Commit graph

716 commits

Author SHA1 Message Date
Graham Christensen 02098d2073 fetchGit: use a better caching scheme
The current usage technically works by putting multiple different
repos in to the same git directory. However, it is very slow as
Git tries very hard to find common commits between the two
repositories. If the two repositories are large (like Nixpkgs and
another long-running project,) it is maddeningly slow.

This change busts the cache for existing deployments, but users
will be promptly repaid in per-repository performance.
2018-08-17 11:27:34 -04:00
Eelco Dolstra fa4def3d46
Require libbrotli 2018-08-06 14:06:54 +02:00
Eelco Dolstra 34c17fdae5
Add a test for LegacySSHStore that doesn't require a VM 2018-08-03 21:20:38 +02:00
Eelco Dolstra bc65e02d96
Merge pull request #2326 from aszlig/fix-symlink-leak
Fix symlink leak in restricted eval mode
2018-08-03 17:01:34 +02:00
aszlig 43e28a1b75
Fix symlink leak in restricted eval mode
In EvalState::checkSourcePath, the path is checked against the list of
allowed paths first and later it's checked again *after* resolving
symlinks.

The resolving of the symlinks is done via canonPath, which also strips
out "../" and "./". However after the canonicalisation the error message
pointing out that the path is not allowed prints the symlink target in
the error message.

Even if we'd suppress the message, symlink targets could still be leaked
if the symlink target doesn't exist (in this case the error is thrown in
canonPath).

So instead, we now do canonPath() without symlink resolving first before
even checking against the list of allowed paths and then later do the
symlink resolving and checking the allowed paths again.

The first call to canonPath() should get rid of all the "../" and "./",
so in theory the only way to leak a symlink if the attacker is able to
put a symlink in one of the paths allowed by restricted evaluation mode.

For the latter I don't think this is part of the threat model, because
if the attacker can write to that path, the attack vector is even
larger.

Signed-off-by: aszlig <aszlig@nix.build>
2018-08-03 06:46:43 +02:00
Samuel Dionne-Riel 438e02529d Allows selectively adding environment variables to pure shells.
Includes documentation and test.
2018-08-02 08:21:30 -04:00
volth 403a76a18f lib.concatMap and lib.mapAttrs to be builtins 2018-07-05 02:54:09 +00:00
Will Dietz cb9ef85852 tests/search.sh: minor fix to unbreak tests after search ux merge 2018-07-03 15:37:06 -05:00
Eelco Dolstra 3b1f54cf06
Add a fromTOML primop
This is primarily useful for processing Cargo.lock files.
2018-07-03 18:39:36 +02:00
Aleksandr Pashkov 5ee4472b8b tests: more robust check for user namespaces availability (canUseSandbox)
Issue https://github.com/NixOS/nix/issues/2165
2018-06-06 00:41:01 +03:00
Eelco Dolstra 4ac4f675df
Don't require --fallback to recover from disappeared binary cache NARs 2018-06-05 16:04:41 +02:00
Eelco Dolstra ca06a9cea7
Improve binary cache fallback test 2018-06-05 15:26:12 +02:00
Eelco Dolstra a936a19da3
Remove non-existent NIX_DEBUG_SUBST 2018-06-05 13:40:45 +02:00
Eelco Dolstra 737ed88f35
Modularize config settings
Allow global config settings to be defined in multiple Config
classes. For example, this means that libutil can have settings and
evaluator settings can be moved out of libstore. The Config classes
are registered in a new GlobalConfig class to which config files
etc. are applied.

