This reverts commit a768871934.
This is too fragile, it breaks at least on:
* ssl dh params
* hostnames in proxypass and upstreams are resolved in the sandbox
On some systems, EFI variables are not supported or otherwise wonky.
bootctl attempting to access them causes failures during bootloader
installations and updates. For such systems, NixOS provides the options
`boot.loader.efi.canTouchEfiVariables` and
`boot.loader.systemd-boot.graceful` which pass flags to bootctl that
change whether and how EFI variables are accessed.
Previously, these flags were only passed to bootctl during an install
operation. However, they also apply during an update operation, which
can cause the same sorts of errors. This change passes the flags during
update operations as well to prevent those errors.
Fixes https://github.com/NixOS/nixpkgs/issues/151336
The update test patches the systemd-boot binary to report a known
version then tests that this is the version updated from. The previous
patch would also search the kernel and initrd binaries, which would
cause sed to write out a temporary file that might cause the disk
to run out of space and the test to fail.
Only attempt to patch binaries which contain systemd-boot (usually
`BOOT<arch>.EFI` and `systemd-boot<arch>.efi` to avoid this problem.
As a bonus, this reduces test time by 20-30%.
The aarch64-linux kernel and initrd recently eclipsed 60M, causing the
boot disk image build to run out of space and fail. Double the size of
the image to 120M to fix the issue.
The disk image is stored in expandable qcow2 format, so only the space
actually used by files in the image is consumed. Therefore, other
architectures are not unfairly penalized, and the output size does not
suddenly double.
This also fixes NixOS tests which use this option, like systemd-boot's.
At some point many months ago, the systemd-boot update script stopped
outputting parentheses around the version being upgraded from, causing
the test to fail. Remove the parentheses from the expected message to
fix the test.
This moves the creation of the bind mount inside the `nixos-enter`
invocation. The command are executed in an unshared mount namespace, so
they can be run as an unprivileged user.
Although we don't really need HTML documentation in the minimal installer,
not including it may cause annoying cache misses in the case of the NixOS manual.
When installing NixOS in the target filesystem /mnt, paths relative to
configuration.nix in `initrd.secrets` are turned by Nix into absolute
paths that reference /mnt. While building the system derivation works,
installing the bootloader fails because the latter process takes place
inside the chroot environment where /mnt does not exist.
Ideally, we would also build the system within chroot, but this greatly
complicates the matter as it requires manually copying over Nix, its
runtime dependencies and all channels. Possibly, this would also break
several assumptions users have about how nixos-install works.
A simpler and safer (but less neat) solution is to temporarily bind
mount all mount points in /mnt under /mnt/mnt to keep the paths
functional while the bootloader is being installed.
This is essentially the workaround described in issue #73404.
The build of initrd-secrets can routinely fail for old boot entries
if the secrets have been removed or renamed in a later generation.
This always happens for generation 1, because it's built from the
NixOS installer and the paths differs by the mount point (i.e. /mnt).
The error is very confusing because it fails to mention it's about
an older generation and that it's somewhat harmless.
This commit turns the error into a warning for all generations but the
current, adds the name of the failed entry to the message and a note
explaining why it can happen.
This commit fixes a papercut in nixos-rebuild where people wanting to
switch to a specialisation (or test one) were forced to manually figure
out the specialisation's path and run its activation script - since now,
there's a dedicated option to do just that.
This is a backwards-compatible change which doesn't affect the existing
behavior, which - to be fair - might still be considered sus by some
people, the painful scenario here being:
- you boot into specialisation `foo`,
- you run `nixos-rebuild switch`,
- whoops, you're no longer at specialisation `foo`, but you're rather
brought back to the base system.
(it's especially painful for cases where specialisation is used to load
extra drivers, e.g. Nvidia, since then launching `nixos-rebuild switch`,
while forgetting that you're inside a specialisation, can cause some
parts of your system to get accidentally unloaded.)
I've tried to mitigate that by improving specialisations so that they
create a dedicated file somewhere in `/run/current-system` containing
the specialisation's name (which `nixos-rebuild` could then use as the
default value for `--specialisation`), but I haven't been able to come
up with anything working (plus it would be a breaking change then).
Closes https://github.com/NixOS/nixpkgs/issues/174065
* Will make it so that GHC.Paths's docdir NIX_GHC_DOCDIR points to an
actual directory.
* Documentation of all packages in the environment is available in
`$out/share/doc`.
This has previously been attempted in #76842 and reverted in #77442,
since documentation can collide when the libraries wouldn't (thanks to
the hash in the lib filename). `symlinkJoin` allows collision, so this
solution should be akin to #77523 (minus `buildEnv`, one step at a
time). `installDocumentation = false` restores the old behavior.
