Nixpkgs issues / PRs:
* https://github.com/NixOS/nixpkgs/pull/368091
* https://github.com/NixOS/nixpkgs/issues/369366
This can be triggered with the postgresql_14 derivation from nixpkgs rev
19305d94dacca226ca048b78e6de00f599c65858
(/nix/store/bxp6g57limvwiga61vdlyvhy7i8rp6wd-postgresql-14.15.drv on
x86_64-linux): for reasons unknown to me, only the `man` and `lib` outputs
are cached on cache.nixos.org:
$ nix derivation show /nix/store/bxp6g57limvwiga61vdlyvhy7i8rp6wd-postgresql-14.15.drv | jq '.[].outputs.[].path' -r | xargs nix path-info --store https://cache.nixos.org
warning: The interpretation of store paths arguments ending in `.drv` recently changed. If this command is now failing try again with '/nix/store/bxp6g57limvwiga61vdlyvhy7i8rp6wd-postgresql-14.15.drv^*'
don't know how to build these paths:
/nix/store/m9vb40xxr6gckjzpfxnqcmjqsks2gx03-postgresql-14.15
/nix/store/nm1415wa53iawar9axwxy0an6ximhayn-postgresql-14.15-dev
/nix/store/v9vrvfhiw9gk8hj9895sb15fxvxnyylj-postgresql-14.15-debug
/nix/store/zi12g1p99g2173i8093ixbqkfh9ng87b-postgresql-14.15-doc
/nix/store/3i3fpz0xss9inampf51gp3pkx24ypxpj-postgresql-14.15-man
/nix/store/db8797h2cp4rm1cnsqrf87apkkxwwdff-postgresql-14.15-lib
error: path '/nix/store/m9vb40xxr6gckjzpfxnqcmjqsks2gx03-postgresql-14.15' does not exist in the store
Also, the derivation uses the `outputChecks` feature (and thus `__structuredAttrs`)
to make sure that e.g. the `out` output doesn't reference the `man`
output:
__structuredAttrs = true;
outputs = [ "out" "dev" "doc" "lib" "man" ];
outputChecks.out.disallowedReferences = [ "dev" "doc" "man" ];
With all that in place, the following error was hit on all CppNix / Lix
versions currently supported when trying to build the derivation above:
error: derivation contains an illegal reference specifier 'man'
The following happened here:
* The `man` & `lib` outputs were substituted at some point.
* When register outputs, the reference checks are made.
* `LocalDerivationGoal::checkOutputs` gets a map of all outputs that
were built and are NOT already registered in the store. In the example
above this means `out`, `dev`, `debug` and `doc`.
* `checkOutputs` tries to resolve the `man` output and fails to do so
because it's a store-path that's already registered and thus not part
of the map passed to `checkOutputs`.
Since the map passed to `checkOutputs` is used in various other places
that appear to assume that the paths aren't registered already, I didn't
write the already registered paths into it. Instead, I created a second
map that contains all already registered outputs and pass it as third
argument to `checkOutputs`. If the other lookups fail, this map will be
now checked before the "illegal reference specifier"-error is thrown.
This fixes the problem with `postgresql_14` for me.
Also wrote a small regression test that fails locally without the patch
in place.
Change-Id: Ieacca80c001fcfbebf6f5fe97e25c49d2724c3ff
So I recently saw it the first time in the wild, I liked that you get
interactively asked about the nix.conf settings from the flake, but
there were a few minor things that I'd like to see changed:
* The `(y/N)` was somewhere in the middle of the line. Moved it to
the end. At first I assumed it was a bug because another thread into
my terminal while I was answering the question.
* I had to say no four times for a single flake with two options. So if
you already know you don't want any of the config for _this_ flake, I
found a `No to all` switch that ignores the rest of the nix.conf
settings a little more ergonomic than having to stop the invocation,
looking up the exact wording of `--no-accept-flake-config` and
restarting it. Hence, I added it.
* Added a note where the choices which settings to trust are persisted.
My initial assumption was that this went into `nix.conf` which is not
writable on NixOS, so I said no there as well.
Change-Id: I0a0d9c403f0662df4707697a77f08e6cd003ec6f
A better fix than in 104448e75d, hence a
revert + the fix.
It turns out that this commit has the side-effect that when having e.g.
`StrictHostKeyChecking=accept-new` for a remote builder, the warnings à la
Warning: Permanently added 'builder' (ED25519) to the list of known hosts.
actually end up in the derivation's log whereas hostkey verification
errors don't, but only in the stderr of the `nix-build` invocation
(which was the motivation for the patch).
This change writes the stderr from the build-hook to
* the daemon's stderr, so that the SSH errors appear in the journal
(which was the case before 104448e75d)
* the client's stderr, as a log message
* NOT to the drv log (this is handled via `handleJSONLogMessage`)
I tried to fix the issue for legacy-ssh as well, but failed and
ultimately decided to not bother.
I know that we'll sooner or later replace the entire component, however
this is the part of the patch I have working for a while, so I figured I
might still submit it for the time being.
Change-Id: I21ca1aa0d8ae281d2eacddf26e0aa825272707e5
While debugging something else I observed that latest `main` ignores
`Control-C` on `sudo nix-build`.
After reading through the capnproto docs, it seems as if the promise
must be fulfilled to actually terminate the `promise.wait()` below.
This also applies to scenarios such as stopping the client
(`nix-build`), but the builders on the daemon-side are still running,
i.e. closes#540
Co-authored-by: eldritch horrors <pennae@lix.systems>
Change-Id: I9634d14df4909fc1b65d05654aad0309bcca8a0a
Since fb38459d6e, each `ref` is appended
with `refs/heads` unless it starts with `refs/` already. This regressed
two use-cases that worked fine before:
* Specifying a commit hash as `ref`: now, if `ref` looks like a commit
hash it will be directly passed to `git fetch`.
