Compare commits

...

188 Commits

Author SHA1 Message Date
Robert Hensing
110a9ef105 Add rl-next/c-api-recoverable-errors 2025-09-10 12:49:41 +02:00
Robert Hensing
6bc358ba38 libexpr: Add recoverable errors
This provides an explicit API for call-fail-retry-succeed evaluation
flows, such as currently used in NixOps4.

An alternative design would simply reset the `Value` to the original
thunk instead of `tFailed` under the condition of catching a
`RecoverableEvalError`.
That is somewhat simpler, but I believe the presence of `tFailed` is
beneficial for possible use in the repl; being able to show the error
sooner, without re-evaluation.

The hasPos method and isEmpty function are required in order to avoid
an include loop.
2025-09-10 12:49:41 +02:00
Robert Hensing
9aa3cc9c8f libexpr: Call destructor on exception_ptr
This uses `gc_cleanup` to call the exception_ptr destructor when the
`Failed` is collected.
I don't know exactly how bad it is to deallocate `std::exception_ptr`
without destruction, but I guess it ranges from small leak to UB and crash.
2025-09-10 12:49:41 +02:00
Robert Hensing
6de4db100f Add tests/f/lang/eval-okay-tryeval-failed-thunk-reeval 2025-09-10 12:49:41 +02:00
Eelco Dolstra
b13143280c Introduce a "failed" value type
In the multithreaded evaluator, it's possible for multiple threads to
wait on the same thunk. If evaluation of the thunk results in an
exception, the waiting threads shouldn't try to re-force the thunk.
Instead, they should rethrow the same exception, without duplicating
any work.

Therefore, there is now a new value type `tFailed` that stores an
std::exception_ptr. If evaluation of a thunk/app results in an
exception, `forceValue()` overwrites the value with a `tFailed`. If
`forceValue()` encounters a `tFailed`, it rethrows the exception. So
you normally never need to check for failed values (since forcing them
causes a rethrow).
2025-09-10 12:49:41 +02:00
Jörg Thalheim
8c789db05b Merge pull request #13956 from NixOS/drop-unused-addMultipleToStoreLegacy
Drop unused LegacySSHStore::addMultipleToStoreLegacy()
2025-09-10 10:52:05 +02:00
Eelco Dolstra
f8b15bfc7f Merge pull request #13951 from NixOS/drop-old-daemon-protocol
Remove support for worker protocol version < 18
2025-09-10 10:43:48 +02:00
Eelco Dolstra
5013b38df4 Drop unused LegacySSHStore::addMultipleToStoreLegacy() 2025-09-10 10:27:07 +02:00
Eelco Dolstra
247d16a530 Merge pull request #13954 from xokdvium/empty-list-elems-fix
libexpr: Fix Value::mkList for empty lists
2025-09-10 10:23:49 +02:00
Eelco Dolstra
d1d3ed6241 Add release note 2025-09-10 10:20:37 +02:00
Jörg Thalheim
7d26bf8cc7 Merge pull request #13906 from obsidiansystems/derivation-builder-simpler
More `DerivationBuilder` simplifications
2025-09-10 09:58:07 +02:00
Jörg Thalheim
9c186c35fa Merge pull request #13932 from NixOS/move-pathInfoCache
Reduce false sharing between pathInfoCache and Store
2025-09-10 09:56:46 +02:00
Sergei Zimmerman
2ed2c79721 libexpr: Fix Value::mkList for empty lists
This code used to save the pointer to a small
list allocated on the stack to the Value, which
is unintended.
2025-09-10 01:37:39 +03:00
Sergei Zimmerman
3c331b7ef3 Merge pull request #13953 from xokdvium/value-alignment
libexpr: Overalign Value to 16 bytes
2025-09-09 20:44:43 +00:00
Sergei Zimmerman
4524235af4 libexpr: Overalign Value to 16 bytes
This is necessary to make use of 128 bit atomics on x86_64 [1],
since MOVAPD, MOVAPS, and MOVDQA need memory operands to be 16-byte
aligned. We are not losing anything here, because Value is already 16-byte
wide and Boehm allocates memory in granules that are 16 bytes by default
on 64 bit systems [2].

[1]: https://patchwork.sourceware.org/project/gcc/patch/YhxkfzGEEQ9KHbBC@tucnak/
[2]: 54ac18ccbc/include/gc/gc_tiny_fl.h (L31-L33)
2025-09-09 22:18:52 +03:00
Eelco Dolstra
86d19956f2 Remove WorkerProto::Op::ImportPaths
This was obsoleted in May 2016 (538a64e8c3).
2025-09-09 15:41:17 +02:00
Eelco Dolstra
4fb61bc5af Remove WorkerProto::Op::ExportPath
This was obsoleted in May 2016 (538a64e8c3).
2025-09-09 15:41:12 +02:00
Eelco Dolstra
137a55122c Remove support for daemon protocol version < 18
Version 18 was introduced in November 2016 (4b8f1b0ec0).
2025-09-09 15:41:01 +02:00
Eelco Dolstra
7658f00bb1 Merge pull request #13941 from NixOS/dependabot/github_actions/actions/labeler-6
build(deps): bump actions/labeler from 5 to 6
2025-09-09 09:37:08 +02:00
Eelco Dolstra
a97c5df47c Merge pull request #13939 from DeterminateSystems/fix-inputs-from-ignoring-dir-param-upstreaming
Pass `dir` in extraAttrs when overriding the registry
2025-09-09 09:36:33 +02:00
Sergei Zimmerman
371623bf0c Merge pull request #13940 from xokdvium/unbracketed-ipv6
libstore: Reallow unbracketed IPv6 addresses in store references
2025-09-08 23:21:34 +00:00
dependabot[bot]
7128abd217 build(deps): bump actions/labeler from 5 to 6
Bumps [actions/labeler](https://github.com/actions/labeler) from 5 to 6.
- [Release notes](https://github.com/actions/labeler/releases)
- [Commits](https://github.com/actions/labeler/compare/v5...v6)

---
updated-dependencies:
- dependency-name: actions/labeler
  dependency-version: '6'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
2025-09-08 22:00:58 +00:00
Sergei Zimmerman
7cc654afa9 libstore: Reallow unbracketed IPv6 addresses in store references
This implements a special back-compat shim to specifically allow
unbracketed IPv6 addresses in store references. This is something
that is relied upon in the wild and the old parsing logic accepted
both ways (brackets were optional). This patch restores this behavior.
As always, we didn't have any tests for this.

Addresses #13937.
2025-09-09 00:41:03 +03:00
Cole Helbling
38663fb434 Pass dir in extraAttrs when overriding the registry
This is handled similarly in the handler for `--override-flake` in
`MixEvalArgs`.
2025-09-08 09:00:59 +02:00
Cole Helbling
ed6ef7cdf4 Test that using --inputs-from with a flakeref that has a dir works
Will not pass until the next commit.
2025-09-08 09:00:59 +02:00
Eelco Dolstra
12db0726e9 Merge pull request #13934 from DeterminateSystems/fix-flake-registry-ignoring-dir-param-upstreaming
Fix flake registry ignoring `dir` parameter
2025-09-08 07:50:54 +02:00
Eelco Dolstra
525245181a Merge pull request #13933 from NixOS/local-store-state
LocalStore::State: Put behind a ref to reduce false sharing
2025-09-08 06:22:04 +02:00
Eelco Dolstra
9302ec5e0e Add comment 2025-09-08 05:57:02 +02:00
Cole Helbling
9c832a08b0 fixup: cached case
I couldn't come up with a test that failed before this, but my existing
test still passes so 🤷
2025-09-07 19:40:24 +02:00
Cole Helbling
bccdb95a86 Fix flake registry ignoring dir parameter
This broke in e3042f10af.
2025-09-07 19:40:24 +02:00
Cole Helbling
258d41bfb6 Test that dir is propagated from registry entry 2025-09-07 19:40:23 +02:00
Tom Westerhout
dbc235cc62 Generalize recognized git url schemas (#13925)
Use `parseUrlScheme` instead of manually parsing `url.scheme`.
2025-09-07 15:22:20 +02:00
Eelco Dolstra
df9b3bfba8 Merge pull request #13845 from NixOS/nix-flake-check-build-test
Add a test for `nix flake check` building checks
2025-09-07 14:53:19 +02:00
Eelco Dolstra
14c001d613 Add a test for nix flake check building checks 2025-09-07 14:41:40 +02:00
Eelco Dolstra
e791ede495 LocalStore::State: Put behind a ref to reduce false sharing 2025-09-07 14:32:24 +02:00
Eelco Dolstra
a73cf447ac Reduce false sharing between pathInfoCache and Store
`perf c2c` shows a lot of cacheline conflicts between purely read-only
Store methods (like `parseStorePath()`) and the Sync classes. So
allocate pathInfoCache separately to avoid that.
2025-09-07 14:27:38 +02:00
Eelco Dolstra
9ff427d7ba Merge pull request #13911 from xokdvium/store-uri-daemon-local
libstore: Do not normalize daemon -> unix://, local -> local://
2025-09-07 14:10:45 +02:00
Eelco Dolstra
4dd27a292c Merge pull request #13929 from NixOS/remove-unused
Remove unused function setChildSignalMask()
2025-09-07 13:30:09 +02:00
Jörg Thalheim
5ae1b5f88b Merge pull request #13916 from sinanmohd/fix/develop-interactive-shell
nix/develop: pass down the interactive shell to subshells
2025-09-07 10:14:25 +02:00
Eelco Dolstra
a7c6a42344 Merge pull request #13923 from NixOS/fix-multithreaded-chroot-hang
Fix hang in enterChroot() draining userNamespaceSync
2025-09-07 09:13:54 +02:00
Eelco Dolstra
f363d958a7 Fix hang in enterChroot() draining userNamespaceSync
Calling `drainFD()` will hang if another process has the write side
open, since then the child won't get an EOF. This can happen if we
have multiple threads doing a build, since in that case another thread
may fork a child process that inherits the write side of the first
thread.

