When explicitly requested by the caller, as suggested in the meeting
(https://github.com/NixOS/nix/pull/8090#issuecomment-1531139324)
> @edolstra: { toPath } vs { fromPath } is too implicit
I've opted for the `inputAddressed = true` requirement, because it
we did not agree on renaming the path attributes.
> @roberth: more explicit
> @edolstra: except for the direction; not immediately clear in which direction the rewriting happens
This is in fact the most explicit syntax and a bit redundant, which is
good, because that redundancy lets us deliver an error message that
reminds expression authors that CA provides a better experience to
their users.
* nix flake check: improve error message if overlay is not a lambda
Suppose you have an overlay like this
{
inputs = { /* ... */ };
outputs = { flake-utils, ... }: flake-utils.lib.eachDefaultSystem
(system: {
overlays.default = final: prev: {
};
});
}
then `nix flake check` (correctly) fails because `overlays` are supposed
to have the structure `overlays.<name> = final: prev: exp`. However, the
error-message is a little bit counter-intuitive:
error: overlay does not take an argument named 'final'
While one might guess where the error actually comes from because the
trace above says `… while checking the overlay 'overlays.x86_64-linux'`
this is still pretty confusing because it complains about an argument
not being named `final` even though that's evidently the case.
With this change, the error-message actually makes it clear what's
wrong:
[ma27@carsten:~/Projects/nix/tmp]$ nix flake check --extra-experimental-features 'nix-command flakes' path:$(pwd)
error:
… while checking flake output 'overlays'
at /nix/store/clgblnxx003hyrq8qkz5ab6kgqkck6qc-source/flake.nix:4:5:
3| outputs = { ... }: {
4| overlays.x86_64-linux.snens = final: prev: {
| ^
5| kek = throw "snens";
… while checking the overlay 'overlays.x86_64-linux'
at /nix/store/clgblnxx003hyrq8qkz5ab6kgqkck6qc-source/flake.nix:4:5:
3| outputs = { ... }: {
4| overlays.x86_64-linux.snens = final: prev: {
| ^
5| kek = throw "snens";
error: overlay is not a lambda, but a set instead
- Improved API docs from comment
- Exit codes are for `nix-build`, not just `nix-store --release`
- Make note in tests so the magic numbers are not surprising
Picking up where #8387 left off.
Previously it was not possible to open a local store when its database is on a read-only filesystem. Obviously a store on a read-only filesystem cannot be modified, but it would still be useful to be able to query it.
This change adds a new read-only setting to LocalStore. When set to true, Nix will skip operations that fail when the database is on a read-only filesystem (acquiring big-lock, schema migration, etc), and the store database will be opened in immutable mode.
Co-authored-by: Ben Radford <benradf@users.noreply.github.com>
Co-authored-by: cidkidnix <cidkidnix@protonmail.com>
Co-authored-by: Dylan Green <67574902+cidkidnix@users.noreply.github.com>
Co-authored-by: John Ericson <git@JohnEricson.me>
Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
If the garbage collector has acquired the global GC lock, but hasn't
created the GC socket yet, then a client attempting to connect would
get ENOENT. Note that this only happens when the GC runs for the first
time on a machine. Subsequently clients will get ECONNREFUSED which
was already handled.
Fixes#7370.
A library shouldn't require changes to the caller's argument handling,
especially if it doesn't have to, and indeed we don't have to.
This changes the lookup order to prioritize the hardcoded path to nix
if it exists. The static executable still finds itself through /proc
and the like.
The remaining constructor RegisterPrimOp::RegisterPrimOp(Info && info)
allows specifying the documentation in .args and .doc members of the
Info structure.
Commit 8ec1ba0210 removed all uses of the removed constructor in the
nix binary. Here, we remove the constructor completely as well as its
use in a plugin test. According to #8515, we didn't promis to maintain
compatibility with external plugins.
Fixes#8515
Previously, for tarball flakes, we recorded the original URL of the
tarball flake, rather than the URL to which it ultimately
redirects. Thus, a flake URL like
http://example.org/patchelf-latest.tar that redirects to
http://example.org/patchelf-<revision>.tar was not really usable. We
couldn't record the redirected URL, because sites like GitHub redirect
to CDN URLs that we can't rely on to be stable.