Relevant to https://github.com/NixOS/nix/issues/2009 in that it
removes the need for ad hoc handling of useCaseHack, which was the
underlying cause of that issue.
2018-05-30 13:28:01 +02:00
volth 88c1ea30e4 add docs and tests 2018-05-24 12:51:34 +00:00
Eelco Dolstra 1ad19232c4
Don't return negative numbers from the flex tokenizer
Fixes #1374.
Closes #2129.
2018-05-11 12:05:12 +02:00
Eelco Dolstra 548ad391d9
Fix builtins.add test
Nix prints the floating point number 4.0 as "4".
2018-05-02 13:49:32 +02:00
Félix Baylac-Jacqué b2f3a7411a
nix-lang: Add deep nested mixed attrs test case. 2018-05-01 14:42:34 +02:00
Tim Sears 9a714f75af
add tests for builtins.add 2018-04-29 12:03:42 -07:00
Shea Levy e2b028353b
Merge branch 'pos-crash-fix' of git://github.com/dezgeg/nix 2018-04-23 08:43:54 -04:00
Shea Levy 0aae411eaa
Merge branch 'add-test-eval-okay-nested-with' of git://github.com/ryantrinkle/nix 2018-04-23 08:43:14 -04:00
Daniel Poelzleithner f6e8ceafa6 add tests for multi search 2018-04-18 21:45:46 +02:00
Félix Baylac-Jacqué 00584bb091
parser: Allow mixed nested and top-level attrpaths
Fixes #2077.
2018-04-18 18:44:38 +02:00
Félix Baylac-Jacqué 10d33452e2
nix-lang parser: Add mixed nested attrs tests.
Part of issue #2077
2018-04-17 09:13:01 +02:00
Shea Levy 7459388448
Merge branch 'fix/ruby-shebang' of git://github.com/samueldr/nix 2018-04-09 10:54:39 -04:00
Samuel Dionne-Riel a4c9b2595e Hardcodes nix-shell instead of /usr/bin/env nix-shell...
... in the ruby shebang test.
2018-04-09 09:36:54 -04:00
Samuel Dionne-Riel 9478f88681 Adds nix-shell test for special-cased ruby interpreter.
The test fakes the interpreter only to verify the arguments it would be
given.
2018-04-08 00:45:28 -04:00
Doug Beardsley 72902ec399 Add missing eval-okay-regex-split.exp test file 2018-04-06 12:18:59 -06:00
Tuomas Tynkkynen af86132e1a libexpr: Make unsafeGetAttrPos not crash on noPos
Currently e.g. `builtins.unsafeGetAttrPos "abort" builtins` will
eventually segfault because pos->file is an unset Symbol.

Found by afl-fuzz.
2018-04-03 15:54:42 +03:00
Ryan Trinkle f2273b11f5 Add test eval-okay-nested-with 2018-03-30 15:41:16 -04:00
Linus Heckemann 12913ccf45 nix-shell: allow symlinks to .drvs
This makes persistent shell environments easier to use.
2018-03-16 22:15:24 +00:00
Eelco Dolstra 64441f0551
Merge pull request #1939 from dezgeg/lexer-fix
libexpr: Recognize newline in more places in lexer
2018-03-16 12:14:58 +01:00
Eelco Dolstra ca14b14200
Use boost::format from the boost package
Note that this only requires headers from boost so it doesn't add a
runtime dependency.

Also, use Nixpkgs 18.03.
2018-03-14 19:24:04 +01:00
Shea Levy 56f2ed0081
Merge pull request #1906 from dtzWill/fix/nix-search
nix search: tests and fix #1893 and part of #1892
2018-03-13 18:43:47 -06:00
Shea Levy aa8bbbf69d
Merge branch 'write-failure-fixes' of git://github.com/lheckemann/nix 2018-03-02 10:59:59 -05:00
Tuomas Tynkkynen a0e38c16bc libexpr: Recognize newline in more places in lexer
Flex's regexes have an annoying feature: the dot matches everything
except a newline. This causes problems for expressions like:

"${0}\
"

where the backslash-newline combination matches this rule instead of the
intended one mentioned in the comment:

    <STRING>\$|\\|\$\\ {
                    /* This can only occur when we reach EOF, otherwise the above
                    (...|\$[^\{\"\\]|\\.|\$\\.)+ would have triggered.
                    This is technically invalid, but we leave the problem to the
                    parser who fails with exact location. */
                    return STR;
                }
However, the parser actually accepts the resulting token sequence
('"' DOLLAR_CURLY 0 '}' STR '"'), which is a problem because the lexer
rule didn't assign anything to yylval. Ultimately this leads to a crash
when dereferencing a NULL pointer in ExprConcatStrings::bindVars().