We could set O_CLOEXEC on the write side (using pipe2()) but it won't
help here since we don't always do an exec() in the child, e.g. in the
case of builtin builders. (We need a "close-on-fork", not a
"close-on-exec".)
2025-09-07 01:12:44 +02:00
Eelco Dolstra
a44dcbff13 Remove unused function setChildSignalMask() 2025-09-06 23:02:57 +02:00
John Ericson
12b6d8d208 Merge pull request #13924 from xokdvium/dead-code
libexpr: Remove decl for undefined overload of Value::mkPath
2025-09-06 10:47:26 -04:00
Sergei Zimmerman
bbdabe4973 libexpr: Remove decl for undefined overload of Value::mkPath 2025-09-06 16:36:16 +03:00
Jörg Thalheim
1d62ccdb3d Merge pull request #13767 from ethanavatar/master
libutil, libexpr: #10542 abstract over getrusage for getting cpuTime stat and implement windows version
2025-09-06 09:26:13 +02:00
Jörg Thalheim
533c6d38aa Merge pull request #13901 from Mic92/fix-macos-hup-detection
Fix macOS HUP detection using kqueue instead of poll
2025-09-06 09:19:51 +02:00
Jörg Thalheim
dbc8d0ab64 Merge pull request #13919 from xokdvium/smaller-bindings
libexpr: Slim down Bindings to 8 bytes (on 64 bit systems)
2025-09-06 09:11:48 +02:00
Sergei Zimmerman
738924b705 libexpr: Slim down Bindings to 8 bytes (on 64 bit systems)
Since the only construction and push_back() calls
to Bindings happen through the `BindingsBuilder` [1] we don't
need to keep `capacity` around on the heap anymore. This saves 8 bytes
(because of the member alignment padding)
per one Bindings allocation. This isn't that much, but it does
save significant memory.

This also shows that the Bindings don't necessarily have to
be mutable, which opens up opportunities for doing small bindings
optimization and storing a 1-element Bindings directly in Value.

For the following scenario:

nix-env --query --available --out-path --file ../nixpkgs --eval-system x86_64-linux

(nixpkgs revision: ddcddd7b09a417ca9a88899f4bd43a8edb72308d)

This patch results in reduction of `sets.bytes` 13115104016 -> 12653087640,
which amounts to 462 MB less bytes allocated for Bindings.

[1]: Not actually, `getBuiltins` does mutate bindings, but this is pretty
     inconsequential and doesn't lead to problems.
2025-09-06 00:23:54 +03:00
sinanmohd
211cbe4abf nix/develop: pass down the interactive shell to subshells 2025-09-05 20:18:25 +05:30
Sergei Zimmerman
3513ab13dc libstore: Do not normalize daemon -> unix://, local -> local://
This is relied upon (specifically the `local` store) by existing
tooling [1] and we broke this in 3e7879e6df (which
was first released in 2.31).

To lessen the scope of the breakage we should not normalize "auto" references
and explicitly specified references like "local" or "daemon". It also makes
sense to canonicalize local://,daemon:// to be more compatible with prior
behavior.

[1]: 05e1b3cba2/lib/NOM/Builds.hs (L60-L64)
2025-09-05 04:14:36 +03:00
John Ericson
49e9c14e2f Merge pull request #13900 from NixOS/fix-mingw-windows-build
Fix downstream MinGW build by not looking for Boost Regex
2025-09-04 21:05:12 -04:00
John Ericson
25d3c197b8 Merge pull request #13902 from NixOS/ssh-master-deadlock
Fix deadlock in SSHMaster::addCommonSSHOpts()
2025-09-03 21:44:06 -04:00
John Ericson
2acb9559d5 Combine DerivationBuilder::{prepareBuild,startBuilder}
After many other cleanups, it turns out there is no reason for these to
be separate methods. We can combine them to simplify things.
2025-09-03 17:58:50 -04:00
John Ericson
14c206f05a DerivationBuilder no more callback soup for logging
`startBuilder` just returns the descriptor for the pipe now.
2025-09-03 17:34:45 -04:00
John Ericson
7f3314a68c DerivationBuilder::initialOutputs make const
At one point I remember it did mutatate `initialOutputs`, but not
anymore!
2025-09-03 17:34:45 -04:00
John Ericson
b69576e2b3 Merge pull request #13905 from obsidiansystems/derivation-building-goal-simplify-0
Derivation building goal simplify -- no `goto`
2025-09-03 17:34:27 -04:00
John Ericson
7b22cd5105 Merge pull request #13839 from Mic92/infra
don't include derivation name in temporary build directories
2025-09-03 17:15:03 -04:00
Jörg Thalheim
1732b4a61b Merge pull request #13885 from netadr/fix-ssh-key-ids
libfetchers: Fix SSH key types for sk type keys
2025-09-03 23:13:48 +02:00
John Ericson
819bf13607 Merge pull request #13880 from Mic92/static-alloc-symbol-ids
libexpr: Convert Symbol comparisons to switch statements
2025-09-03 17:13:12 -04:00
Jörg Thalheim
81e068ab8a Merge pull request #13904 from NixOS/c-ffi-improvements
C ffi improvements
2025-09-03 23:10:27 +02:00
John Ericson
a30bf96349 DerivationBuildingGoal::initialOutputs make local variable
Also inline `assertPathValidity` in the process.
2025-09-03 17:04:13 -04:00
John Ericson
c0c2a89f05 DerivationBuildingGoal::initialOutputs move initialization down to tryToBuild
Will help us make this a local variable.
2025-09-03 17:04:08 -04:00
John Ericson
450633aa8c Move machineName from DerivationBuildingGoal to HookInstance
Exactly why is is correct is a little subtle, because sometimes the
worker is owned by the worker. But the commit message in
e437b08250 explained the situation well
enough: I made that commit message part of the ABI docs, and now it
should be understandable to the next person.
2025-09-03 17:03:56 -04:00
netadr
671c21db9f libfetchers: Fix SSH key identifiers for sk type keys
libfetchers: Mark ssh-ecdsa-sk key type mapping as a TODO for now
2025-09-03 22:56:33 +02:00
John Ericson
8089102164 Separate internal from non-internal unit tests of the C API
This helps us make sure that the external C API is sufficient for the
tasks that we think it is sufficient for.
2025-09-03 22:50:42 +02:00
John Ericson
f6bc47bc50 nix_store_realise: Improve typing of store path
Use `StorePath *` not `const char *`.
2025-09-03 22:50:42 +02:00
John Ericson
fa76b6e215 nix store benchmarks: Only get unit test dir from env var 2025-09-03 22:50:42 +02:00
John Ericson
44d096f68d nix_store_is_valid_path param path should be const 2025-09-03 22:50:42 +02:00
John Ericson
7e4608a3f8 More extern "C" for FFI
This allows us to catch the header and file getting out of sync, because
we are not doing overloading by mistake.
2025-09-03 22:50:42 +02:00
John Ericson
eb56b181ae DerivationBuildingGoal: Make almost everything private 2025-09-03 16:25:12 -04:00
John Ericson
c6ba120000 DerivationBuildingGoal::started make local (lambda) variable 2025-09-03 16:19:35 -04:00
John Ericson
3b9c510ab1 DerivationBuildingGoal::outputLocks make local variable 2025-09-03 16:19:35 -04:00
John Ericson
a63ac8d98b Inline DerivationBuildingGoal::hookDone 2025-09-03 16:19:35 -04:00
John Ericson
51dadaded4 Move up assert(!hook);
We don't need to keep doing this every loop iteration, hook stuff it is only set
above.
2025-09-03 16:19:35 -04:00
John Ericson
7c1e5b3345 In DerivationBuildingGoal Demote actLock to local variable
It doesn't need to be a field any more, because we just use it with two
loops.
2025-09-03 16:19:35 -04:00
John Ericson
4c44a213a3 Get rid of a tryToBuild tail recursive call with loop
This will make it easier to convert somethings to RAII.
2025-09-03 16:19:35 -04:00
John Ericson
95c5779880 DerivationBuildingGoal::tryToBuild pull hook waiting out of switch
Do this with a new `useHook` boolean we carefully make sure is set in
all cases. This change isn't really worthwhile by itself, but it allows
us to make further refactors (see later commits) which are
well-motivated.
2025-09-03 16:19:35 -04:00
Eelco Dolstra
c7603c61c8 Mark tmpDir as const 2025-09-03 20:17:42 +02:00
Eelco Dolstra
2fe629c5d4 Fix deadlock in SSHMaster::addCommonSSHOpts()
When useMaster is true, startMaster() acquires the state lock, then
calls isMasterRunning(), which calls addCommonSSHOpts(), which tries
to acquire the state lock again, causing a deadlock.