* Specifying a tag without `refs/tags` as prefix: now, the fetcher prepends
`refs/*` to a ref that doesn't start with `refs/` and doesn't look
like a commit hash. That way, both a branch and a tag specified in
`ref` can be fetched.
The order of preference in git is
* file in `refs/` (e.g. `HEAD`)
* file in `refs/tags/`
* file in `refs/heads` (i.e. a branch)
After fetching `refs/*`, ref is resolved the same way as git does.
Change-Id: Idd49b97cbdc8c6fdc8faa5a48bef3dec25e4ccc3
We don't support GCC anymore for building, so the overlay currently
fails to evaluate with
error: assertion '((stdenv).cc.isClang || lintInsteadOfBuild)' failed
`clangStdenv` seems like a reasonable default now.
Noticed while upgrading Lix for our Hydra fork.
Change-Id: I948a7c03b3e5648fc7c596f96e1b8053a9e7f92f
See https://git.lix.systems/lix-project/lix/issues/496.
The core idea is to be able to do e.g.
nix-instantiate -A some-nonfree-thing --arg config.allowUnfree true
which is currently not possible since `config.allowUnfree` is
interpreted as attribute name with a dot in it.
In order to change that (probably), Jade suggested to find out if there
are any folks out there relying on this behavior.
For such a use-case, it may still be possible to accept strings, i.e.
`--arg '"config.allowUnfree"'.
Change-Id: I986c73619fbd87a95b55e2f0ac03feaed3de2d2d
Closes#460
I managed to trigger the issue by having the following inputs (shortened):
authentik-nix.url = "github:nix-community/authentik-nix";
authentik-nix.inputs.poetry2nix.inputs.nixpkgs.follows = "nixpkgs";
When evaluating this using
nix-eval-jobs --flake .#hydraJobs
I got the following error:
error: cannot update unlocked flake input 'authentik-nix/poetry2nix' in pure mode
The issue we have here is that `authentik-nix/poetry2nix` was written
into the `overrideMap` which caused Nix to assume it's a new input and
tried to refetch it (#460) or errored out in pure mode
(nix-eval-jobs / Hydra).
The testcase unfortunately only involves checking for the output log
and makes sure that something *is* logged on the first fetch so that
the test doesn't rot when the logging changes since I didn't
manage to trigger the error above with the reproducer from #460. In
fact, I only managed to trigger the `cannot update unlocked flake input`
error in this context with `nix-eval-jobs`.
Change-Id: Ifd00091eec9a0067ed4bb3e5765a15d027328807
The original attempt at this introduced a regression; this commit
reverts the revert and fixes the regression.
This reverts commit 3e151d4d77.
Fix to the regression:
flakeref: fix handling of `?dir=` param for flakes in subdirs
As reported in #419[1], accessing a flake in a subdir of a Git
repository fails with the previous commit[2] applied with the error
error: unsupported Git input attribute 'dir'
The problem is that the `dir`-param is inserted into the parsed URL if a
flake is fetched from the subdir of a Git repository. However, for the
fetching part this isn't even needed. The fix is to just pass `subdir`
as second argument to `FlakeRef` (which needs a `basedir` that can be
empty) and leave the parsedURL as-is.
Added a regression test to make sure we don't run into this again.
[1] https://git.lix.systems/lix-project/lix/issues/419
[2] e22172aaf6b6a366cecd3c025590e68fa2b91bcc,
originally 3e151d4d77
Change-Id: I2c72d5a32e406a7ca308e271730bd0af01c5d18b
This comes quite often when the available job slots on all remote
builders are exhausted and this is pretty spammy.
This isn't really an issue, but expected behavior.
A better way to display this is a nom-like approach where all scheduled
builds are shown in a tree and pending builds are being marked as such
IMHO.
Change-Id: I6bc14e6054f84e3eb0768127b490e263d8cdcf89
The original idea was to fix lix#174, but for a user friendly solution,
I figured that we'd need more consistency:
* Invalid query params will cause an error, just like invalid
attributes. This has the following two consequences:
* The `?dir=`-param from flakes will be removed before the URL to be
fetched is passed to libfetchers.
* The tarball fetcher doesn't allow URLs with custom query params
anymore. I think this was questionable anyways given that an
arbitrary set of query params was silently removed from the URL you
wanted to fetch. The correct way is to use an attribute-set
with a key `url` that contains the tarball URL to fetch.
* Same for the git & mercurial fetchers: in that case it doesn't even
matter though: both fetchers added unused query params to the URL
that's passed from the input scheme to the fetcher (`url2` in the code).
It turns out that this was never used since the query parameters were
erased again in `getActualUrl`.
* Validation happens for both attributes and URLs. Previously, a lot of
fetchers validated e.g. refs/revs only when specified in a URL and
the validity of attribute names only in `inputFromAttrs`.
Now, all the validation is done in `inputFromAttrs` and `inputFromURL`
constructs attributes that will be passed to `inputFromAttrs`.
* Accept all attributes as URL query parameters. That also includes
lesser used ones such as `narHash`.
And "output" attributes like `lastModified`: these could be declared
already when declaring inputs as attribute rather than URL. Now the
behavior is at least consistent.
Personally, I think we should differentiate in the future between
"fetched input" (basically the attr-set that ends up in the lock-file)
and "unfetched input" earlier: both inputFrom{Attrs,URL} entrypoints
are probably OK for unfetched inputs, but for locked/fetched inputs
a custom entrypoint should be used. Then, the current entrypoints
wouldn't have to allow these attributes anymore.
