Since systemd version 232 the install subcommand of bootctl opens the
loader.conf with fopen() modes "wxe", where the "e" stands for
exclusive, so the call will fail if the file exists.
For installing the boot loader just once this is fine, but if we're
using NIXOS_INSTALL_BOOTLOADER on a systemd where the bootloader is
already present this will fail.
Exactly this is done within the simpleUefiGummiboot installer test,
where nixos-install is called twice and thus the bootloader is also
installed twice, resulting in an error during the fopen call:
Failed to open loader.conf for writing: File exists
Removing the file prior to calling bootctl should fix this.
I've tested this using the installer.simpleUefiGummiboot test and it now
succeeds.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @edolstra, @shlevy, @mic92
Fixes: #22925
This leads to the following error when trying to install a new machine
where the machine ID wasn't yet initialized during boot:
Failed to get machine did: No such file or directory
In addition this was also detected by the simpleUefiGummiboot installer
test.
So let's generate a fallback machine ID by using
systemd-machine-id-setup before actually running bootctl.
Tested this by running the installer.simpleUefiGummiboot test, it still
fails but not because of the machine ID.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @edolstra, @shlevy, @mic92
Fixes: #22561
So this was suggested as [long ago as October, 2015](https://github.com/NixOS/nixpkgs/issues/10376#issuecomment-147734898).
Despite being fairly ignorant of the nix Python support, I decided
that I really wanted this; this change brings in what I believe are
the necessary components---I have, at least, successfully run `beet
replaygain` and `beet bpd`---but it may not do it in the best way; I'm
happy to consider input on that front.
I can at least state that all three changes are necessary---leave any
one of them out and gstreamer support doesn't work.
First of all, these "documents" are not really documentation, so it
really doesn't make much sense to put it into $doc.
The main point however is that the installer tests are failing since
this was introduced in ac0cdc1952.
One way to circumvent this is putting dbus.doc into
system.extraDependencies of the installer tests, but given the first
point this sounds a bit odd to me.
So I went for the second way of putting it into $out, because it's now
basically necessary to build a NixOS system.
With this the NixOS installer tests should now work again, although I
have only tested this with the installer.simple test.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @abbradar
The problem with using libxslt as buildInputs is that the dev output is
used for building the dbus config.
This is one of the reasons why the installer tests are failing since
ac0cdc1952, because the tests do not have
libxslt.dev in their closure and really shouldn't.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @abbradar
This is the output of the builder:
building path(s) `/nix/store/khkcfb8433i9mabb6wnb8ik6p9skg644-dbus-1'
error : connection refused
error : connection refused
However, even when using --nonet we'd still get this:
I/O error : Attempt to load network entity
http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd
So in order to avoid this, we now provide an XML catalog file, mapping
the public URLs to the local DTD paths inside the store instead of using
--path (which doesn't seem to work with xsltproc).
Tested this by comparing the SHA256 (nix-hash --type sha256) of the
output path generated by:
nix-build -E '(import ./. {}).makeDBusConf {
suidHelper = "SUIDHELPER";
serviceDirectories = [ "SERVICEDIR1" "SERVICEDIR1" ];
}'
... with the SHA256 of the generated output path prior to this commit
and they have the same hash:
6f3f9594b12fddbff9407b85252b6f649da11f56b7fd514f761966c11399a7ab
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @abbradar