The solution is to move tmpDir out of the state. It doesn't need to be
there in the first place because it never changes.
2025-09-03 17:49:24 +02:00
Jörg Thalheim
1286d5db78 Fix macOS HUP detection using kqueue instead of poll
On macOS, poll() is fundamentally broken for HUP detection. It loses event
subscriptions when EVFILT_READ fires without matching the requested events
in the pollfd. This causes daemon processes to linger after client disconnect.

This commit replaces poll() with kqueue on macOS, which is what poll()
uses internally but without the bugs. The kqueue implementation uses
EVFILT_READ which works for both sockets and pipes, avoiding EVFILT_SOCK
which only works for sockets.

On Linux and other platforms, we continue using poll() with the standard
POSIX behavior where POLLHUP is always reported regardless of requested events.

Based on work from the Lix project (https://git.lix.systems/lix-project/lix)
commit 69ba3c92db3ecca468bcd5ff7849fa8e8e0fc6c0

Fixes: https://github.com/NixOS/nix/issues/13847
Related: https://git.lix.systems/lix-project/lix/issues/729
Apple bugs: rdar://37537852 (poll), FB17447257 (poll)

Co-authored-by: Jade Lovelace <jadel@mercury.com>
2025-09-03 11:33:23 +02:00
Jörg Thalheim
cbcb434cb3 libexpr: Convert Symbol comparisons to switch statements
Now that Symbols are statically allocated at compile time with known IDs,
we can use switch statements instead of if-else chains for Symbol comparisons.
This provides better performance through compiler optimizations like jump tables.

Changes:
- Add public getId() method to Symbol class to access the internal ID
- Convert if-else chains comparing Symbol values to switch statements
  in primops.cc's derivationStrictInternal function
- Simplify control flow by removing the 'handled' flag and moving the
  default attribute handling into the switch's default case

The static and runtime Symbol IDs are guaranteed to match by the
copyIntoSymbolTable implementation which asserts this invariant.

Co-authored-by: John Ericson <git@JohnEricson.me>
2025-09-03 10:13:12 +02:00
Sergei Zimmerman
1935c19705 Merge pull request #13890 from xokdvium/mkstring-no-copy
Re-introduce mkStringNoCopy (revised)
2025-09-02 17:13:17 +00:00
John Ericson
6bdb5e8e09 Fix downstream MinGW build by not looking for Boost Regex 2025-09-02 10:41:39 -04:00
John Ericson
b806440808 Merge pull request #13894 from NixOS/more-url-testing
More URL testing
2025-09-02 00:10:13 -04:00
John Ericson
7f91e91876 More URL testing
More parameterized tests, we can have more coverage.
2025-09-01 18:26:21 -04:00
John Ericson
ab095c029c Merge pull request #13891 from NixOS/another-url-test
Add another `fixGitURL` test
2025-09-01 17:36:18 -04:00
John Ericson
7195250fc4 Add another fixGitURL test
Also improve a similar `parseURL` test.
2025-09-01 17:19:26 -04:00
Sergei Zimmerman
34181afc6a libexpr: Use mkStringNoCopy in prim_typeOf
This would lead to an unnecessary allocation. Not
a significant issue by any means, but it doesn't
have to allocate for most cases.
2025-09-02 00:16:11 +03:00
Eelco Dolstra
d62cfc1c97 Re-introduce mkStringNoCopy (revised)
In b70d22b `mkStringNoCopy()` was renamed to
`mkString()`, but this is a bit risky since in code like

    vStringRegular.mkString("regular");

we want to be sure that the right overload is picked. (This is
especially problematic since the overload that takes an
`std::string_view` *does* allocate.)  So let's be explicit.

(Rebased from https://github.com/NixOS/nix/pull/11551)
2025-09-02 00:16:06 +03:00
John Ericson
3a19ea96d9 Merge pull request #13888 from NixOS/old-busted-git-url-with-tests
Old busted git url with tests
2025-09-01 16:46:01 -04:00
Farid Zakaria
2b310aee13 A few more URL tests
Adapted from commit 04ad66af5f
2025-09-01 16:31:45 -04:00
John Ericson
d2f1860ee5 Revert "Improve Git URI handling"
I (@Ericson2314) messed up. We were supposed to test the status quo
before landing any new chnages, and also there is one change that is not
quite right (relative paths).

I am reverting for now, and then backporting the test suite to the old
situation.

This reverts commit 04ad66af5f.
2025-09-01 16:13:32 -04:00
Jörg Thalheim
a0ce514769 Merge pull request #13866 from obsidiansystems/more-derivation-builder-cleanup
Even more `DerivationBuilder` cleanup
2025-09-01 20:35:16 +02:00
Jörg Thalheim
0d300112fa Merge pull request #13862 from obsidiansystems/build-failure-content-vs-presentation
Properly separater builder failure content and presentation
2025-09-01 20:25:50 +02:00
Jörg Thalheim
de7f137f31 Merge pull request #13860 from obsidiansystems/derivation-building-resources-code-cleanup
Derivation building resources code cleanup
2025-09-01 20:22:30 +02:00
John Ericson
7fde4f7d6f Merge pull request #13821 from fzakaria/fzakaria/improve-fixgiturl
Improve Git URI handling
2025-09-01 14:15:17 -04:00
Jörg Thalheim
dc29cdf66d Merge pull request #13858 from obsidiansystems/no-more-defered-exception
Get rid of `delayedException` in `DerivationBuilder`
2025-09-01 20:11:51 +02:00
Jörg Thalheim
3e0fb3f8d2 Merge pull request #13881 from xokdvium/pass-url-verbatim
lib{store,fetchers}: Pass URLs specified directly verbatim to FileTra…
2025-09-01 20:06:50 +02:00
Farid Zakaria
04ad66af5f Improve Git URI handling
Git URI can also support scp style links similar to git itself.

This change augments the function fixGitURL to better handle the scp
style urls through a minimal parser rather than regex which has been
found to be brittle.

* Support for IPV6 added
* New test cases added for fixGitURL
* Clearer documentation on purpose and goal of function
* More `std::string_view` for performance
* A few more URL tests

Fixes #5958
2025-09-01 14:04:04 -04:00
Jörg Thalheim
fea4a29c0a Merge pull request #13883 from xokdvium/toml-timestamps-reapply
Reapply "Merge pull request #13741 from xokdvium/toml-timestamps"
2025-09-01 09:12:33 +02:00
Sergei Zimmerman
e548700010 lib{store,fetchers}: Pass URLs specified directly verbatim to FileTransferRequest
The URL should not be normalized before handing it off to cURL, because
builtin fetchers like fetchTarball/fetchurl are expected to work with
arbitrary URLs, that might not be RFC3986 compliant. For those cases
Nix should not normalize URLs, though validation is fine. ParseURL and
cURL are supposed to match the set of acceptable URLs, since they implement
the same RFC.
2025-09-01 02:22:23 +03:00
Emily
acd627fa46 tests/functional/lang: Add tests for builtins.fromTOML overflow
This adds regression tests for fromTOML overflow/underflow behavior.
Previous versions of toml11 used to saturate, but this was never an
intended behavior (and Snix/Nix 2.3/toml11 >= 4.0 validate this).