Change-Id: I1be1992249f7af8287cfc37891ab505ddaa2e8cd
Basically I'd expect the same behavior as with `nix-build`, i.e.
with `--keep-going` the hash-mismatch error of each failing
fixed-output derivation is shown.
The approach is derived from `Store::buildPaths` (`entry-point.cc`):
instead of throwing the first build-result, check if there are any build
errors and if so, display all of them and throw after that.
Unfortunately, the BuildResult struct doesn't have an `ErrorInfo`
(there's a FIXME for that at least), so I have to construct my own here.
This is a rather cheap bugfix and I decided against touching too many
parts of libstore for that (also I don't know if that's in line with the
ongoing refactoring work).
Closes https://git.lix.systems/lix-project/lix/issues/302
Change-Id: I378ab984fa271e6808c6897c45e0f070eb4c6fac
This includes the update to libseccomp 2.5.5[1], so we don't need to
override it on our own.
[1] https://nixpk.gs/pr-tracker.html?pr=306070
Change-Id: I1fa9c7fcc23e501d75f774745107c6bb086ced70
This has the following downsides:
* you cannot build Lix against nixos-unstable.
* this will immediately break as soon as libseccomp will hit
nixos-23.11 (given that people will probably use the package.nix via
our overlay or override nixpkgs via `follows`).
Hence, removing the assert again and add a better FIXME comment.
Change-Id: I284e10cf08e1873fef70ed869a1638aa89792422
When I added the warning that an input X has an override for a
non-existent input, the recursive flake input override fix wasn't
implemented yet[1].
This patch tests that both work together.
[1] https://github.com/NixOS/nix/pull/6663
Change-Id: I90dc032029b7160ab4a97d28c480c59d3a6f0150
This reverts commit a8b3d777fb.
This undoes the revert of PR#6621, which allows nested `follows`, i.e.
{
inputs = {
foo.url = "github:bar/foo";
foo.inputs.bar.inputs.nixpkgs = "nixpkgs";
};
}
does the expected thing now. This is useful to avoid the 1000 instances
of nixpkgs problem without having each flake in the dependency tree to
expose all of its transitive dependencies for modification.
This was in fact part of Nix before and the C++ changes applied w/o
conflicts. However, it got reverted then because people didn't want to
merge lazy-trees against it which was supposed to be merged soon back in
October 2022.
Fixes: https://git.lix.systems/lix-project/lix/issues/201
Change-Id: I5ddef914135b695717b2ef88862d57ced5e7aa3c
With Linux kernel >=6.6 & glibc 2.39 a `fchmodat2(2)` is available that
isn't filtered away by the libseccomp sandbox.
Being able to use this to bypass that restriction has surprising results
for some builds such as lxc[1]:
> With kernel ≥6.6 and glibc 2.39, lxc's install phase uses fchmodat2,
> which slips through https://github.com/NixOS/nix/blob/9b88e5284608116b7db0dbd3d5dd7a33b90d52d7/src/libstore/build/local-derivation-goal.cc#L1650-L1663.
> The fixupPhase then uses fchmodat, which fails.
> With older kernel or glibc, setting the suid bit fails in the
> install phase, which is not treated as fatal, and then the
> fixup phase does not try to set it again.
Please note that there are still ways to bypass this sandbox[2] and this is
mostly a fix for the breaking builds.
This change works by creating a syscall filter for the `fchmodat2`
syscall (number 452 on most systems). The problem is that glibc 2.39
is needed to have the correct syscall number available via
`__NR_fchmodat2` / `__SNR_fchmodat2`, but this flake is still on
nixpkgs 23.11. To have this change everywhere and not dependent on the
glibc this package is built against, I added a header
"fchmodat2-compat.hh" that sets the syscall number based on the
architecture. On most platforms its 452 according to glibc with a few
exceptions:
$ rg --pcre2 'define __NR_fchmodat2 (?!452)'
sysdeps/unix/sysv/linux/x86_64/x32/arch-syscall.h
58:#define __NR_fchmodat2 1073742276
sysdeps/unix/sysv/linux/mips/mips64/n32/arch-syscall.h
67:#define __NR_fchmodat2 6452
sysdeps/unix/sysv/linux/mips/mips64/n64/arch-syscall.h
62:#define __NR_fchmodat2 5452
sysdeps/unix/sysv/linux/mips/mips32/arch-syscall.h
70:#define __NR_fchmodat2 4452
sysdeps/unix/sysv/linux/alpha/arch-syscall.h
59:#define __NR_fchmodat2 562
I added a small regression-test to the setuid integration-test that
attempts to set the suid bit on a file using the fchmodat2 syscall.
I confirmed that the test fails without the change in
local-derivation-goal.
Additionally, we require libseccomp 2.5.5 or greater now: as it turns
out, libseccomp maintains an internal syscall table and
validates each rule against it. This means that when using libseccomp
2.5.4 or older, one may pass `452` as syscall number against it, but
since it doesn't exist in the internal structure, `libseccomp` will refuse
to create a filter for that. This happens with nixpkgs-23.11, i.e. on
stable NixOS and when building Lix against the project's flake.
To work around that
* a backport of libseccomp 2.5.5 on upstream nixpkgs has been
scheduled[3].
* the package now uses libseccomp 2.5.5 on its own already. This is to
provide a quick fix since the correct fix for 23.11 is still a staging cycle
away.
We still need the compat header though since `SCMP_SYS(fchmodat2)`
internally transforms this into `__SNR_fchmodat2` which points to
`__NR_fchmodat2` from glibc 2.39, so it wouldn't build on glibc 2.38.
