbcomp does not find the file evdev, which is part of
the xkeyboard-config package. So I added a configure flag
       (list (string-append "--with-xkb-config-root="
                            (assoc-ref %build-inputs "xkeyboard-config")
                            "/share/X11/xkb"))))
to xkbcomp. Now xkeyboard-config needs xkbcomp, so I added an additional
package xkbcomp-tmp to build xkeyboard-config, then the final xkbcomp can
use xkeyboard-config as an input. See the attached patch.

From guix, the tests fail with the same error message as before.

When I become root outside the chroot, source the environment variables and
run ./xtest, the tests succeed. The strace output ends like this:
open("/nix/store/hylfnc1nxg7b8rgnz5iyh0xpszkikaf0-xkeyboard-config-2.6/share/X11/xkb/rules/evdev", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0444, st_size=39508, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7feab8a26000
read(3, "// DO NOT EDIT THIS FILE - IT WA"..., 4096) = 4096
read(3, "\n  $azerty       =       +aliase"..., 4096) = 4096
read(3, "pl_altquotes)\t\t\t=\tpc+pl(dvorak_a"..., 4096) = 4096
read(3, "v[1])\n  *\t\tkan\t\t\t=\tpc+in(kan)\n  "..., 4096) = 4096
read(3, "\tro\t\tacademic\t\t=\tpc+ro(std)\n  *\t"..., 4096) = 4096
read(3, "_probhat):3\n  *\t\tdev(basic)\t=\t+i"..., 4096) = 4096
brk(0x1f3e000)                          = 0x1f3e000
read(3, "4\n  *\t\tyu(unicode)\t=\t+srp(latinu"..., 4096) = 4096
read(3, "s):4\n  *\t\tdzdwi\t\tbasic\t=\t+bt(bas"..., 4096) = 4096
read(3, "ft_caps_toggle)\n  grp:shift_caps"..., 4096) = 4096
read(3, "pdl:momayyezoss      =\t+kpdl(mom"..., 4096) = 2644
read(3, "", 4096)                       = 0
close(3)                                = 0
munmap(0x7feab8a26000, 4096)            = 0
brk(0x1f0d000)                          = 0x1f0d000
brk(0x1f0c000)                          = 0x1f0c000
access("/tmp/", W_OK|X_OK)              = 0
pipe([3, 4])                            = 0
rt_sigaction(SIGALRM, {SIG_IGN, [], SA_RESTORER, 0x7feab7e1c350}, {SIG_DFL, [], 0}, 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x7feab8a27b50) = 24792
rt_sigprocmask(SIG_BLOCK, [ALRM CHLD TSTP TTIN TTOU VTALRM WINCH IO], [], 8) = 0
fcntl(4, F_GETFL)                       = 0x1 (flags O_WRONLY)
fstat(4, {st_mode=S_IFIFO|0600, st_size=0, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7feab8a26000
lseek(4, 0, SEEK_CUR)                   = -1 ESPIPE (Illegal seek)
close(3)                                = 0
write(4, "xkb_keymap \"default\" {\n    xkb_k"..., 314) = 314
close(4)                                = 0
munmap(0x7feab8a26000, 4096)            = 0
wait4(24792, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0, NULL) = 24792
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
--- SIGCHLD (Child exited) @ 0 (0) ---
access("/tmp/", W_OK|X_OK)              = 0
open("/tmp/server-(null).xkm", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=11248, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7feab8a26000
read(3, "\17mkx\26\10\377\7\177\0\0\0\6\0\1\0\214\0D\0\4\0\1\0\f\5\320\0\0\0\1\0"..., 4096) = 4096
lseek(3, 0, SEEK_SET)                   = 0
read(3, "\17mkx\26\10\377\7\177\0\0\0\6\0\1\0\214\0D\0\4\0\1\0\f\5\320\0\0\0\1\0"..., 4096) = 4096
lseek(3, 4096, SEEK_SET)                = 4096
lseek(3, 4096, SEEK_SET)                = 4096
lseek(3, 4096, SEEK_SET)                = 4096
read(3, "\n\376\0\0\377\201\t\0\6\0\377\0\0\0\0\0\f\376\0\0\377\1\377\0\6\4\0\0\0\0\0\0"..., 4096) = 4096
lseek(3, 8192, SEEK_SET)                = 8192
read(3, "\1\1\0\0V\377\10\20\0\0\0\0\0\0\0\0\1\1\0\0x\377\10\20\1\1\0\0y\377\10\20"..., 4096) = 3056
lseek(3, 11248, SEEK_SET)               = 11248
lseek(3, 11248, SEEK_SET)               = 11248
close(3)                                = 0
munmap(0x7feab8a26000, 4096)            = 0
unlink("/tmp/server-(null).xkm")        = 0
exit_group(0)                           = ?

So it looks like XKM is the right prefix in the header file, since a file
named /tmp/server-(null).xkm is used.

Andreas

[xkb-config.h (text/x-chdr, attachment)]

[xorg.patch (text/x-diff, attachment)]

Message #23 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: Andreas Enge <andreas <at> enge.fr>
To: Ludovic Courtès <ludo <at> gnu.org>
Cc: 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Sun, 22 Sep 2013 11:09:36 +0200

On Sun, Sep 22, 2013 at 11:01:12AM +0200, Andreas Enge wrote:
> When I become root outside the chroot, source the environment variables and
> run ./xtest, the tests succeed.

Even more: Without being root, sourcing the environment variables and running
"make check" inside the build directory also runs all tests successfully.