(cherry picked from Lix [1,2])

[1]: 7ee442079d
[2]: 4de09b6b54
2025-09-01 01:49:15 +03:00
Sergei Zimmerman
8251305aff Reapply "Merge pull request #13741 from xokdvium/toml-timestamps"
This reverts commit 75740fbd75.
2025-09-01 01:26:14 +03:00
Jörg Thalheim
73cdfe7066 Merge pull request #13878 from urbas/hacking-instructions
hacking.md: set installation outputs as well
2025-08-31 13:58:36 +02:00
Jörg Thalheim
1f7d43e5bd Merge pull request #13879 from xokdvium/static-alloc-symbol-ids
libexpr: Statically allocate commonly used symbols
2025-08-31 13:40:55 +02:00
Sergei Zimmerman
363620dd24 libexpr: Statically allocate commonly used symbols
The motivation for this change is two-fold:

1. Commonly used Symbol values can be referred to
   quite often and they can be assigned at compile-time
   rather than runtime.

2. This also unclutters EvalState constructor, which was
   getting very long and unreadable.

Spiritually similar to https://gerrit.lix.systems/c/lix/+/2218,
though that patch doesn't allocate the Symbol at compile time.

Co-authored-by: eldritch horrors <pennae@lix.systems>
2025-08-31 13:24:06 +02:00
Matej Urbas
112f311c50 hacking.md: set installation outputs as well 2025-08-31 09:53:14 +01:00
John Ericson
2746985d90 Merge pull request #13877 from xokdvium/opt-print-string
libstore: Get rid of allocations in printString, allocate 2K bytes on the stack
2025-08-31 00:42:18 -04:00
Sergei Zimmerman
e1c9bc0ef6 libstore: Get rid of allocations in printString, allocate 2K bytes on the stack
Looking at perf:

   0.21 │       push   %rbp
   0.99 │       mov    %rsp,%rbp
        │       push   %r15
   0.25 │       push   %r14
        │       push   %r13
   0.49 │       push   %r12
   0.66 │       push   %rbx
   1.23 │       lea    -0x10000(%rsp),%r11
   0.23 │ 15:   sub    $0x1000,%rsp
   1.01 │       orq    $0x0,(%rsp)
  59.12 │       cmp    %r11,%rsp
   0.27 │     ↑ jne    15

Seems like 64K is too much to have on the stack for each invocation, considering
that only a minuscule number of allocations are actually larger than 4K.

There's actually no good reason this function should use so much stack space. Or
use small_string at all. Everything can be done in small chunks that don't require
any memory allocations and use up 2K bytes on the stack.

This patch also adds a microbenchmark for tracking the unparsing performance. Here
are the results for this change:

(Before)

BM_UnparseRealDerivationFile/hello         7275 ns         7247 ns        96093 bytes_per_second=232.136Mi/s
BM_UnparseRealDerivationFile/firefox      40538 ns        40376 ns        17327 bytes_per_second=378.534Mi/s

(After)

BM_UnparseRealDerivationFile/hello         3228 ns         3218 ns       215671 bytes_per_second=522.775Mi/s
BM_UnparseRealDerivationFile/firefox      39724 ns        39584 ns        17617 bytes_per_second=386.101Mi/s

This translates into nice evaluation performance improvements (compared to 18c3d2348f):

Benchmark 1: GC_INITIAL_HEAP_SIZE=8G old-nix/bin/nix-instantiate ../nixpkgs -A nixosTests.gnome --readonly-mode
  Time (mean ± σ):      3.111 s ±  0.021 s    [User: 2.513 s, System: 0.580 s]
  Range (min … max):    3.083 s …  3.143 s    10 runs

Benchmark 2: GC_INITIAL_HEAP_SIZE=8G result/bin/nix-instantiate ../nixpkgs -A nixosTests.gnome --readonly-mode
  Time (mean ± σ):      3.037 s ±  0.038 s    [User: 2.461 s, System: 0.558 s]
  Range (min … max):    2.960 s …  3.086 s    10 runs
2025-08-31 00:48:37 +03:00
John Ericson
18c3d2348f Merge pull request #13875 from xokdvium/restore-weird-flakeref-path
libfetchers: Restore path separator ignoring behavior for indirect an…
2025-08-30 11:17:42 -04:00
Sergei Zimmerman
a38ebdd511 libfetchers: Restore path separator ignoring behavior for indirect and git-archive flakerefs
Old versions of nix happily accepted a lot of weird flake references,
which we didn't have tests for, so this was accidentally broken in
c436b7a32a.

This patch restores previous behavior and adds a plethora of tests
to ensure we don't break this in the future.

These test cases are aligned with how 2.18/2.28 parsed flake references.
2025-08-30 14:40:56 +03:00
John Ericson
401e7fe3ad Merge pull request #13873 from xokdvium/fix-mingw
libfetchers: Fix mingw build
2025-08-29 19:56:20 -04:00
Sergei Zimmerman
b88a22504f libfetchers: Fix mingw build 2025-08-30 02:36:16 +03:00
John Ericson
511d885d60 Merge pull request #13872 from xokdvium/fix-indirect-flake-refs
libflake: Fix flake id flake refs with revisions
2025-08-29 18:51:25 -04:00
Sergei Zimmerman
3ef3f525c3 libflake: Fix flake id flake refs with revisions
Starting from c436b7a32a
this used to lead to assertion failures like:

> std::string nix::ParsedURL::renderAuthorityAndPath() const: Assertion `path.empty() || path.front().empty()' failed.

This has the bugfix for the issue and regressions tests
so that this gets properly tested in the future.
2025-08-30 01:26:51 +03:00
John Ericson
53a7d87b93 Merge pull request #13871 from obsidiansystems/fix-refactor-bug
`DerivationBuildingGoal::done*` restore `outputLocks.unlock()`
2025-08-29 18:25:43 -04:00
John Ericson
a8c4cfae26 DerivationBuildingGoal::done* restore outputLocks.unlock()
This was accidentally removed in
169033001d.
2025-08-29 17:49:11 -04:00
John Ericson
d50d4b01c7 Merge pull request #13867 from xokdvium/fix-13482
nix/develop: Fix misleading ignored error when run with --arg/--argstr
2025-08-29 17:17:25 -04:00
Sergei Zimmerman
b6f98b52a4 nix/develop: Fix misleading ignored error when run with --arg/--argstr
This would print erroneous and misleading diagnostics like:

> error (ignored): error: '--arg' and '--argstr' are incompatible with flakes

When run with --expr/--file. Since this installable is used to get the
bash package it doesn't make sense to check this.
2025-08-30 00:03:54 +03:00
John Ericson
d7ed86ceb1 Move deleting redirected outputs in to cleanupBuild
It is only done in the `force = true` case, and the only
`cleanupBuild(true)` call is right after where it used to be, so this
has the exact same behavior as before.
2025-08-29 16:10:25 -04:00
John Ericson
76125f8eb1 Get rid of Finally in DerivationBuilderImpl::unprepareBuild
Calling `reset` on this `std::optional` field of `DerivationBuilderImpl`
is also what the (automatically created) destructor of
`DerivationBuilderImpl` will do. We should be making sure that the
derivation builder is cleaned up by the goal anyways, and if we do that,
then this `Finally` is no longer needed.
2025-08-29 13:22:36 -04:00
Jörg Thalheim
0d006aedd6 Merge pull request #13854 from obsidiansystems/register-outputs-slight-simplify
Simplify handling of statuses for build errors
2025-08-29 07:20:55 +02:00
Jörg Thalheim
04d2122de2 Merge pull request #13861 from xokdvium/terminate-for-unreachable
libutil: Try to call std::terminate for panic, use C++20 std::source_location
2025-08-29 07:15:49 +02:00
John Ericson
8825bfa7fe Properly separater builer failure content and presentation
Before, had a very ugly `appendLogTailErrorMsg` callback. Now, we
instead have a `fixupBuilderFailureErrorMessage` that is just used by
`DerivationBuildingGoal`, and `DerivationBuilder` just returns the raw
data needed by this.
2025-08-28 22:17:15 -04:00
Sergei Zimmerman
d59b959c87 libutil: Use std::source_location for unreachable
Make unreachable a function instead of a macro, since
C++20 provides a convenience class as a replacement for
older __FILE__, __LINE__ macros.
2025-08-29 00:21:07 +03:00
John Ericson
47cae1f72b Merge pull request #13850 from obsidiansystems/factor-out-drv-env-desugar
Factor out a new `DesugaredEnv` from `DerivationBuildingGoal`
2025-08-28 17:10:48 -04:00
Sergei Zimmerman
1f607b5def libutil: Try to call std::terminate for panic
We now have a terminate handler that prints a
stack trace, which is useful to have when encountering
an unreachable.
2025-08-29 00:02:13 +03:00
John Ericson
53c31c8b29 Factor out a new DesugaredEnv from DerivationBuildingGoal
Now we have better separation of the core logic --- an integral part of
the store layer spec even --- from the goal mechanism and other
minutiae.