The updated syscall table from libseccomp 2.5.5 is NOT used for that
step, but used later, so we need both, our compat header and their
syscall table 🤷
Relevant PRs in CppNix:
* https://github.com/NixOS/nix/pull/10591
* https://github.com/NixOS/nix/pull/10501
[1] https://github.com/NixOS/nixpkgs/issues/300635#issuecomment-2031073804
[2] https://github.com/NixOS/nixpkgs/issues/300635#issuecomment-2030844251
[3] https://github.com/NixOS/nixpkgs/pull/306070
(cherry picked from commit ba6804518772e6afb403dd55478365d4b863c854)
Change-Id: I6921ab5a363188c6bff617750d00bb517276b7fe
That's expected by `build-remote` and makes sure that errors are
correctly forwarded to the user. For instance, let's say that the
host-key of `example.org` is unknown and
nix-build ../nixpkgs -A hello -j0 --builders 'ssh-ng://example.org'
is issued, then you get the following output:
cannot build on 'ssh-ng://example.org?&': error: failed to start SSH connection to 'example.org'
Failed to find a machine for remote build!
derivation: yh46gakxq3kchrbihwxvpn5bmadcw90b-hello-2.12.1.drv
required (system, features): (x86_64-linux, [])
2 available machines:
[...]
The relevant information (`Host key verification failed`) ends up in the
daemon's log, but that's not very obvious considering that the daemon
isn't very chatty normally.
This can be fixed - the same way as its done for legacy-ssh - by passing
fd 4 to the SSH wrapper. Now you'd get the following error:
cannot build on 'ssh-ng://example.org': error: failed to start SSH connection to 'example.org': Host key verification failed.
Failed to find a machine for remote build!
[...]
...and now it's clear what's wrong.
Please note that this is won't end up in the derivation's log.
For previous discussion about this change see
https://github.com/NixOS/nix/pull/7659.
Change-Id: I5790856dbf58e53ea3e63238b015ea06c347cf92
My main motivation for this change is to limit the amount of compile
jobs to make sure my machine is still usable for something else when
building a fresh Lix locally.
Also made `build` a dependency of `install`: this is analogous to
`make install` in CppNix where this both recompiles changed files and
installs the artifacts into `outputs/out`. May be a little more pleasant
to work with that, especially when you're used to contributing to
CppNix.
Change-Id: I321e2b0daf1c5e20f82c04e2dd158056c80ed86c
I haven't checked when this was exactly introduced, but on Nix 2.16 I
realized that the additional lines inserted when using `--precise` are
completely separated from the tree:
nix why-depends /nix/store/ccgr4faaxys39s091qridxg1947lggh4-evcxr-0.14.2 /nix/store/b7hvml0m3qmqraz1022fwvyyg6fc1vdy-gcc-12.2.0 --precise --extra-experimental-features nix-command
/nix/store/ccgr4faaxys39s091qridxg1947lggh4-evcxr-0.14.2
→ /nix/store/lcf37pgp3rgww67v9x2990hbfwx96c1w-gcc-wrapper-12.2.0
→ /nix/store/b7hvml0m3qmqraz1022fwvyyg6fc1vdy-gcc-12.2.0
└───bin/evcxr: …':'}.PATH=${PATH/':''/nix/store/lcf37pgp3rgww67v9x2990hbfwx96c1w-gcc-wrapper-12.2.0/bin'':'/':'}…
└───bin/cpp: …k disable=SC2193.[[ "/nix/store/b7hvml0m3qmqraz1022fwvyyg6fc1vdy-gcc-12.2.0/bin/cpp" = *++ ]] &&…
This is apparently because `std::cout` is buffered and flushed in the
end whereas the rest of the output isn't. The fix is rather simple, just
use `logger->cout` as it's already the case for the rest of the code.
This way we also don't need to insert additional newlines in the `hits`
map since that's something the logger takes care of.
Also added a small test to make sure that the layout of this is somehow
tested to reduce the risk of further regressions here.
* nix flake check: improve error message if overlay is not a lambda
Suppose you have an overlay like this
{
inputs = { /* ... */ };
outputs = { flake-utils, ... }: flake-utils.lib.eachDefaultSystem
(system: {
overlays.default = final: prev: {
};
});
}
then `nix flake check` (correctly) fails because `overlays` are supposed
to have the structure `overlays.<name> = final: prev: exp`. However, the
error-message is a little bit counter-intuitive:
error: overlay does not take an argument named 'final'
While one might guess where the error actually comes from because the
trace above says `… while checking the overlay 'overlays.x86_64-linux'`
this is still pretty confusing because it complains about an argument
not being named `final` even though that's evidently the case.
With this change, the error-message actually makes it clear what's
wrong:
[ma27@carsten:~/Projects/nix/tmp]$ 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
Since #7478 it's mandatory that `initLibStore()` is called for store
operations. However that's not the case when running `openStore()` in
Perl using the perl-bindings. That breaks e.g. `hydra-eval-jobset` when
built against Nix 2.13 which uses small portions of the store API.
I recently got fairly confused why the following expression didn't have
any effect
{
description = "Foobar";
inputs.sops-nix = {
url = github:mic92/sops-nix;
inputs.nixpkgs_22_05.follows = "nixpkgs";
};
}
until I found out that the input was called `nixpkgs-22_05` (please note
the dash vs. underscore).
IMHO it's not a good idea to not throw an error in that case and
probably leave end-users rather confused, so I implemented a small check
for that which basically checks whether `follows`-declaration from
overrides actually have corresponding inputs in the transitive flake.
In fact this was done by accident already in our own test-suite where
the removal of a `follows` was apparently forgotten[1].
Since the key of the `std::map` that holds the `overrides` is a vector
and we have to find the last element of each vector (i.e. the override)
this has to be done with a for loop in O(n) complexity with `n` being
the total amount of overrides (which shouldn't be that large though).