Collision in the documentation only happen if the dependency closure of
the given packages has more than one different derivation for the same
library of the very same version. I'm personally inclined not to claim
that our infrastructure does anything sensible in this case.
Additionally, the documentation is likely largely the same in such
cases (unless it is heavily patched).
Resolves#150666.
Resolves#76837.
Closes#150968.
Closes#77523.
If `runtime_dir` is not set, gitaly will use `/tmp` as a place for sockets
and packed binaries like `git2go`. If the gitlab instance does not experience
much traffic and gitlay is not restarted regularly, systemd-tmpfiles will
remove the binaries. This breaks some gitlab functionality until gitaly is
restarted manually.
This is a followup of #148921, to allow local builds when
`--target-host` is used again. It also documents the change in
behavior, regarding the specialty of the `localhost` value.
By removing the special handling of an empty `buildHost` and non empty
`targetHost`, this change also slightly alters the behavior of
`nixos-rebuild`.
Originally by specifying `--target-host target --build-host ""`, the
now removed special case would transform those arguments to
`--target-host target --build-host target`.
Now the empty `--build-host` would result in a local build.
This provides an easy way to specify exclude patterns in config. It was
already possible via extraBackupOptions; this change creates a simpler,
similar to other backup services, way to specify them.
This commit also moves the indicator files out of the directory that's
being backed up, so that the directory remains static throughout the
backup operation.
The agent has not been updated for a very long time. In addition to
updating to the newest tagged version the change creates a package for
it.
The existing version has issues with the new python2.7 package not
containing crypt.so file. And the commit
6910a4eea0 I believe introduced
regression that caused the shebang to not be updated.
apparently pandoc has changed behavior over the past releases, so the
files are no longer in sync. occasionally this requires edits
to the markdown source to not remove an anchor that was there
before (albeit wth a very questionable id), or where things were simply
being misrendered due to syntax errors.
we only have three uses at the moment, all of them in code blocks where
they could just as well (or maybe better) be comments. markdown can't do
callouts without another pandoc filter, so we'll turn them into comments
instead.
synapse would've benefited from inline links, but referencing an
external numbered list as plain text (instead of clickable links, like
callout lists had) seems even worse than putting urls into comments as
plain text.
markdown doesn't really have examples as a first-class construct. we'll
keep all examples that are referenced around for now, but all
unreferenced examples turn into invisible anchors. (turning them into
fourth-level headings in their files, as would be necessary for emacs,
removes them from the TOC anyway.)
productname, application, acronym, guilabel, and guibutton were so far
not rendered specially and can go away completely.
replaceable does render differently, but since it was only used twice
and in places where the intent should be clear without the extra markup
it can go as well.
makes sure that program listing tags are separated from their contents
by exactly a newline character. this makes the markdown translation
easier to verify (since no new newlines need to be inserted), and
there's no rendering difference anyway.
MD can only do the latter, so change them all over now to keeps diffs reviewable.
this also includes <literal><xref> -> <xref> where options are referenced since
the reference will implicitly add an inner literal tag.
markdown cannot represent those links. remove them all now instead of in
each chapter conversion to keep the diff for each chapter small and more
understandable.
The `serviceOverrides` module option is commonly used to loosen the
systemd unit's hardening. This commit merges the `serviceConfig` with
`mkMerge` instead of using the update operator `//` which discards all
existing values on conflict. To avoid a breaking change which requires
defining each option with a higher priority (e.g., through `mkForce`),
this commit prefixes hardening values with `mkDefault`.
Notable exceptions are list hardening options which use `mkBefore`
instead of `mkDefault`. This allows for easy extension of the existing
settings. Resetting redefinitions are still possible through `mkForce`.
When this module was first introduced, it processed the runtime option
in a way that nested the resulting files and directories under an etc
directory.
https://github.com/NixOS/nixpkgs/pull/98506/files#diff-685092dbb1852fbf30857fe3505d25dc471dc79d0f31c466523b5f5822b68127R11-R21
That implementation relied on nixos/modules/system/etc/make-etc.sh, a
script that was later removed.
eb7120dc79
The implementation was updated to use linkFarm, which changed the
behavior slightly, in that the configured files and directories are no
longer automatically nested under an etc directory.
307b1253a7
But the module still configures neovim's runtimepath in a way that
assumes the old nesting behavior.
04f574a1c0/nixos/modules/programs/neovim.nix (L173)
Restore the original behavior, nesting runtime files and directories
under an etc directory.