Co-authored-by: Jeremy Kolb <kjeremy@gmail.com>
2025-08-28 16:45:45 -04:00
Sergei Zimmerman
731349639f Merge pull request #13524 from gmarti/fix_cacertificate
Add /etc/ssl/certs/ca-certificates.crt in docker.nix
2025-08-28 23:28:53 +03:00
John Ericson
f019f1b75a Merge pull request #13838 from NixOS/parse-url-path
Fix `ParsedURL` handling of `%2F` in URL paths
2025-08-28 16:07:25 -04:00
Jörg Thalheim
c436b7a32a Fix ParsedURL handling of %2F in URL paths
See the new extensive doxygen in `url.hh`.
This fixes fetching gitlab: flakes.

Paths are now stored as a std::vector of individual path
segments, which can themselves contain path separators '/' (%2F).
This is necessary to make the Gitlab's /projects/ API work.

Co-authored-by: John Ericson <John.Ericson@Obsidian.Systems>
Co-authored-by: Sergei Zimmerman <sergei@zimmerman.foo>
2025-08-28 22:20:04 +03:00
Sergei Zimmerman
6839f3de55 libutil-tests: Add more URL tests 2025-08-28 14:58:17 -04:00
John Ericson
3e0b1705c1 Move markContentsGood to after DerivationBuilder finishes
I think this should be fine for repairing. If anything, it is better,
because it would be weird to "mark and output good" only for it to then
fail output checks.
2025-08-28 14:54:11 -04:00
Sergei Zimmerman
c2782d7b84 Merge pull request #13853 from obsidiansystems/no-old-debugging-aid
Revert "Add a crude tracing mechansim for the build results"
2025-08-28 21:18:10 +03:00
John Ericson
bde745cb3f Move killChild call from ~DerivationBuildingGoal to ~DerivationBuilder
Sadly we cannot unexpose `DerivationBuilder::killChild` yet, because
`DerivationBuildingGoal` calls it elsewhere, but we can at least haave a
better division of labor between the two destructors.
2025-08-28 14:01:24 -04:00
John Ericson
c632c823ce Take DerivationBuilder::pid private 2025-08-28 14:01:20 -04:00
John Ericson
4388e3dcb5 Create DerivationBuilder::killChild
Then the derivation building goal doesn't need to snoop around as much.
2025-08-28 14:01:17 -04:00
John Ericson
49da508f46 Write a destructor for DerivationBuilderImpl
This allows `DerivationBuildingGoal` to know less.
2025-08-28 14:01:14 -04:00
John Ericson
557bbe969e Combine cleanupBuild and deleteTmpDir
It's hard to tell if I changed any behavior, but if I did, I think I
made it better, because now we explicitly move stuff out of the chroot
(if we were going to) before trying to delete the chroot.
2025-08-28 14:01:11 -04:00
John Ericson
4db6bf96b7 Give DerivationBuilderImpl::cleanupBuild bool arg
Do this to match `DerivationBuilder::deleteTmpDir`, which we'll want to
combine it with next.

Also chenge one caller from `deleteTmpDir(true)` to `cleanupBuild(true)`
now that this is done, because it will not make a difference.

This should be a pure refactor with no behavioral change.
2025-08-28 14:01:08 -04:00
John Ericson
8dd289099c Simplify DerivationGoal::unprepareBuild::diskFull
We only need it defined in the narrower scope
2025-08-28 14:01:05 -04:00
John Ericson
374f8e79a1 DerivationBuilderImpl::unprepareBuild Just throw error
Aftet the previous simplifications, there is no reason to catch the
error and immediately return it with a `std::variant` --- just let the
caller catch it instead.
2025-08-28 14:00:35 -04:00
John Ericson
0b85b023d8 Get rid of delayedException in DerivationBuilder
Instead of that funny business, the fixed output checks are not put in
`checkOutputs`, with the other (newer) output checks, where they also
better belong. The control flow is reworked (with comments!) so that
`checkOutputs` also runs in the `bmCheck` case.

Not only does this preserve existing behavior of `bmCheck`
double-checking fixed output hashes with less tricky code, it also makes
`bmCheck` better by also double-checking the other output checks, rather
than just assuming they pass if the derivation is deterministic.
2025-08-28 11:44:18 -04:00
John Ericson
ff961fd9e2 Get rid of DerivationBuilder::note*Mismatch
It's fine to set these worker flags a little later in the control flow,
since we'll be sure to reach those points in the error cases. And doing
that is much nicer than having these tangled callbacks.

I originally made the callbacks to meticulously recreate the exact
behavior which I didn't quite understand. Now, thanks to cleaning up the
error handling, I do understand what is going on, so I can be confident
that this change is safe to make.
2025-08-28 11:44:18 -04:00
Sergei Zimmerman
2eacb3c36f Merge pull request #13851 from lovesegfault/http-binary-cache-store-once
refactor(libstore/http-binary-cache-store): pragma once
2025-08-28 03:44:17 +03:00
John Ericson
169033001d Simplify handling of statuses for build errors
Instead of passing them around separately, or doing finicky logic in a
try-catch block to recover them, just make `BuildError` always contain a
status, and make it the thrower's responsibility to set it. This is much
more simple and explicit.

Once that change is done, split the `done` functions of `DerivationGoal`
and `DerivationBuildingGoal` into separate success and failure
functions, which ends up being easier to understand and hardly any
duplication.

Also, change the handling of failures in resolved cases to use
`BuildResult::DependencyFailed` and a new message. This is because the
underlying derivation will also get its message printed --- which is
good, because in general the resolved derivation is not unique. One dyn
drv test had to be updated, but CA (and dyn drv) is experimental, so I
do not mind.

Finally, delete `SubstError` because it is unused.
2025-08-27 20:05:06 -04:00
John Ericson
0590b13156 Revert "Add a crude tracing mechansim for the build results"
The commit says it was added for CA testing --- manual I assume, since
there is no use of this in the test suite. I don't think we need it any
more, and I am not sure whether it was ever supposed to have made it to
`master` either.

This reverts commit 2eec2f765a.
2025-08-27 19:36:02 -04:00
Bernardo Meurer Costa
241abcca86 refactor(libstore/http-binary-cache-store): pragma once 2025-08-27 21:13:59 +00:00
John Ericson
35978ca47b Merge pull request #13848 from obsidiansystems/factor-out-drv-check
Factor out `checkOutputs`
2025-08-27 16:50:38 -04:00
John Ericson
d1bdaef04e Factor out checkOutputs
We currently just use this during the build of a derivation, but there is no
reason we wouldn't want to use it elsewhere, e.g. to check the outputs
of someone else's build after the fact.

Moreover, I like pulling things out of `DerivationBuilder` that are
simple and don't need access to all that state. While
`DerivationBuilder` is unix-only, this refactor also make the code more
portable "for free".

The header is private, at Eelco's request.
2025-08-27 16:25:46 -04:00
John Ericson
6c8f5ef9f7 Merge pull request #13802 from obsidiansystems/post-build-hook-later
Move `runPostBuildHook` out of `DerivationBuilder`
2025-08-27 15:48:05 -04:00
John Ericson
193ad73ce2 Merge pull request #13808 from obsidiansystems/derivation-builder-kvm
Create `StringSet DerivationBuilderParams::systemFeatures`
2025-08-27 15:19:06 -04:00
John Ericson
f4a0161cb1 Create StringSet DerivationBuilderParams::systemFeatures
Do this to avoid checking "system features" from the store config
directly, because we rather not have `DerivationBuilder` depend on
`Store`.
2025-08-27 12:38:15 -04:00
John Ericson
79211b6110 Merge pull request #13846 from obsidiansystems/derivation-builder-params-aggregate-initialize
No more `DerivationBuilderParams:` constructor!
2025-08-27 12:30:49 -04:00
John Ericson
f5f9e32f54 No more DerivationBuilderParams: constructor!
I am not sure how/why this started working. C++23?
2025-08-27 11:40:02 -04:00
Jörg Thalheim
725a2f379f don't include derivation name in temporary build directories
With the migration to /nix/var/nix/builds we now have failing builds
when the derivation name is too long.
This change removes the derivation name from the temporary build to have
a predictable prefix length:

Also see: https://github.com/NixOS/infra/pull/764
for context.
2025-08-27 09:48:31 +02:00
Jörg Thalheim
564593bcb9 Merge pull request #13837 from xokdvium/bump-nixpkgs
flake: Update nixpkgs
2025-08-27 09:33:04 +02:00
Sergei Zimmerman
8ee74792fe Merge pull request #13819 from obsidiansystems/relative-url
Implement `parseURLRelative`, use in `HttpBinaryCacheStore`
2025-08-27 03:34:57 +03:00
John Ericson
e82210b3b2 Implement parseURLRelative, use in HttpBinaryCacheStore
This allows us to replace some very hacky and not correct string
concatentation in `HttpBinaryCacheStore`. It will especially be useful
with #13752, when today's hacks started to cause problems in practice,
not just theory.