Please note that this doesn't work with nested expressions, i.e.
inputs.fenix.inputs.nixpkgs.follows = "...";
which is a known problem[2].
For the expression demonstrated above, an error like this will be
thrown:
error: sops-nix has a `follows'-declaration for a non-existant input nixpkgs_22_05!
[1] 2664a216e5
[2] https://github.com/NixOS/nix/issues/5790
Basically an attempt to resume fixing #5543 for a breakage introduced
earlier[1]. Basically, when evaluating an older `nixpkgs` with
`nix-shell` the following error occurs:
λ ma27 [~] → nix-shell -I nixpkgs=channel:nixos-18.03 -p nix
error: anonymous function at /nix/store/zakqwc529rb6xcj8pwixjsxscvlx9fbi-source/pkgs/top-level/default.nix:20:1 called with unexpected argument 'inNixShell'
at /nix/store/zakqwc529rb6xcj8pwixjsxscvlx9fbi-source/pkgs/top-level/impure.nix:82:1:
81|
82| import ./. (builtins.removeAttrs args [ "system" "platform" ] // {
| ^
83| inherit config overlays crossSystem;
This is a problem because one of the main selling points of Nix is that
you can evaluate any old Nix expression and still get the same result
(which also means that it *still evaluates*). In fact we're deprecating,
but not removing a lot of stuff for that reason such as unquoted URLs[2]
or `builtins.toPath`. However this property was essentially thrown away
here.
The change is rather simple: check if `inNixShell` is specified in the
formals of an auto-called function. This means that
{ inNixShell ? false }:
builtins.trace inNixShell
(with import <nixpkgs> { }; makeShell { name = "foo"; })
will show `trace: true` while
args@{ ... }:
builtins.trace args.inNixShell
(with import <nixpkgs> { }; makeShell { name = "foo"; })
will throw the following error:
error: attribute 'inNixShell' missing
This is explicitly needed because the function in
`pkgs/top-level/impure.nix` of e.g. NixOS 18.03 has an ellipsis[3], but
passes the attribute-set on to another lambda with formals that doesn't
have an ellipsis anymore (hence the error from above). This was perhaps
a mistake, but we can't fix it anymore. This also means that there's
AFAICS no proper way to check if the attr-set that's passed to the Nix
code via `EvalState::autoCallFunction` is eventually passed to a lambda
with formals where `inNixShell` is missing.
However, this fix comes with a certain price. Essentially every
`shell.nix` that assumes `inNixShell` to be passed to the formals even
without explicitly specifying it would break with this[4]. However I think
that this is ugly, but preferable:
* Nix 2.3 was declared stable by NixOS up until recently (well, it still
is as long as 21.11 is alive), so most people might not have even
noticed that feature.
* We're talking about a way shorter time-span with this change being
in the wild, so the fallout should be smaller IMHO.
[1] https://github.com/NixOS/nix/commit/9d612c393abc3a73590650d24bcfe2ee57792872
[2] https://github.com/NixOS/rfcs/pull/45#issuecomment-488232537
[3] https://github.com/NixOS/nixpkgs/blob/release-18.03/pkgs/top-level/impure.nix#L75
[4] See e.g. the second expression in this commit-message or the changes
for `tests/ca/nix-shell.sh`.
To quote Eelco in #5867:
> Unfortunately we can't do
>
> evalSettings.pureEval.setDefault(false);
>
> because then we have to do the same in main.cc (where
> pureEval is set to true), and that would allow pure-eval
> to be disabled globally from nix.conf.
Instead, a command should specify that it should be impure by
default. Then, `evalSettings.pureEval` will be set to `false;` unless
it's overridden by e.g. a CLI flag.
In that case it's IMHO OK to be (theoretically) able to override
`pure-eval` via `nix.conf` because it doesn't have an effect on commands
where `forceImpureByDefault` returns `false` (i.e. everything where pure
eval actually matters).
Closes#5867
I'm afraid I missed a few problematic `git(1)`-calls while implementing
PR #6440, sorry for that! Upon investigating what went wrong, I realized
that I only tested against the "cached"-case by accident because my
git-checkout with my system's flake was apparently cached during my
debugging.
I managed to trigger the original issue again by running:
$ git commit --allow-empty -m "tmp"
$ sudo nixos-rebuild switch --flake .# -L --builders ''
Since `repoDir` points to the checkout that's potentially owned by
another user, I decided to add `--git-dir` to each call affecting
`repoDir`.
Since the `tmpDir` for the temporary submodule-checkout is created by
Nix itself, it doesn't seem to be an issue.
Sorry for that, it should be fine now.
The `--git-dir=` must be `.` in some cases (for cached repos that are
"bare" repos in `~/.cache/nix/gitv3`). With this fix we can add
`--git-dir` to each `git`-invokation needed for `nixos-rebuild`.
To demonstrate the problem:
* You need a `git` at 2.33.3 in your $PATH
* An expression like this in a git repository:
``` nix
{
outputs = { self, nixpkgs }: {
packages.foo.x86_64-linux = with nixpkgs.legacyPackages.x86_64-linux;
runCommand "snens" { } ''
echo ${(builtins.fetchGit ./.).lastModifiedDate} > $out
'';
};
}
```
Now, when instantiating the package via `builtins.getFlake`, it fails on
Nix 2.7 like this:
$ nix-instantiate -E '(builtins.getFlake "'"$(pwd)"'").packages.foo.x86_64-linux'
fatal: unsafe repository ('/nix/store/a7j3125km4h8l0p71q6ssfkxamfh5d61-source' is owned by someone else)
To add an exception for this directory, call:
git config --global --add safe.directory /nix/store/a7j3125km4h8l0p71q6ssfkxamfh5d61-source
error: program 'git' failed with exit code 128
(use '--show-trace' to show detailed location information)
This breaks e.g. `nixops`-deployments using flakes with similar
expressions as shown above.