So now we use the redirected URL only if the server returns the
`x-nix-is-immutable` or `x-amz-meta-nix-is-immutable` headers in its
response.
Currently `fromTOML` throws an exception when encountering a timestamp
since the Nix language lacks a way to represent them.
This patch changes this beaviour and makes `fromTOML` parse timestamps as
attrsets of the format
{ _type = "timestamp"; value = "1979-05-27T07:32:00Z"; }
This is guarded by an experimental feature flag to leave room for iterating on the representation.
The primop `builtins.replaceStrings` currently always strictly evaluates the
replacement strings, however time and space are wasted for their computation
if the corresponding pattern do not occur in the input string. This commit
makes the evaluation of the replacement strings lazy by deferring their
evaluation to when the corresponding pattern are matched and memoize the result
for efficient retrieval on subsequent matches.
The testcases for replaceStrings was updated to check for lazy evaluation
of the replacements. A note was also added in the release notes to
document the behavior change.
- If the element comes from a flake, print the full flakeref (with the
fragment part) and not just the reference to the flake itself
- If the element doesn't come from a flake, print its store path(s)
This is a bit too verbose, but has the advantages of being correct (and
not crashing), so it's strictly better than the previous situation
Fix https://github.com/NixOS/nix/issues/8284
As discussed in #7417, it would be good to make more string values work
as installables. That is to say, if an installable refers to a value,
and the value is a string, it used to not work at all, since #7484, it
works somewhat, and this PR make it work some more.
The new cases that are added for `BuiltPath` contexts:
- Fixed input- or content-addressed derivation:
```
nix-repl> hello.out.outPath
"/nix/store/jppfl2bp1zhx8sgs2mgifmsx6dv16mv2-hello-2.12"
nix-repl> :p builtins.getContext hello.out.outPath
{ "/nix/store/c7jrxqjhdda93lhbkanqfs07x2bzazbm-hello-2.12.drv" = { outputs = [ "out" ]; }; }
The string matches the specified single output of that derivation, so
it should also be valid.
- Floating content-addressed derivation:
```
nix-repl> (hello.overrideAttrs (_: { __contentAddressed = true; })).out.outPath
"/1a08j26xqc0zm8agps8anxpjji410yvsx4pcgyn4bfan1ddkx2g0"
nix-repl> :p builtins.getContext (hello.overrideAttrs (_: { __contentAddressed = true; })).out.outPath
{ "/nix/store/qc645pyf9wl37c6qvqzaqkwsm1gp48al-hello-2.12.drv" = { outputs = [ "out" ]; }; }
```
The string is not a path but a placeholder, however it also matches
the context, and because it is a CA derivation we have no better
option. This should also be valid.
We may also want to think about richer attrset based values (also
discussed in that issue and #6507), but this change "completes" our
string-based building blocks, from which the others can be desugared
into or at least described/document/taught in terms of.
Progress towards #7417
Co-authored-by: Robert Hensing <roberth@users.noreply.github.com>
Fixes#8309
This regression was because both `CmdDevelop` and `CmdPrintDevEnv` were
switched to be `InstallableValueCommand` subclasses, but actually
neither should have been.
The `nixpkgsFlakeRef` method should indeed not be on the base
installable class, because "flake refs" and "nixpkgs" are not
installable-wide notions, but that doesn't mean these commands should
only accept installable values.
This fixes a bug in commands like `nix eval' which would emit invalid attribute
sets if they contained reserved keywords such as "assert", "let", etc.
These keywords will not be quoted when printed, making them valid expressions.
All keywords recognized by the lexer are quoted except "or", which does not
require quotation.
fe5509df caused only wanted outputs to be passed to the
post-build-hook, which resulted in paths being built
without ever going into the hook.
This commit adds a (currently failing) test for this.
Previously, we relied on the `shutdown()` function to terminate `accept()`
calls on a listening socket. However, this approach did not work on macOS as
the waiting `accept()` call is not considered a connected socket, resulting in
an `ENOTCONN` error. Instead, we now close the listening socket to terminate
the `accept()` call.