Also make `fixGitURL` returned a `ParsedURL`.
2025-08-26 19:45:10 -04:00
Sergei Zimmerman
625477a7df flake: Update nixpkgs
• Updated input 'nixpkgs':
    'github:NixOS/nixpkgs/cd32a774ac52caaa03bcfc9e7591ac8c18617ced?narHash=sha256-VtMQg02B3kt1oejwwrGn50U9Xbjgzfbb5TV5Wtx8dKI%3D' (2025-08-17)
  → 'github:NixOS/nixpkgs/d98ce345cdab58477ca61855540999c86577d19d?narHash=sha256-O2CIn7HjZwEGqBrwu9EU76zlmA5dbmna7jL1XUmAId8%3D' (2025-08-26)

This update contains d1266642a8722f2a05e311fa151c1413d2b9653c, which
is necessary for the TOML timestamps to get tested via nixpkgsLibTests job.
2025-08-27 02:23:05 +03:00
Sergei Zimmerman
231f3af535 Merge pull request #13835 from obsidiansystems/better-string-split
Better `stringSplit`
2025-08-27 01:20:46 +03:00
John Ericson
cc4aa70e6e Better stringSplit
I need this for some `ParseURL` improvements, but I figure this is
better to send as its own PR.

I changed the tests willy-nilly to sometimes use
`std::list<std::string_view>` instead of `Strings` (which is
`std::list<std::string>`).

Co-Authored-By: Sergei Zimmerman <sergei@zimmerman.foo>
2025-08-26 18:03:23 -04:00
John Ericson
0bd9d6a28e Merge pull request #13832 from kip93/fix/empty-ports
Handle empty ports with new URL parsing
2025-08-26 13:55:59 -04:00
Leandro Reina
7989e3192d Handle empty ports 2025-08-26 17:41:27 +02:00
Robert Hensing
1e16a54ee5 Merge pull request #13828 from NixOS/readme-meeting-times
Update work meeting time in README
2025-08-26 01:37:07 +02:00
Robert Hensing
afade27123 Update work meeting time in README 2025-08-26 00:50:12 +02:00
John Ericson
0250d50df3 Move runPostBuildHook out of DerivationBuilder
It is suppposed to be "post build" not "during the build" after all. Its
location now matches that for the hook case (see elsewhere in
`DerivationdBuildingGoal`).

It was in a try-catch before, and now it isn't, but I believe that it is
impossible for it to throw `BuildError`, which is sufficient for this
code motion to be correct.
2025-08-25 18:29:24 -04:00
Robert Hensing
c1e2396d58 Merge pull request #13826 from xokdvium/sqlite-zfs-hack
SQLite: fsync db.sqlite-shm before opening the database
2025-08-26 00:05:40 +02:00
John Ericson
ca94905593 Merge pull request #13825 from obsidiansystems/slight-optimize-s3ToHttpsURL
`ParsedS3URL::toHttpsUrl` Slight optimize
2025-08-25 17:51:09 -04:00
Eelco Dolstra
e492c64c8e SQLite: fsync db.sqlite-shm before opening the database
This is a workaround for https://github.com/NixOS/nix/issues/13515
(opening the SQLite DB randomly taking a couple of seconds on ZFS).

(cherry picked from commit a7fceb5eec)
2025-08-26 00:42:18 +03:00
John Ericson
e4e8a615fa ParsedS3URL::toHttpsUrl Slight optimize
I didn't want to block that PR on further code review while I figured
out these new (to us) C++23 goodies.
2025-08-25 16:53:39 -04:00
John Ericson
fac34ad20f Merge pull request #13824 from xokdvium/fix-formatting
libexpr: Fix weird formatting after treewide reformat
2025-08-25 15:26:17 -04:00
John Ericson
024d3954af Merge pull request #13823 from lovesegfault/extract-s3ToHttpsURL
feat(libstore/s3): add toHttpsUrl
2025-08-25 15:11:38 -04:00
Sergei Zimmerman
f0e4af4365 libexpr: Fix weird formatting after treewide reformat 2025-08-25 22:09:18 +03:00
Bernardo Meurer Costa
5985d67906 feat(libstore/s3): add toHttpsUrl
This is extracted from the work in #13752
2025-08-25 18:48:19 +00:00
Eelco Dolstra
9bee0fa6ac Merge pull request #13822 from NixOS/bump-2.32.0
Bump version to 2.32
2025-08-25 17:22:57 +02:00
Eelco Dolstra
adec28bf85 Update release-process.md 2025-08-25 10:30:21 +02:00
Eelco Dolstra
f5e09d9b58 Update mergify.yml 2025-08-25 10:28:47 +02:00
Eelco Dolstra
f67daa4a87 Bump version 2025-08-25 10:27:46 +02:00
Ethan Evans
7b8ceb5d2d libutil, libexpr: #10542 abstract over getrusage for getting cpuTime stat and implement windows version
Update src/libutil/windows/current-process.cc

Prefer `nullptr` over `NULL`

Co-authored-by: Sergei Zimmerman <sergei@zimmerman.foo>

Update src/libutil/unix/current-process.cc

Prefer C++ type casts

Co-authored-by: Sergei Zimmerman <sergei@zimmerman.foo>

Update src/libutil/windows/current-process.cc

Prefer C++ type casts

Co-authored-by: Sergei Zimmerman <sergei@zimmerman.foo>

Update src/libutil/unix/current-process.cc

Don't allocate exception

Co-authored-by: Sergei Zimmerman <sergei@zimmerman.foo>
2025-08-24 18:45:33 -07:00
John Ericson
c9211b0b2d Merge pull request #13803 from obsidiansystems/more-parsed-urls
Make more URLs parsed, most notably `FileTransferRequest::url`
2025-08-23 10:54:39 -04:00
John Ericson
2fa2c0b09f Merge pull request #13812 from obsidiansystems/url-parse-leniency
Limit to lenient parsing of non-standard URLs only where needed
2025-08-23 10:53:59 -04:00
Jörg Thalheim
ebf1cf5227 Merge pull request #13807 from roberth/release-notes-todo
maintainers: Add script for release notes todo list
2025-08-23 08:52:22 +02:00
John Ericson
3e86d75c9d Make more URLs parsed, most notably FileTransferRequest::url
Trying to gradually replace the use of strings with better types in ways
that makes sense.
2025-08-22 12:42:48 -04:00
John Ericson
72a548ed6a Limit to lenient parsing of non-standard URLs only where needed
This allows us to put `parseURL` in more spots without furthering
technical debt.
2025-08-22 12:37:37 -04:00
John Ericson
4083eff0c0 decodeQuery Take std::string_view not string ref 2025-08-22 12:26:48 -04:00
Robert Hensing
a1b3934a78 maintainers: Add script for release notes todo list 2025-08-21 14:19:22 +02:00
Grégory marti
f0c7fbcdab Add /etc/ssl/certs/ca-certificates.crt in docker.nix 2025-07-22 17:39:29 +02:00
185 changed files with 4856 additions and 2452 deletions

View File

@@ -18,7 +18,7 @@ jobs:
runs-on: ubuntu-24.04
if: github.repository_owner == 'NixOS'
steps:
- uses: actions/labeler@v5
- uses: actions/labeler@v6
with:
repo-token: ${{ secrets.GITHUB_TOKEN }}
sync-labels: false

View File

@@ -161,3 +161,14 @@ pull_request_rules:
labels:
- automatic backport
- merge-queue
- name: backport patches to 2.31
conditions:
- label=backport 2.31-maintenance
actions:
backport:
branches:
- "2.31-maintenance"
labels:
- automatic backport
- merge-queue

View File

@@ -1 +1 @@
2.31.0
2.32.0

View File

@@ -0,0 +1,23 @@
---
synopsis: "C API: Errors returned from your primops are not treated as recoverable by default"
prs: [13930]
---
Nix 2.32 by default remembers the error in the thunk that triggered it.
Previously the following sequence of events worked:
1. Have a thunk that invokes a primop that's defined through the C API
2. The primop returns an error
3. Force the thunk again
4. The primop returns a value
5. The thunk evaluated successfully
**Resolution**
C API consumers that rely on this must change their recoverable error calls:
```diff
-nix_set_err_msg(context, NIX_ERR_*, msg);
+nix_set_err_msg(context, NIX_ERR_RECOVERABLE, msg);
```

View File

@@ -0,0 +1,6 @@
---
synopsis: "Removed support for daemons and clients older than Nix 2.0"
prs: [13951]
---
We have dropped support in the daemon worker protocol for daemons and clients that don't speak at least version 18 of the protocol. This first Nix release that supports this version is Nix 2.0, released in February 2018.