The cause for this is that `git(1)` tries to find the highest
`.git`-directory in the directory tree and if it finds a such a
directory, but with another owning user (root vs. the user who evaluates
the expression), it fails as above. This was changed recently to fix
CVE-2022-24765[1].
By explicitly specifying `--git-dir`, Git assumes to be in the top-level
directory and doesn't attempt to look for a `.git`-directory in the
parent directories and thus the code-path leading to said error is never
reached.
[1] https://lore.kernel.org/git/xmqqv8veb5i6.fsf@gitster.g/
When importing e.g. a local `nixpkgs` in a flake to test a change like
{
inputs.nixpkgs.url = path:/home/ma27/Projects/nixpkgs;
outputs = /* ... */
}
then the input is missing a `lastModified`-field that's e.g. used in
`nixpkgs.lib.nixosSystem`. Due to the missing `lastMoified`-field, the
mtime is set to 19700101:
result -> /nix/store/b7dg1lmmsill2rsgyv2w7b6cnmixkvc1-nixos-system-nixos-22.05.19700101.dirty
With this change, the `path`-fetcher now sets a `lastModified` attribute
to the `mtime` just like it's the case in the `tarball`-fetcher already.
When building NixOS systems with `nixpkgs` being a `path`-input and this
patch, the output-path now looks like this:
result -> /nix/store/ld2qf9c1s98dxmiwcaq5vn9k5ylzrm1s-nixos-system-nixos-22.05.20220217.dirty
A few notes:
* The `echo hi` is needed to make sure that a file that can be read by
`nix log` is properly created (i.e. some output is needed). This is
known and to be fixed in #6051.
* We explicitly ignore the floating-CA case here: the `$out` of `input3`
depends on `$out` of `input2`. This means that there are actually two
derivations - I assume that this is because at eval time (i.e.
`nix-instantiate -A`) the hash of `input2` isn't known yet and the
other .drv is created as soon as `input2` was built. This is another
issue on its own, so we ignore the case here explicitly.
To avoid that JSON messages are parsed twice in case of
remote builds with `ssh-ng://`, I split up the original
`handleJSONLogMessage` into three parts:
* `parseJSONMessage(const std::string&)` checks if it's a message in the
form of `@nix {...}` and tries to parse it (and prints an error if the
parsing fails).
* `handleJSONLogMessage(nlohmann::json&, ...)` reads the fields from the
message and passes them to the logger.
* `handleJSONLogMessage(const std::string&, ...)` behaves as before, but
uses the two functions mentioned above as implementation.
In case of `ssh-ng://`-logs the first two methods are invoked manually.
Right now when building a derivation remotely via
$ nix build -j0 -f . hello -L --builders 'ssh://builder'
it's possible later to read through the entire build-log by running
`nix log -f . hello`. This isn't possible however when using `ssh-ng`
rather than `ssh`.
The reason for that is that there are two different ways to transfer
logs in Nix through e.g. an SSH tunnel (that are used by `ssh`/`ssh-ng`
respectively):
* `ssh://` receives its logs from the fd pointing to `builderOut`. This
is directly passed to the "log-sink" (and to the logger on each `\n`),
hence `nix log` works here.
* `ssh-ng://` however expects JSON-like messages (i.e. `@nix {log data
in here}`) and passes it directly to the logger without doing anything
with the `logSink`. However it's certainly possible to extract
log-lines from this format as these have their own message-type in the
JSON payload (i.e. `resBuildLogLine`).
This is basically what I changed in this patch: if the code-path for
`builderOut` is not reached and a `logSink` is initialized, the
message was successfully processed by the JSON logger (i.e. it's in
the expected format) and the line is of the expected type (i.e.
`resBuildLogLine`), the line will be written to the log-sink as well.
Closes#5079
On Nix 2.6 the output of `nix why-depends --all` seems to be somewhat
off:
$ nix why-depends /nix/store/kn47hayxab8gc01jhr98dwyywbx561aq-nixos-system-roflmayr-21.11.20220207.6c202a9.drv /nix/store/srn5jbs1q30jpybdmxqrwskyny659qgc-nix-2.6.drv --derivation --extra-experimental-features nix-command --all
/nix/store/kn47hayxab8gc01jhr98dwyywbx561aq-nixos-system-roflmayr-21.11.20220207.6c202a9.drv
└───/nix/store/g8bpgfjhh5vxrdq0w6r6s64f9kkm9z6c-etc.drv
│ └───/nix/store/hm0jmhp8shbf3cl846a685nv4f5cp3fy-nspawn-inst.drv
| [...]
└───/nix/store/2d6q3ygiim9ijl5d4h0qqx6vnjgxywyr-system-units.drv
└───/nix/store/dil014y1b8qyjhhhf5fpaah5fzdf0bzs-unit-systemd-nspawn-hydra.service.drv
└───/nix/store/a9r72wwx8qrxyp7hjydyg0gsrwnn26zb-activate.drv
└───/nix/store/99hlc7i4gl77wq087lbhag4hkf3kvssj-nixos-system-hydra-21.11pre-git.drv
Please note that `[...]-system-units.drv` is supposed to be a direct
child of `[...]-etc.drv`.
The reason for that is that each new level printed by `printNode` is
four spaces off in comparison to `nix why-depends --precise` because the
recursive `printNode()` only prints the path and not the `tree*`-chars in
the case of `--precise` and in this format the path is four spaces further
indented, i.e. on a newline, but on the same level as the path's children, i.e.