Additionally, we fixed a resource management issue where we set the
`daemonSocket` variable to -1, triggering resource cleanup and causing the
`stopDaemon` function to be called twice. This resulted in errors as the socket
was already closed by the time the second `stopDaemon` call was made. Instead of
setting `daemonSocket` to -1, we now release the socket using the `release()`
method on a unique pointer. This properly transfers ownership and allows for
correct resource cleanup.
These changes ensure proper behavior and resource management for the
recursive-nix feature on macOS.
We finally test the status quo of remote build trust in a number of
ways. We create a new experimental feature on `nix-daemon` to do so.
PR #3921, which improves the situation with trustless remote building,
will build upon these changes. This code / tests was pull out of there
to make this, so everything is easier to review, and in particular we
test before and after so the new behavior in that PR is readily apparent
from the testsuite diff alone.
Issues:
1. Features gated on disabled experimental settings should warn and be
ignored, not silently succeed.
2. Experimental settings in the same config "batch" (file or env var)
as the enabling of the experimental feature should work.
3. For (2), the order should not matter.
These are analogous to the issues @roberth caught with my changes for
arg handling, but they are instead for config handling.
Co-authored-by: Robert Hensing <roberth@users.noreply.github.com>
This is somewhat hacky fix just for 2.15. I unintentionally hid them
from the manual, when no one wanted to hide them that (including
myself). I also required the experimental feature to be enabled in an
order-dependent way, which is not good.
The simplest fix for this immanent release is just to always show them,
and always allow them to be set.
Effectively undoes some changes from aa663b7e89
This function returns true or false depending on whether the Nix client
is trusted or not. Mostly relevant when speaking to a remote store with
a daemon.
We include this information in `nix ping store` and `nix doctor`
Co-Authored-By: John Ericson <John.Ericson@Obsidian.Systems>
This introduces the SourcePath type from lazy-trees as an abstraction
for accessing files from inputs that may not be materialized in the
real filesystem (e.g. Git repositories). Currently, however, it's just
a wrapper around CanonPath, so it shouldn't change any behaviour. (On
lazy-trees, SourcePath is a <InputAccessor, CanonPath> tuple.)
Prior to this, there was an ad-hoc whitelist in `main.cc`. Now, every
command states its stability.
In a future PR, we will adjust the manual to take advantage of this new
information in the JSON.
(It will be easier to do that once we have some experimental feature
docs to link too; see #5930 and #7798.)
Was probably an overlook of when the tests were first added, but that
now messes-up with the `nix-channel --update` that happens down the line
(and can't access the network since we're inside a Nix build)
See the note in the test.
We don't want these flags showing up for commands where they are
irrelevant.
Eventually, this needs a proper fix, but it need not be a blocker for
stabilize: for a quick-n-dirty punt, just put these flags behind the
`nix-command` unstable feature.
This is fine because they are only relevant for commands which we don't
need to stabilize for a while.
`legacyPackages` of nixpkgs trigger eval errors in `hasContent`, causing
the whole `legacyPackages` being skipped. We should treat it as
has-content in that case.
I noticed a regression in the lazy-trees branch, which I'm trying to
capture with this test. While the tests succeeds in master, the
lazy-trees branch gives the following error message:
error: access to path
'/build/nix-test/tests/flakes/flake-in-submodule/rootRepo/submodule/flake.nix'
is forbidden because it is not under Git control; maybe you should
'git add' it to the repository
'/build/nix-test/tests/flakes/flake-in-submodule/rootRepo'?
- Try not to put cryptic "99" in many places
Factor out `exit 99` into `skipTest` function
- Alows make sure skipping a test is done with a reason
`skipTest` takes a mandatory argument
- Separate pure conditionals vs side-effectful test skipping.
"require daemon" already had this, but "sandbox support" did not.
Add the --base64 and --sri flags for the Base64 and SRI format output.
Add the --base16 flag to explicitly specify the hexadecimal format.
Add the --to-base64 and --to-sri flag to convert a hash to the above
mentioned format.
Use `set -u` and `set -o pipefail` to catch accidental mistakes and
failures more strongly.
- `set -u` catches the use of undefined variables
- `set -o pipefail` catches failures (like `set -e`) earlier in the
pipeline.
This makes the tests a bit more robust. It is nice to read code not
worrying about these spurious success paths (via uncaught) errors
undermining the tests. Indeed, I caught some bugs doing this.