View File

@@ -0,0 +1,6 @@
---
synopsis: "Temporary build directories no longer include derivation names"
prs: [13839]
---
Temporary build directories created during derivation builds no longer include the derivation name in their path to avoid build failures when the derivation name is too long. This change ensures predictable prefix lengths for build directories under `/nix/var/nix/builds`.

View File

@@ -34,7 +34,7 @@ $ nix-shell --attr devShells.x86_64-linux.native-clangStdenvPackages
To build Nix itself in this shell:
```console
[nix-shell]$ mesonFlags+=" --prefix=$(pwd)/outputs/out"
[nix-shell]$ out="$(pwd)/outputs/out" dev=$out debug=$out mesonFlags+=" --prefix=${out}"
[nix-shell]$ dontAddPrefix=1 configurePhase
[nix-shell]$ buildPhase
```

View File

@@ -281,7 +281,10 @@ let
# may get replaced by pkgs.dockerTools.caCertificates
mkdir -p $out/etc/ssl/certs
# Old NixOS compatibility.
ln -s /nix/var/nix/profiles/default/etc/ssl/certs/ca-bundle.crt $out/etc/ssl/certs
# NixOS canonical location
ln -s /nix/var/nix/profiles/default/etc/ssl/certs/ca-bundle.crt $out/etc/ssl/certs/ca-certificates.crt
cat $passwdContentsPath > $out/etc/passwd
echo "" >> $out/etc/passwd

6
flake.lock generated
View File

@@ -63,11 +63,11 @@
},
"nixpkgs": {
"locked": {
"lastModified": 1755442223,
"narHash": "sha256-VtMQg02B3kt1oejwwrGn50U9Xbjgzfbb5TV5Wtx8dKI=",
"lastModified": 1756178832,
"narHash": "sha256-O2CIn7HjZwEGqBrwu9EU76zlmA5dbmna7jL1XUmAId8=",
"owner": "NixOS",
"repo": "nixpkgs",
"rev": "cd32a774ac52caaa03bcfc9e7591ac8c18617ced",
"rev": "d98ce345cdab58477ca61855540999c86577d19d",
"type": "github"
},
"original": {

View File

@@ -46,7 +46,7 @@ The team meets twice a week (times are denoted in the [Europe/Amsterdam](https:/
- mark it as draft if it is blocked on the contributor
- escalate it back to the team by moving it to To discuss, and leaving a comment as to why the issue needs to be discussed again.
- Work meeting: Mondays 14:00-16:00 Europe/Amsterdam see [calendar](https://calendar.google.com/calendar/u/0/embed?src=b9o52fobqjak8oq8lfkhg3t0qg@group.calendar.google.com).
- Work meeting: Mondays 18:00-20:00 Europe/Amsterdam; see [calendar](https://calendar.google.com/calendar/u/0/embed?src=b9o52fobqjak8oq8lfkhg3t0qg@group.calendar.google.com).
1. Code review on pull requests from [In review](#in-review).
2. Other chores and tasks.

58
maintainers/release-notes-todo Executable file
View File

@@ -0,0 +1,58 @@
#!/usr/bin/env bash
set -euo pipefail
# debug:
# set -x
START_REF="${1}"
END_REF="${2:-upstream/master}"
# Get the merge base
MERGE_BASE=$(git merge-base "$START_REF" "$END_REF")
unset START_REF
# Get date range
START_DATE=$(git show -s --format=%cI "$MERGE_BASE")
END_DATE=$(git show -s --format=%cI "$END_REF")
echo "Checking PRs merged between $START_DATE and $END_DATE" >&2
# Get all commits between merge base and HEAD
COMMITS=$(git rev-list "$MERGE_BASE..$END_REF")
# Convert to set for fast lookup
declare -A commit_set
for commit in $COMMITS; do
commit_set["$commit"]=1
done
# Get the current changelog
LOG_DONE="$(changelog-d doc/manual/rl-next)"
is_done(){
local nr="$1"
echo "$LOG_DONE" | grep -E "^- .*/pull/$nr)"
}
# Query merged PRs in date range
gh pr list \
--repo NixOS/nix \
--state merged \
--limit 1000 \
--json number,title,author,mergeCommit \
--search "merged:$START_DATE..$END_DATE" | \
jq -r '.[] | [.number, .mergeCommit.oid, .title, .author.login] | @tsv' | \
while IFS=$'\t' read -r pr_num merge_commit _title author; do
# Check if this PR's merge commit is in our branch
if [[ -n "${commit_set[$merge_commit]:-}" ]]; then
# Full detail, not suitable for comment due to mass ping and duplicate title
# echo "- #$pr_num $_title (@$author)"
echo "- #$pr_num ($author)"
if is_done "$pr_num"
then
echo " - [x] has note"
else
echo " - [ ] has note"
fi
echo " - [ ] skip"
fi
done

View File

@@ -24,6 +24,12 @@ release:
* In a checkout of the Nix repo, make sure you're on `master` and run
`git pull`.
* Compile a release notes to-do list by running
```console
$ ./maintainers/release-notes-todo PREV_RELEASE HEAD
```
* Compile the release notes by running
```console
@@ -127,6 +133,8 @@ release:
Commit and push this to the maintenance branch.
* Create a backport label.
* Bump the version of `master`:
```console
@@ -134,6 +142,7 @@ release:
$ git pull
$ NEW_VERSION=2.13.0
$ echo $NEW_VERSION > .version
$ ... edit .mergify.yml to add the previous version ...
$ git checkout -b bump-$NEW_VERSION
$ git commit -a -m 'Bump version'
$ git push --set-upstream origin bump-$NEW_VERSION
@@ -141,10 +150,6 @@ release:
Make a pull request and auto-merge it.
* Create a backport label.
* Add the new backport label to `.mergify.yml`.
* Post an [announcement on Discourse](https://discourse.nixos.org/c/announcements/8), including the contents of
`rl-$VERSION.md`.

View File

@@ -76,6 +76,16 @@ scope: {
prevAttrs.postInstall;
});
toml11 = pkgs.toml11.overrideAttrs rec {
version = "4.4.0";
src = pkgs.fetchFromGitHub {
owner = "ToruNiina";
repo = "toml11";
tag = "v${version}";
hash = "sha256-sgWKYxNT22nw376ttGsTdg0AMzOwp8QH3E8mx0BZJTQ=";
};
};
# TODO Hack until https://github.com/NixOS/nixpkgs/issues/45462 is fixed.
boost =
(pkgs.boost.override {

View File

@@ -105,8 +105,8 @@ DerivedPathsWithInfo InstallableFlake::toDerivedPaths()
std::optional<NixInt::Inner> priority;
if (attr->maybeGetAttr(state->sOutputSpecified)) {
} else if (auto aMeta = attr->maybeGetAttr(state->sMeta)) {
if (attr->maybeGetAttr(state->s.outputSpecified)) {
} else if (auto aMeta = attr->maybeGetAttr(state->s.meta)) {
if (auto aPriority = aMeta->maybeGetAttr("priority"))
priority = aPriority->getInt().value;
}
@@ -119,12 +119,12 @@ DerivedPathsWithInfo InstallableFlake::toDerivedPaths()
overloaded{
[&](const ExtendedOutputsSpec::Default & d) -> OutputsSpec {
StringSet outputsToInstall;
if (auto aOutputSpecified = attr->maybeGetAttr(state->sOutputSpecified)) {
if (auto aOutputSpecified = attr->maybeGetAttr(state->s.outputSpecified)) {
if (aOutputSpecified->getBool()) {
if (auto aOutputName = attr->maybeGetAttr("outputName"))
outputsToInstall = {aOutputName->getString()};
}
} else if (auto aMeta = attr->maybeGetAttr(state->sMeta)) {
} else if (auto aMeta = attr->maybeGetAttr(state->s.meta)) {
if (auto aOutputsToInstall = aMeta->maybeGetAttr("outputsToInstall"))
for (auto & s : aOutputsToInstall->getListOfStrings())
outputsToInstall.insert(s);

View File

@@ -178,10 +178,16 @@ MixFlakeOptions::MixFlakeOptions()
for (auto & [inputName, input] : flake.lockFile.root->inputs) {
auto input2 = flake.lockFile.findInput({inputName}); // resolve 'follows' nodes
if (auto input3 = std::dynamic_pointer_cast<const flake::LockedNode>(input2)) {
fetchers::Attrs extraAttrs;
if (!input3->lockedRef.subdir.empty()) {
extraAttrs["dir"] = input3->lockedRef.subdir;
}
overrideRegistry(
fetchers::Input::fromAttrs(fetchSettings, {{"type", "indirect"}, {"id", inputName}}),
input3->lockedRef.input,
{});
extraAttrs);
}
}
}},