/nix/store/kn47hayxab8gc01jhr98dwyywbx561aq-nixos-system-roflmayr-21.11.20220207.6c202a9.drv
└───/: …1-p8.drv",["out"]),("/nix/store/g8bpgfjhh5vxrdq0w6r6s64f9kkm9z6c-etc.drv",["out"]),("/nix/store/…
→ /nix/store/g8bpgfjhh5vxrdq0w6r6s64f9kkm9z6c-etc.drv
As you can see `[...]-etc.drv` is a direct child of the root, but four
spaces indented. This logic was directly applied to the code-path with
`precise=false` which resulted in `tree*` being printed four spaces too
deep.
In case of no `--precise`, `hits[hash]` is empty and the path itself
should be printed rather than hits using the same logic as for `hits[hash]`.
With this fix, the output looks correct now:
/nix/store/kn47hayxab8gc01jhr98dwyywbx561aq-nixos-system-roflmayr-21.11.20220207.6c202a9.drv
└───/nix/store/g8bpgfjhh5vxrdq0w6r6s64f9kkm9z6c-etc.drv
├───/nix/store/hm0jmhp8shbf3cl846a685nv4f5cp3fy-nspawn-inst.drv
| [...]
└───/nix/store/2d6q3ygiim9ijl5d4h0qqx6vnjgxywyr-system-units.drv
└───/nix/store/dil014y1b8qyjhhhf5fpaah5fzdf0bzs-unit-systemd-nspawn-hydra.service.drv
└───/nix/store/a9r72wwx8qrxyp7hjydyg0gsrwnn26zb-activate.drv
└───/nix/store/99hlc7i4gl77wq087lbhag4hkf3kvssj-nixos-system-hydra-21.11pre-git.drv
When I stop a download with Ctrl-C in a `nix repl` of a flake, the REPL
refuses to do any other downloads:
nix-repl> builtins.getFlake "nix-serve"
[0.0 MiB DL] downloading 'https://api.github.com/repos/edolstra/nix-serve/tarball/e9828a9e01a14297d15ca41 error: download of 'https://api.github.com/repos/edolstra/nix-serve/tarball/e9828a9e01a14297d15ca416e5a9415d4972b0f0' was interrupted
[0.0 MiB DL]
nix-repl> builtins.getFlake "nix-serve"
error: interrupted by the user
[0.0 MiB DL]
To fix this issue, two changes were necessary:
* Reset the global `_isInterrupted` variable: only because a single
operation was aborted, it should still be possible to continue the
session.
* Recreate a `fileTransfer`-instance if the current one was shut down by
an abort.
I found it somewhat confusing to have an error like
error: attribute 'getFlake' missing
if the required experimental-feature (`flakes`) is not enabled. Instead,
I'd expect Nix to throw an error just like it's the case when using e.g. `nix
flake` without `flakes` being enabled.
With this change, the error looks like this:
$ nix-instantiate -E 'builtins.getFlake "nixpkgs"'
error: Cannot call 'builtins.getFlake' because experimental Nix feature 'flakes' is disabled. You can enable it via '--extra-experimental-features flakes'.
at «string»:1:1:
1| builtins.getFlake "nixpkgs"
| ^
I didn't use `settings.requireExperimentalFeature` here on purpose
because this doesn't contain a position. Also, it doesn't seem as if we
need to catch the error and check for the missing feature here since
this already happens at evaluation time.
This actually bit me quite recently in `nixpkgs` because I assumed that
`nix-build --check` would also error out if hashes don't match anymore[1]
and so I wrongly assumed that I couldn't reproduce the mismatch error.
The fix is rather simple, during the output registration a so-called
`delayedException` is instantiated e.g. if a FOD hash-mismatch occurs.
However, in case of `nix-build --check` (or `--rebuild` in case of `nix
build`), the code-path where this exception is thrown will never be
reached.
By adding that check to the if-clause that causes an early exit in case
of `bmCheck`, the issue is gone. Also added a (previously failing)
test-case to demonstrate the problem.
[1] https://github.com/NixOS/nixpkgs/pull/139238, the underlying issue
was that `nix-prefetch-git` returns different hashes than `fetchgit`
because the latter one fetches submodules by default.
When doing e.g.
nix-build -A package --keep-failed --option \
builders \
'ssh://mfhydra?remote-store=/home/bosch/store x86_64-linux - 10 4 big-parallel'
this doesn't work properly because this build-setting is ignored.
I changed this behavior by passing the `settings.keepFailed` through the
serve-protocol to remote machines to make sure that I can introspect the
build-directory (which is particularly helpful when I have to look at a
`config.log` from a failed build for instance).
This way no derivation has to expect that these files are in the `cwd`
during the build. This is problematic for `nix-shell` where these files
would have to be inserted into the nix-shell's `cwd` which can become
problematic with e.g. recursive `nix-shell`.
To remain backwards-compatible, the location inside the build sandbox
will be kept, however using these files directly should be deprecated
from now on.
This is needed to push the adoption of structured attrs[1] forward. It's
now checked if a `__json` exists in the environment-map of the derivation
to be openend in a `nix-shell`.
Derivations with structured attributes enabled also make use of a file
named `.attrs.json` containing every environment variable represented as
JSON which is useful for e.g. `exportReferencesGraph`[2]. To
provide an environment similar to the build sandbox, `nix-shell` now
adds a `.attrs.json` to `cwd` (which is mostly equal to the one in the
build sandbox) and removes it using an exit hook when closing the shell.
To avoid leaking internals of the build-process to the `nix-shell`, the
entire logic to generate JSON and shell code for structured attrs was
moved into the `ParsedDerivation` class.
[1] https://nixos.mayflower.consulting/blog/2020/01/20/structured-attrs/
[2] https://nixos.org/manual/nix/unstable/expressions/advanced-attributes.html#advanced-attributes
I guess I misunderstood John's initial explanation about why wildcards
for outputs are sent to older stores[1]. My `nix-daemon` from 2021-03-26
also has version 1.29, but misses the wildcard[2]. So bumping seems to
be the right call.