Andreas

Message #26 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: ludo <at> gnu.org (Ludovic Courtès)
To: Andreas Enge <andreas <at> enge.fr>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Sun, 22 Sep 2013 22:35:54 +0200

Andreas Enge <andreas <at> enge.fr> skribis:

> Running xtest outside of the chroot of the guix build system, there is an
> error message that xkbcomp does not find the file evdev, which is part of
> the xkeyboard-config package. So I added a configure flag
>        (list (string-append "--with-xkb-config-root="
>                             (assoc-ref %build-inputs "xkeyboard-config")
>                             "/share/X11/xkb"))))
> to xkbcomp. Now xkeyboard-config needs xkbcomp, so I added an additional
> package xkbcomp-tmp to build xkeyboard-config, then the final xkbcomp can
> use xkeyboard-config as an input. See the attached patch.

Wow.

> From guix, the tests fail with the same error message as before.
>
> When I become root outside the chroot, source the environment variables and
> run ./xtest, the tests succeed. The strace output ends like this:

Could it be that xtest does something not possible in the chroot?  Like
host name lookups, or similar (check the strace log for accesses to
/etc/hosts or connections to /var/run/nscd/socket.)

HTH,
Ludo’.

Message #29 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: Andreas Enge <andreas <at> enge.fr>
To: Ludovic Courtès <ludo <at> gnu.org>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Mon, 23 Sep 2013 19:22:56 +0200

[Message part 1 (text/plain, inline)]

On Sun, Sep 22, 2013 at 10:35:54PM +0200, Ludovic Courtès wrote:
> Could it be that xtest does something not possible in the chroot?  Like
> host name lookups, or similar (check the strace log for accesses to
> /etc/hosts or connections to /var/run/nscd/socket.)

Not that I could tell; I am attaching the strace output in case I overlooked
something. The only access I see outside of the nix store is in /tmp (consis-
tently with my configure flag that writes the compiled keymaps there), and
from what I grepped in the nix builder code, this is in the chroot.

Andreas

[trace (text/plain, attachment)]

Message #32 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: ludo <at> gnu.org (Ludovic Courtès)
To: Andreas Enge <andreas <at> enge.fr>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Mon, 23 Sep 2013 21:39:12 +0200

Andreas Enge <andreas <at> enge.fr> skribis:

> clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x7f9c31ee3b50) = 4633

Could you try ‘strace -f’ so we see what goes on in the child?

Thanks,
Ludo’.

Message #35 received at 15392 <at> debbugs.gnu.org (full text, mbox):

From: Andreas Enge <andreas <at> enge.fr>
To: Ludovic Courtès <ludo <at> gnu.org>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392 <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Wed, 25 Sep 2013 09:00:45 +0200

On Mon, Sep 23, 2013 at 09:39:12PM +0200, Ludovic Courtès wrote:
> Could you try ‘strace -f’ so we see what goes on in the child?

Thanks to joint work with Ludovic, the problem should be solved. Currently, 
hydra.gnu.org does not pick up the git repository, but at least, locally
xorg-server compiles and passes all its tests.

There is one question I still have concerning a path set by a configure flag,
which I started out by setting arbitrarily:
   (string-append "--with-xkb-output="
                  "/tmp") ; FIXME: This is a bit doubtful; where should
                          ; the compiled keyboard maps go?

Do you have any knowledge or opinion on what we should do here?
I suppose that we need a path to which the xserver can write at runtime,
so it should be outside the store.

Andreas

Message #40 received at 15392-done <at> debbugs.gnu.org (full text, mbox):

From: ludo <at> gnu.org (Ludovic Courtès)
To: Andreas Enge <andreas <at> enge.fr>
Cc: Mark H Weaver <mhw <at> netris.org>, 15392-done <at> debbugs.gnu.org
Subject: Re: bug#15392: xorg-server FTBFS
Date: Wed, 25 Sep 2013 15:11:44 +0200

Andreas Enge <andreas <at> enge.fr> skribis:

> On Mon, Sep 23, 2013 at 09:39:12PM +0200, Ludovic Courtès wrote:
>> Could you try ‘strace -f’ so we see what goes on in the child?
>
> Thanks to joint work with Ludovic, the problem should be solved. Currently, 
> hydra.gnu.org does not pick up the git repository, but at least, locally
> xorg-server compiles and passes all its tests.

For the record, it was another /bin/sh issue, fixed by 2365338.

Closing the bug now.  :-)

> There is one question I still have concerning a path set by a configure flag,
> which I started out by setting arbitrarily:
>    (string-append "--with-xkb-output="
>                   "/tmp") ; FIXME: This is a bit doubtful; where should
>                           ; the compiled keyboard maps go?
>
> Do you have any knowledge or opinion on what we should do here?
> I suppose that we need a path to which the xserver can write at runtime,
> so it should be outside the store.

Looking at configure.ac, this is used to #define XKM_OUTPUT_DIR, which
ddxLoad.c describes like this:

--8<---------------cut here---------------start------------->8---
        /*
         * If XKM_OUTPUT_DIR specifies a path without a leading slash, it is
         * relative to the top-level XKB configuration directory.
         * Making the server write to a subdirectory of that directory
         * requires some work in the general case (install procedure
         * has to create links to /var or somesuch on many machines),
         * so we just compile into /usr/tmp for now.
         */
#ifndef XKM_OUTPUT_DIR
#define	XKM_OUTPUT_DIR	"compiled/"
#endif
--8<---------------cut here---------------end--------------->8---

I interpret it as /tmp being a good value.

At the same time, I wonder if the X server normally calls xkbcomp at
run time.

Thanks,
Ludo’.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.