The fix does change the syntax of the language in some corner cases
but I think it's only turning previously invalid (or crashing) syntax
to valid syntax. E.g.

"a\
b"

and

''a''\
b''

were previously syntax errors but now both result in "a\nb".

Found by afl-fuzz.
2018-03-02 17:30:48 +02:00
Will Dietz c577186f59 tests/search.sh: simplify, don't rely on consistent ordering 2018-03-01 13:25:06 -06:00
Tuomas Tynkkynen 77e9e1ed91 libexpr: Fix prim_replaceStrings() to work on an empty source string
Otherwise, running e.g.

nix-instantiate --eval -E --strict 'builtins.replaceStrings [""] ["X"] "abc"'

would just hang in an infinite loop.

Found by afl-fuzz.

First attempt of this was reverted in e2d71bd186 because it caused
another infinite loop, which is fixed now and a test added.
2018-02-26 19:49:13 +02:00
Will Dietz 8282c60d74 tests: test nix search behavior 2018-02-25 16:40:05 -06:00
Eelco Dolstra de4c03d201
Merge branch 'fix/dry-run-partially' of https://github.com/dtzWill/nix 2018-02-22 12:18:20 +01:00
Eelco Dolstra e2d71bd186
Revert "libexpr: Fix prim_replaceStrings() to work on an empty source string"
This reverts commit 4ea9707591.

It causes an infinite loop in Nixpkgs evaluation,
e.g. "nix-instantiate -A hello" hung.

PR #1886.
2018-02-21 15:35:28 +01:00
Tuomas Tynkkynen 4ea9707591 libexpr: Fix prim_replaceStrings() to work on an empty source string
Otherwise, running e.g.

nix-instantiate --eval -E --strict 'builtins.replaceStrings [""] ["X"] "abc"'

would just hang in an infinite loop.

Found by afl-fuzz.
2018-02-19 23:20:26 +02:00
Eelco Dolstra 75a1d96cfd
Merge branch 'register-settings' of https://github.com/shlevy/nix 2018-02-19 13:58:34 +01:00
Shea Levy b095c06139
Add splitVersion primop.
Fixes #1868.
2018-02-14 09:55:43 -05:00
Shea Levy de4934ab3b
Allow plugins to define new settings. 2018-02-13 14:43:32 -05:00
Shea Levy 3fe9767dd3
Fix plugin tests on darwin 2018-02-13 12:49:14 -05:00
Linus Heckemann 78ac3eb4eb Fix #1762
nix-store --export, nix-store --dump, and nix dump-path would previously
fail silently if writing the data out failed, because
 a) FdSink::write ignored exceptions, and
 b) the commands relied on FdSink's destructor, which ignores
    exceptions, to flush the data out.

This could cause rather opaque issues with installing nixos, because
nix-store --export would happily proceed even if it couldn't write its
data out (e.g. if nix-store --import on the other side of the pipe
failed).

This commit adds tests that expose these issues in the nix-store
commands, and fixes them for all three.
2018-02-13 14:26:19 +00:00
Shea Levy 6eb1040e90
Allow includes from nix.conf 2018-02-13 08:16:32 -05:00
Eelco Dolstra 7828dca9e8
Merge branch 'register-constant' of https://github.com/shlevy/nix 2018-02-13 12:24:48 +01:00
Eelco Dolstra 1c10a74c73
Merge branch 'plugins' of https://github.com/shlevy/nix 2018-02-13 12:15:27 +01:00
Shea Levy 081f14a169
Allow using RegisterPrimop to define constants.
This enables plugins to add new constants, as well as new primops.
2018-02-08 14:35:50 -05:00
Shea Levy 88cd2d41ac
Add plugins to make Nix more extensible.
All plugins in plugin-files will be dlopened, allowing them to
statically construct instances of the various Register* types Nix
supports.
2018-02-08 12:44:37 -05:00
Eelco Dolstra ad97a21834
nix-env: Fix parsing of --system
https://hydra.nixos.org/build/68827814
2018-02-08 15:25:03 +01:00
Eelco Dolstra 444bae44ef
dsa -> ed25519
DSS is disabled by default in NixOS 18.03.