There are a few tests where we run a command that should fail, and then
search its output to make sure the failure message is one that we
expect. Before, since the `grep` was the last command in the pipeline
the exit code of those failing programs was silently ignored. Now with
`set -o pipefail` it won't be, and we have to do something so the
expected failure doesn't accidentally fail the test.
To do that we use `expect` and a new `expectStderr` to check for the
exact failing exit code. See the comments on each for why.
`grep -q` is replaced with `grepQuiet`, see the comments on that
function for why.
`grep -v` when we just want the exit code is replaced with `grepInverse,
see the comments on that function for why.
`grep -q -v` together is, surprise surprise, replaced with
`grepQuietInverse`, which is both combined.
Co-authored-by: Robert Hensing <roberth@users.noreply.github.com>
The release notes document the change in behavior, I don't include it
here so there is no risk to it getting out of sync.
> Motivation
>> Plumbing CLI should be simple
Store derivation installations are intended as "plumbing": very simple
utilities for advanced users and scripts, and not what regular users
interact with. (Similarly, regular Git users will use branch and tag
names not explicit hashes for most things.)
The plumbing CLI should prize simplicity over convenience; that is its
raison d'etre. If the user provides a path, we should treat it the same
way not caring what sort of path it is.
>> Scripting
This is especially important for the scripting use-case. when arbitrary
paths are sent to e.g. `nix copy` and the script author wants consistent
behavior regardless of what those store paths are. Otherwise the script
author needs to be careful to filter out `.drv` ones, and then run `nix
copy` again with those paths and `--derivation`. That is not good!
>> Surprisingly low impact
Only two lines in the tests need changing, showing that the impact of
this is pretty light.
Many command, like `nix log` will continue to work with just the
derivation passed as before. This because we used to:
- Special case the drv path and replace it with it's outputs (what this
gets rid of).
- Turn those output path *back* into the original drv path.
Now we just skip that entire round trip!
> Context
Issue #7261 lays out a broader vision for getting rid of `--derivation`,
and has this as one of its dependencies. But we can do this with or
without that.
`Installable::toDerivations` is changed to handle the case of a
`DerivedPath::Opaque` ending in `.drv`, which is new: it simply doesn't
need to do any extra work in that case. On this basis, commands like
`nix {show-derivation,log} /nix/store/...-foo.drv` still work as before,
as described above.
When testing older daemons, the post-build-hook will be run against the
old CLI, so we need the old version of the post-build-hook to support
that use-case.
Co-authored-by: Travis A. Everett <travis.a.everett@gmail.com>
Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
Split `common.sh` into the vars and functions definitions vs starting
the daemon (and possibly other initialization logic). This way,
`init.sh` can just `source` the former. Trying to start the daemon
before `nix.conf` is written will fail because `nix daemon` requires
`--experimental-features 'nix-command'`.
`killDaemon` is idempotent, so it's safe to call when no daemon is
running.
`startDaemon` and `killDaemon` use the PID (which is now exported to
subshells) to decide whether there is work to be done, rather than
`NIX_REMOTE`, which might conceivably be set differently even if a
daemon is running.
`startDaemon` and `killDaemon` can save/restore the old `NIX_REMOTE` as
`NIX_REMOTE_OLD`.
`init.sh` kills daemon before deleting everything (including the daemon
socket).
`init.sh` is tested on its own. We used to do that. I deleted it in
4720853129 but I am not sure why. Better
to just restore it; at one point working on this every other test
passed, so seems good to check whether `init.sh` can be run twice.
We don't *need* to run `init.sh` twice, but I want to try to make our
tests as robust as possible so that manual debugging (where tests for
better or worse might be run ways that we didn't expect) is less
fragile.
It would be incorrect to say that the `sourceInfo` has an `outPath`
that isn't the root. `sourceInfo` is about the root, whereas only
the flake may not be about the root. Thanks Eelco for pointing that
out.
- `nixpkgsFor` does all of native, static, cross, and the different stdenvs.
- The main Nix derivation is no longer duplicated for static.
- DRY nixpkgs.lib and lib.genAttrs calls.
Adding a test to ensure there is no regression.
The tests that are split out of `tests/build.sh` are ones that don't yet
work with CA derivation. I have not yet evaluated whether they should or
not.