[1] https://github.com/NixOS/nix/pull/4759#issuecomment-830812464
[2] 255d145ba7
As described in #4745 it's otherwise fairly hard to understand where
this is coming from. Say you have an expression which uses e.g.
`types.package`:
``` nix
{ outputs = { self, nixpkgs }: {
packages.x86_64-linux.hello = let
foo = nixpkgs.lib.evalModules {
modules = [
{
options.foo.bar = with nixpkgs.lib; mkOption { type = types.package; };
}
{
foo.bar = ./.;
}
];
};
in builtins.trace foo.config.foo.bar.outPath nixpkgs.legacyPackages.x86_64-linux.hello;
defaultPackage.x86_64-linux = self.packages.x86_64-linux.hello;
};
}
```
Then you'll get an error trace like this:
```
error: 'builtins.storePath' is not allowed in pure evaluation mode
at /nix/store/p4h2x6r80njkb0j2rc1xjhhl99yri3zb-source/lib/attrsets.nix:328:15:
327| let
328| path' = builtins.storePath path;
| ^
329| res =
… while evaluating the attribute 'config.foo.bar.outPath'
at /nix/store/p4h2x6r80njkb0j2rc1xjhhl99yri3zb-source/lib/attrsets.nix:332:11:
331| name = sanitizeDerivationName (builtins.substring 33 (-1) (baseNameOf path'));
332| outPath = path';
| ^
333| outputs = [ "out" ];
… while evaluating the attribute 'packages.x86_64-linux.hello'
at /nix/store/6c1rfsqzrhjw1235palzjmf5vihcpci7-source/flake.nix:3:5:
2| { outputs = { self, nixpkgs }: {
3| packages.x86_64-linux.hello = let
| ^
4| foo = nixpkgs.lib.evalModules {
```
Fixes#4745
I think that it's not very helpful to get "cached failures" in a wrong
`flake.nix`. This can be very confusing when debugging a Nix expression.
See for instance NixOS/nixpkgs#118115.
In fact, the eval cache allows a forced reevaluation which is used for
e.g. `nix eval`.
This change makes sure that this is the case for `nix build` as well. So
rather than
λ ma27 [~/Projects/exp] → ../nix/outputs/out/bin/nix build -L --rebuild --experimental-features 'nix-command flakes'
error: cached failure of attribute 'defaultPackage.x86_64-linux'
the evaluation of already-evaluated (and failed) attributes looks like
this now:
λ ma27 [~/Projects/exp] → ../nix/outputs/out/bin/nix build -L --rebuild --experimental-features 'nix-command flakes'
error: attribute 'hell' missing
at /nix/store/mrnvi9ss8zn5wj6gpn4bcd68vbh42mfh-source/flake.nix:6:35:
5|
6| packages.x86_64-linux.hello = nixpkgs.legacyPackages.x86_64-linux.hell;
| ^
7|
(use '--show-trace' to show detailed location information)
When working on some more complex Nix code, there are sometimes rather
unhelpful or misleading error messages, especially if coerce-errors are
thrown.
This patch is a first steps towards improving that. I'm happy to file
more changes after that, but I'd like to gather some feedback first.
To summarize, this patch does the following things:
* Attrsets (a.k.a. `Bindings` in `libexpr`) now have a `Pos`. This is
helpful e.g. to identify which attribute-set in `listToAttrs` is
invalid.
* The `Value`-struct has a new method named `determinePos` which tries
to guess the position of a value and falls back to a default if that's
not possible.
This can be used to provide better messages if a coercion fails.
* The new `determinePos`-API is used by `builtins.concatMap` now. With
that change, Nix shows the exact position in the error where a wrong
value was returned by the lambda.
To make sure it's still obvious that `concatMap` is the problem,
another stack-frame was added.
* The changes described above can be added to every other `primop`, but
first I'd like to get some feedback about the overall approach.
* The position of the `name`-attribute appears in the trace.
* If e.g. `meta` has no `outPath`-attribute, a `cannot coerce set to
string` error will be thrown where `pos` points to `name =` which is
highly misleading.
According to RFC4007[1], IPv6 addresses can have a so-called zone_id
separated from the actual address with `%` as delimiter. In contrast to
Nix 2.3, the version on `master` doesn't recognize it as such:
$ nix ping-store --store ssh://root@fe80::1%18 --experimental-features nix-command
warning: 'ping-store' is a deprecated alias for 'store ping'
error: --- Error ----------------------------------------------------------------- nix
don't know how to open Nix store 'ssh://root@fe80::1%18'
I modified the IPv6 match-regex accordingly to optionally detect this
part of the address. As we don't seem to do anything special with it, I
decided to leave it as part of the URL for now.
Fixes#4490
[1] https://tools.ietf.org/html/rfc4007
While the progress dots during the copying of the store work fine on a
normal terminal, those look pretty off if the script is run inside a
provisioning script of e.g. `vagrant` or `packer` where `stderr` and
`stdout` are captured:
default: .
default: ..
default: .
default: .
default: .
To work around this, the script checks with `-t 0` if it's
running on an actual terminal and doesn't show the progress if that's not
the case.
Sometimes it's necessary to fetch a git repository at a revision and
it's unknown which ref contains the revision in question. An example
would be a Cargo.lock which only provides the URL and the revision when
using a git repository as build input.
However it's considered a bad practice to perform a full checkout of a
repository since this may take a lot of time and can eat up a lot of
disk space. This patch makes a full checkout explicit by adding an
`allRefs` argument to `builtins.fetchGit` which fetches all refs if
explicitly set to true.
Closes#2409