https://hydra.nixos.org/build/68788560
2018-02-08 13:46:23 +01:00
Will Dietz c7e0be1bfc build-dry: disable failing portion of test until issue is fixed 2018-02-07 15:20:54 -06:00
Will Dietz 3780435a0e tests: Add (failing) tests for reported --dry-run issues. 2018-02-07 15:20:54 -06:00
Shea Levy 48c192ca2d
builtins.path test: Don't rely on shlevy's XDG_RUNTIME_DIR 2018-02-07 10:26:53 -05:00
Eelco Dolstra 84989d3af2
Improve filtering of ANSI escape sequences in build logs
All ANSI sequences except color setting are now filtered out. In
particular, terminal resets (such as from NixOS VM tests) are filtered
out.

Also, fix the completely broken tab character handling.
2018-02-07 15:23:57 +01:00
Eelco Dolstra abe6be578b
Merge pull request #1816 from shlevy/add-path
Add path primop.
2018-02-07 13:32:35 +01:00
Shea Levy 69d82e5c58
Add path primop.
builtins.path allows specifying the name of a path (which makes paths
with store-illegal names now addable), allows adding paths with flat
instead of recursive hashes, allows specifying a filter (so is a
generalization of filterSource), and allows specifying an expected
hash (enabling safe path adding in pure mode).
2018-02-06 16:48:08 -05:00
Eelco Dolstra 43f8ef73c6
realiseContext(): Add derivation outputs to the allowed paths
This makes import-from-derivation work in restricted mode again.
2018-02-06 15:38:45 +01:00
Eelco Dolstra f539085e65
Fix evaluation 2018-02-05 21:48:09 +01:00
Eelco Dolstra 84722d67d2
Remove nix-build --hash
Instead, if a fixed-output derivation produces has an incorrect output
hash, we now unconditionally move the outputs to the path
corresponding with the actual hash and register it as valid. Thus,
after correcting the hash in the Nix expression (e.g. in a fetchurl
call), the fixed-output derivation doesn't have to be built again.

It would still be good to have a command for reporting the actual hash
of a fixed-output derivation (instead of throwing an error), but
"nix-build --hash" didn't do that.
2018-02-03 10:08:05 +01:00
Eelco Dolstra 855699855f
Remove obsolete references to manifests
Closes #323.
2018-02-01 10:39:16 +01:00
Eelco Dolstra 98f3c75a0e
Merge pull request #1797 from dezgeg/userns-tests-fix
Fix tests using user namespaces on kernels that don't have it
2018-01-24 17:17:48 +01:00
Eelco Dolstra c382866cd2
Fix test
https://hydra.nixos.org/build/67806811
2018-01-22 17:04:08 +01:00
Eelco Dolstra 9304fde8de
Don't access tarballs.nixos.org in a test
https://hydra.nixos.org/build/67806811
2018-01-22 16:56:28 +01:00
Eelco Dolstra 3c4c30eadd
Rewrite builtin derivation environment
Also add a test.

Fixes #1803.
Closes #1805.
2018-01-19 15:00:38 +01:00
Eelco Dolstra 87e3d142cc
Add a test for --check / --repeat 2018-01-19 13:58:28 +01:00
Eelco Dolstra d4dcffd643
Add pure evaluation mode
In this mode, the following restrictions apply:

* The builtins currentTime, currentSystem and storePath throw an
  error.

* $NIX_PATH and -I are ignored.

* fetchGit and fetchMercurial require a revision hash.

* fetchurl and fetchTarball require a sha256 attribute.

* No file system access is allowed outside of the paths returned by
  fetch{Git,Mercurial,url,Tarball}. Thus 'nix build -f ./foo.nix' is
  not allowed.