This behavior, reported missing in issue #4661, already got fixed in
PR #4818, but didn't get a test case then.
XDG Base Directory is a standard for locations for storing various
files. Nix has a few files which seem to fit in the standard, but
currently use a custom location directly in the user's ~, polluting
it:
- ~/.nix-profile
- ~/.nix-defexpr
- ~/.nix-channels
This commit adds a config option (use-xdg-base-directories) to follow
the XDG spec and instead use the following locations:
- $XDG_STATE_HOME/nix/profile
- $XDG_STATE_HOME/nix/defexpr
- $XDG_STATE_HOME/nix/channels
If $XDG_STATE_HOME is not set, it is assumed to be ~/.local/state.
Co-authored-by: Théophane Hufschmitt <7226587+thufschmitt@users.noreply.github.com>
Co-authored-by: Tim Fenney <kodekata@gmail.com>
Co-authored-by: pasqui23 <pasqui23@users.noreply.github.com>
Co-authored-by: Artturin <Artturin@artturin.com>
Co-authored-by: John Ericson <Ericson2314@Yahoo.com>
For frameworks it's important that structures are as lazy as possible
to prevent infinite recursions, performance issues and errors that
aren't related to the thing to evaluate. As a consequence, they have
to emit more attributes than strictly (sic) necessary.
However, these attributes with empty values are not useful to the user
so we omit them.
Previously, getDefaultNixPath was called too early: at initialisation
time, before CLI and config have been processed, when `restrictEval` and
`pureEval` both have their default value `false`. Call it when
initialising the EvalState instead, and use `setDefault`.
Add an `$` at the end of the `grep` regex. Without it, `checkRef foo`
would always imply `checkRef foo.drv`. We want to tell these situations
apart to more precisely test what is going on.
Allows checking directory entry type of a single file/directory.
This was added to optimize the use of `builtins.readDir` on some
filesystems and operating systems which cannot detect this information
using POSIX's `readdir`.
Previously `builtins.readDir` would eagerly use system calls to lookup
these filetypes using other interfaces; this change makes these
operations lazy in the attribute values for each file with application
of `builtins.readFileType`.
The original `builtins.getContext` test from
1d757292d0 would have caught this. The
problem is that b30be6b450 adding
`builtins.appendContext` modified that test to make it test too much at
once, rather than adding a separate test.
We now have isolated tests for both functions, and also a property test
showing everything put together (in the form of an eta rule for strings
with context). This is better coverage and properly reproduces the bug.
Rather than using `/nix/var/nix/{profiles,gcroots}/per-user/`, put the user
profiles and gcroots under `$XDG_DATA_DIR/nix/{profiles,gcroots}`.
This means that the daemon no longer needs to manage these paths itself
(they are fully handled client-side). In particular, it doesn’t have to
`chown` them anymore (removing one need for root).
This does change the layout of the gc-roots created by nix-env, and is
likely to break some stuff, so I’m not sure how to properly handle that.
Prior to this change, we had a bunch of ad-hoc string manipulation code
scattered around. This made it hard to figure out what data model for
string contexts is.
Now, we still store string contexts most of the time as encoded strings
--- I was wary of the performance implications of changing that --- but
whenever we parse them we do so only through the
`NixStringContextElem::parse` method, which handles all cases. This
creates a data type that is very similar to `DerivedPath` but:
- Represents the funky `=<drvpath>` case as properly distinct from the
others.
- Only encodes a single output, no wildcards and no set, for the
"built" case.
(I would like to deprecate `=<path>`, after which we are in spitting
distance of `DerivedPath` and could maybe get away with fewer types, but
that is another topic for another day.)
Adds a new boolean structured attribute
`outputChecks.<output>.unsafeDiscardReferences` which disables scanning
an output for runtime references.
__structuredAttrs = true;
outputChecks.out.unsafeDiscardReferences = true;
This is useful when creating filesystem images containing their own embedded Nix
store: they are self-contained blobs of data with no runtime dependencies.
Setting this attribute requires the experimental feature
`discard-references` to be enabled.