Thus, the evaluation result is completely reproducible from the
command line arguments. E.g.

  nix build --pure-eval '(
    let
      nix = fetchGit { url = https://github.com/NixOS/nixpkgs.git; rev = "9c927de4b179a6dd210dd88d34bda8af4b575680"; };
      nixpkgs = fetchGit { url = https://github.com/NixOS/nixpkgs.git; ref = "release-17.09"; rev = "66b4de79e3841530e6d9c6baf98702aa1f7124e4"; };
    in (import (nix + "/release.nix") { inherit nix nixpkgs; }).build.x86_64-linux
  )'

The goal is to enable completely reproducible and traceable
evaluation. For example, a NixOS configuration could be fully
described by a single Git commit hash. 'nixos-rebuild' would do
something like

  nix build --pure-eval '(
    (import (fetchGit { url = file:///my-nixos-config; rev = "..."; })).system
  ')

where the Git repository /my-nixos-config would use further fetchGit
calls or Git externals to fetch Nixpkgs and whatever other
dependencies it has. Either way, the commit hash would uniquely
identify the NixOS configuration and allow it to reproduced.
2018-01-16 19:23:18 +01:00
Tuomas Tynkkynen 59086e459c Fix tests using user namespaces on kernels that don't have it
Disable various tests if the kernel doesn't support unprivileged user
namespaces (e.g. Arch Linux disables them) or disable them via a sysctl
(Debian, Ubuntu).

Fixes #1521
Fixes #1625
2018-01-13 19:08:41 +02:00
Will Dietz 84d9e213d2 fetchGit.sh: Test we don't "corrupt" cache if invoke w/o git avail 2018-01-09 08:58:19 -06:00
Eelco Dolstra 44272d8719
Rename "use-substitutes" to "substitute"
Commit c2154d4c84 renamed
"build-use-substitutes" to "use-substitutes", but that broke
"nix-copy-closure --use-substitutes".
2018-01-04 16:58:39 +01:00
Eelco Dolstra f68c2b5a78
Merge pull request #1770 from dtzWill/fix/run-test-sandbox-ubuntu
run.sh: include lib64 in sandbox-paths to fix on ubuntu 16.XX
2018-01-02 11:32:18 +01:00
Will Dietz 5afee18726 run.sh: include lib64 in sandbox-paths to fix on ubuntu 16.XX
(cc #1769)
2017-12-30 22:41:49 -06:00
Will Dietz 9dd2b8ac7b use libbrotli directly when available
* Look for both 'brotli' and 'bro' as external command,
  since upstream has renamed it in newer versions.
  If neither are found, current runtime behavior
  is preserved: try to find 'bro' on PATH.
* Limit amount handed to BrotliEncoderCompressStream
  to ensure interrupts are processed in a timely manner.
  Testing shows negligible performance impact.
  (Other compression sinks don't seem to require this)
2017-12-30 20:26:33 -06:00
Will Dietz 2e6f06c37e fetchGit: Fix handling of local repo when not using 'master' branch
Add tests checking this behavior.
2017-12-22 15:29:52 -06:00
Eelco Dolstra 6d80870832
release.nix: Use fetchTarball and fetchGit
In particular, using fetchGit means we don't need hackery to clean the
source tree when building from an unclean tree.
2017-12-22 11:35:32 +01:00
Eelco Dolstra e3013543d3 Fix test 2017-12-07 01:07:07 +01:00
Eelco Dolstra 338f29dbd4 nix ls-{nar,store}: Return offset of files in the NAR if known
E.g.

  $ nix ls-store --json --recursive --store https://cache.nixos.org /nix/store/b0w2hafndl09h64fhb86kw6bmhbmnpm1-blender-2.79 \
    | jq .entries.bin.entries.blender.narOffset
  400
2017-12-07 01:07:07 +01:00
Shea Levy 11a7f8ce14
Merge branch 'fetchGit-fast-revision-update' 2017-12-05 11:16:39 -05:00
aszlig 6567ab95a0
build-remote: Fix missing log output
The storeUri variable in the build-remote hook is declared very much to
the start of the main function and a bunch of lines later, the same
variable gets checked via hasPrefix() but it gets assigned *after* that
check when the most suitable machine for the build was choosen.

So I guess this was just a typo in d16fd24973
and what we really want is to either checkd the prefix *after* assigning
storeUri or use bestMachine->storeUri directly.

I choose the latter, because the former could introduce even more
regressions if the try block where the variable gets assigned terminates
early.

Nevertheless, the reason why the log output didn't work is because
hasPrefix() checked for "ssh://" in front of storeUri, but if the
storeUri isn't set correctly (or at all), we don't get the log file
descriptor set up properly, leading to no log output.

I've adjusted the remote-builds test to include a regression test for
this, so that we can make sure we get a build output when using remote
builds.

In addition to that I've tested this with two of my build farms and the
build logs are emitted correctly again.

Signed-off-by: aszlig <aszlig@nix.build>
2017-11-25 01:34:18 +01:00
Eelco Dolstra 90948a4e3a
nix-shell/nix-build: Support .drv files again
Fixes #1663.

Also handle '!<output-name>' (#1694).
2017-11-24 18:08:35 +01:00
Shea Levy eedbc4e06c
fetchGit: Ignore tarballTtl if rev is set and not in the repo.
Fixes #1697.
2017-11-24 06:09:24 -05:00
Eelco Dolstra 0fc3e581e0
tests/run.sh: Fix /build conflict when building in a sandbox
https://hydra.nixos.org/build/64519355
2017-11-21 19:43:12 +01:00
Eelco Dolstra d7da6c9ea9
fetchGit/fetchMercurial: Fix directory inclusion check
E.g. the existence of .gitignore would cause .git to be included.
2017-11-21 19:34:46 +01:00
Eelco Dolstra 4fcf44825f
Add tests for verifying/copying content-addressed paths
These don't require signatures.
2017-11-20 19:11:02 +01:00
Eelco Dolstra 0c9718aabc
Add tests for signature checking when copying between local stores 2017-11-20 19:02:57 +01:00
Eelco Dolstra 4cde04f476
Add tests for "nix run" 2017-11-20 18:36:36 +01:00
Eelco Dolstra 193330d321
Test: Replace --option with the corresponding flag 2017-11-20 17:50:49 +01:00
Eelco Dolstra 91a1987607
signed-binary-caches -> require-sigs
Unlike signed-binary-caches (which could only be '*' or ''),
require-sigs is a proper Boolean option. The default is true.
2017-11-20 17:44:07 +01:00
Eelco Dolstra 7a2b64e55c
binary-cache-public-keys -> trusted-public-keys
The name had become a misnomer since it's not only for substitution
from binary caches, but when adding/copying any
(non-content-addressed) path to a store.
2017-11-20 17:32:34 +01:00
Eelco Dolstra 27131866c9
Doh 2017-11-15 12:23:31 +01:00
Eelco Dolstra ec5b04862b
nix sign-paths: Support binary caches 2017-11-14 18:44:05 +01:00
Eelco Dolstra d6dbda7004
Add tests for "nix verify", "nix sign-paths" etc. 2017-11-14 18:24:20 +01:00
Eelco Dolstra 2c39e4eca0
Revert "Don't parse "x:x" as a URI"
This reverts commit f90f660b24.

This broke Hydra's release.nix, which contained

  preCheck = ''export LOGNAME=${LOGNAME:-foo}'';
2017-11-14 15:10:52 +01:00
Eelco Dolstra 5773d667ee
nix ls-{nar,store}: Don't abort on missing files 2017-11-14 14:49:06 +01:00
Eelco Dolstra b8d446829e
Add some tests 2017-11-14 14:37:39 +01:00
Eelco Dolstra c9c3fc710b
Rename tests/nar-index -> tests/nar-access 2017-11-14 13:27:40 +01:00
Eelco Dolstra fd10f6f241
Show when tests are skipped
Also, don't depend on tput (ncurses). It's really not needed since
ANSI escape sequences have been standardized for 35 years or so.
2017-11-07 12:09:57 +01:00