Without the change checks issue the fllowing warning:
$ nix flake check
trace: warning: The option `nix.useSandbox' defined in `makeTest parameters' has been renamed to `nix.settings.sandbox'.
trace: warning: The option `nix.useSandbox' defined in `makeTest parameters' has been renamed to `nix.settings.sandbox'.
trace: warning: The option `nix.maxJobs' defined in `makeTest parameters' has been renamed to `nix.settings.max-jobs'.
...
First, logic is consolidated in the shell script instead of being spread
between them and makefiles. That makes understanding what is going on a
little easier.
This would not be super interesting by itself, but it gives us a way to
debug tests more easily. *That* in turn I hope is much more compelling.
See the updated manual for details.
Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
Co-authored-by: Eelco Dolstra <edolstra@gmail.com>
Co-authored-by: Robert Hensing <roberth@users.noreply.github.com>
This makes 'nix build' work on paths (which will be copied to the
store) and store paths (returned as is). E.g. the following flake
output attributes can be built using 'nix build .#foo':
foo = ./src;
foo = self.outPath;
foo = builtins.fetchTarball { ... };
foo = (builtins.fetchTree { .. }).outPath;
foo = builtins.fetchTree { .. } + "/README.md";
foo = builtins.storePath /nix/store/...;
Note that this is potentially risky, e.g.
foo = /.;
will cause Nix to try to copy the entire file system to the store.
What doesn't work yet:
foo = self;
foo = builtins.fetchTree { .. };
because we don't handle attrsets with an outPath attribute in it yet,
and
foo = builtins.storePath /nix/store/.../README.md;
since result symlinks have to point to a store path currently (rather
than a file inside a store path).
Fixes#7417.
This makes the position object used in exceptions abstract, with a
method getSource() to get the source code of the file in which the
error originated. This is needed for lazy trees because source files
don't necessarily exist in the filesystem, and we don't want to make
libutil depend on the InputAccessor type in libfetcher.
Fix#6209
When trying to run `nix log <installable>`, try first to resolve the derivation pointed to
by `<installable>` as it is the resolved one that holds the build log.
This has a couple of shortcomings:
1. It’s expensive as it requires re-reading the derivation
2. It’s brittle because if the derivation doesn’t exist anymore or can’t
be resolved (which is the case if any one of its build inputs is missing),
then we can’t access the log anymore
However, I don’t think we can do better (at least not right now).
The alternatives I see are:
1. Copy the build log for the un-resolved derivation. But that means a
lot of duplication
2. Store the results of the resolving in the db. Which might be the best
long-term solution, but leads to a whole new class of potential
issues.
This is a really old test case (which was originally written before the
proper Nix syntax). The tested deep comparison behavior was implemented
and reverted soon after due to performance problems, but it has been
restored in today's Nix again (thanks to the derivation comparison
optimization, presumably).
* Clarify the documentation of foldl': That the arguments are forced
before application (?) of `op` is necessarily true. What is important
to stress is that we force every application of `op`, even when the
value turns out to be unused.
* Move the example before the comment about strictness to make it less
confusing: It is a general example and doesn't really showcase anything
about foldl' strictness.
* Add test cases which nail down aspects of foldl' strictness:
* The initial accumulator value is not forced unconditionally.
* Applications of op are forced.
* The list elements are not forced unconditionally.
This commit adds a test covering the discrepancy between parseDrvName's
implementation and documentation (the discrepancy was eliminated in the previous
commit).
I just had a colleague get confused by the previous phrase for good
reason. "valid" sounds like an *objective* criterion, e.g. and *invalid
signature* would be one that would be trusted by no one, e.g. because it
misformatted or something.
What is actually going is that there might be a signature which is
perfectly valid to *someone else*, but not to the user, because they
don't trust the corresponding public key. This is a *subjective*
criterion, because it depends on the arbitrary and personal choice of
which public keys to trust.
I therefore think "trustworthy" is a better adjective to use. Whether
something is worthy of trust is clearly subjective, and then "trust"
within that word nicely evokes `trusted-public-keys` and friends.
This runs the installer in a QEMU VM. Unlike the old installer test
that ran inside a declaratively built RedHat/Debian image, this uses
an image from Vagrant.
`--override-input` id snarky because it takes two arguments, so it
doesn't play well when completed in the middle of the CLI (since the
argument just after gets interpreted as its second argument). So use
`--update-input` instead