View File

@@ -40,6 +40,8 @@ static T * unsafe_new_with_self(F && init)
return new (p) T(init(static_cast<T *>(p)));
}
extern "C" {
nix_err nix_libexpr_init(nix_c_context * context)
{
if (context)
@@ -287,3 +289,5 @@ void nix_gc_register_finalizer(void * obj, void * cd, void (*finalizer)(void * o
GC_REGISTER_FINALIZER(obj, finalizer, cd, 0, 0);
#endif
}
} // extern "C"

View File

@@ -8,6 +8,8 @@
#include "nix_api_value.h"
#include "nix/expr/search-path.hh"
extern "C" {
struct nix_eval_state_builder
{
nix::ref<nix::Store> store;
@@ -61,4 +63,6 @@ struct nix_realised_string
std::vector<StorePath> storePaths;
};
} // extern "C"
#endif // NIX_API_EXPR_INTERNAL_H

View File

@@ -14,6 +14,8 @@
#include <nlohmann/json.hpp>
extern "C" {
void nix_set_string_return(nix_string_return * str, const char * c)
{
str->str = c;
@@ -40,6 +42,8 @@ nix_err nix_external_add_string_context(nix_c_context * context, nix_string_cont
NIXC_CATCH_ERRS
}
} // extern "C"
class NixCExternalValue : public nix::ExternalValueBase
{
NixCExternalValueDesc & desc;
@@ -170,6 +174,8 @@ public:
virtual ~NixCExternalValue() override {};
};
extern "C" {
ExternalValue * nix_create_external_value(nix_c_context * context, NixCExternalValueDesc * desc, void * v)
{
if (context)
@@ -198,3 +204,5 @@ void * nix_get_external_value_content(nix_c_context * context, ExternalValue * b
}
NIXC_CATCH_ERRS_NULL
}
} // extern "C"

View File

@@ -1,4 +1,5 @@
#include "nix/expr/attr-set.hh"
#include "nix/expr/eval-error.hh"
#include "nix/util/configuration.hh"
#include "nix/expr/eval.hh"
#include "nix/store/globals.hh"
@@ -89,8 +90,13 @@ static void nix_c_primop_wrapper(
f(userdata, &ctx, (EvalState *) &state, (nix_value **) args, (nix_value *) &vTmp);
if (ctx.last_err_code != NIX_OK) {
/* TODO: Throw different errors depending on the error code */
state.error<nix::EvalError>("Error from custom function: %s", *ctx.last_err).atPos(pos).debugThrow();
if (ctx.last_err_code == NIX_ERR_RECOVERABLE) {
state.error<nix::RecoverableEvalError>("Recoverable error from custom function: %s", *ctx.last_err)
.atPos(pos)
.debugThrow();
} else {
state.error<nix::EvalError>("Error from custom function: %s", *ctx.last_err).atPos(pos).debugThrow();
}
}
if (!vTmp.isValid()) {
@@ -111,6 +117,8 @@ static void nix_c_primop_wrapper(
v = vTmp;
}
extern "C" {
PrimOp * nix_alloc_primop(
nix_c_context * context,
PrimOpFun fun,
@@ -175,6 +183,8 @@ ValueType nix_get_type(nix_c_context * context, const nix_value * value)
switch (v.type()) {
case nThunk:
return NIX_TYPE_THUNK;
case nFailed:
return NIX_TYPE_FAILED;
case nInt:
return NIX_TYPE_INT;
case nFloat:
@@ -592,7 +602,7 @@ nix_err nix_bindings_builder_insert(nix_c_context * context, BindingsBuilder * b
context->last_err_code = NIX_OK;
try {
auto & v = check_value_not_null(value);
nix::Symbol s = bb->builder.state.symbols.create(name);
nix::Symbol s = bb->builder.state.get().symbols.create(name);
bb->builder.insert(s, &v);
}
NIXC_CATCH_ERRS
@@ -651,3 +661,5 @@ const StorePath * nix_realised_string_get_store_path(nix_realised_string * s, si
{
return &s->storePaths[i];
}
} // extern "C"

View File

@@ -32,7 +32,8 @@ typedef enum {
NIX_TYPE_ATTRS,
NIX_TYPE_LIST,
NIX_TYPE_FUNCTION,
NIX_TYPE_EXTERNAL
NIX_TYPE_EXTERNAL,
NIX_TYPE_FAILED,
} ValueType;
// forward declarations

View File

@@ -54,7 +54,7 @@ TEST_F(JSONValueTest, IntNegative)
TEST_F(JSONValueTest, String)
{
Value v;
v.mkString("test");
v.mkStringNoCopy("test");
ASSERT_EQ(getJSONValue(v), "\"test\"");
}
@@ -62,7 +62,7 @@ TEST_F(JSONValueTest, StringQuotes)
{
Value v;
v.mkString("test\"");
v.mkStringNoCopy("test\"");
ASSERT_EQ(getJSONValue(v), "\"test\\\"\"");
}

View File

@@ -55,6 +55,7 @@ sources = files(
'nix_api_expr.cc',
'nix_api_external.cc',
'nix_api_value.cc',
'nix_api_value_internal.cc',
'primops.cc',
'search-path.cc',
'trivial.cc',

View File

@@ -1,7 +1,5 @@
#include "nix_api_store.h"
#include "nix_api_store_internal.h"
#include "nix_api_util.h"
#include "nix_api_util_internal.h"
#include "nix_api_expr.h"
#include "nix_api_value.h"
@@ -151,8 +149,8 @@ TEST_F(nix_api_expr_test, nix_expr_realise_context_bad_value)
assert_ctx_ok();
auto r = nix_string_realise(ctx, state, value, false);
ASSERT_EQ(nullptr, r);
ASSERT_EQ(ctx->last_err_code, NIX_ERR_NIX_ERROR);
ASSERT_THAT(ctx->last_err, testing::Optional(testing::HasSubstr("cannot coerce")));
ASSERT_EQ(nix_err_code(ctx), NIX_ERR_NIX_ERROR);
ASSERT_THAT(nix_err_msg(nullptr, ctx, nullptr), testing::HasSubstr("cannot coerce"));
}
TEST_F(nix_api_expr_test, nix_expr_realise_context_bad_build)
@@ -168,8 +166,8 @@ TEST_F(nix_api_expr_test, nix_expr_realise_context_bad_build)
assert_ctx_ok();
auto r = nix_string_realise(ctx, state, value, false);
ASSERT_EQ(nullptr, r);
ASSERT_EQ(ctx->last_err_code, NIX_ERR_NIX_ERROR);
ASSERT_THAT(ctx->last_err, testing::Optional(testing::HasSubstr("failed with exit code 1")));
ASSERT_EQ(nix_err_code(ctx), NIX_ERR_NIX_ERROR);
ASSERT_THAT(nix_err_msg(nullptr, ctx, nullptr), testing::HasSubstr("failed with exit code 1"));
}
TEST_F(nix_api_expr_test, nix_expr_realise_context)
@@ -381,12 +379,11 @@ TEST_F(nix_api_expr_test, nix_expr_primop_bad_no_return)
nix_value * result = nix_alloc_value(ctx, state);
assert_ctx_ok();
nix_value_call(ctx, state, primopValue, three, result);
ASSERT_EQ(ctx->last_err_code, NIX_ERR_NIX_ERROR);
ASSERT_EQ(nix_err_code(ctx), NIX_ERR_NIX_ERROR);
ASSERT_THAT(
ctx->last_err,
testing::Optional(
testing::HasSubstr("Implementation error in custom function: return value was not initialized")));
ASSERT_THAT(ctx->last_err, testing::Optional(testing::HasSubstr("badNoReturn")));
nix_err_msg(nullptr, ctx, nullptr),
testing::HasSubstr("Implementation error in custom function: return value was not initialized"));
ASSERT_THAT(nix_err_msg(nullptr, ctx, nullptr), testing::HasSubstr("badNoReturn"));
}
static void primop_bad_return_thunk(
@@ -419,12 +416,11 @@ TEST_F(nix_api_expr_test, nix_expr_primop_bad_return_thunk)
assert_ctx_ok();
NIX_VALUE_CALL(ctx, state, result, primopValue, toString, four);
ASSERT_EQ(ctx->last_err_code, NIX_ERR_NIX_ERROR);
ASSERT_EQ(nix_err_code(ctx), NIX_ERR_NIX_ERROR);
ASSERT_THAT(
ctx->last_err,
testing::Optional(
testing::HasSubstr("Implementation error in custom function: return value must not be a thunk")));
ASSERT_THAT(ctx->last_err, testing::Optional(testing::HasSubstr("badReturnThunk")));
nix_err_msg(nullptr, ctx, nullptr),
testing::HasSubstr("Implementation error in custom function: return value must not be a thunk"));
ASSERT_THAT(nix_err_msg(nullptr, ctx, nullptr), testing::HasSubstr("badReturnThunk"));
}
TEST_F(nix_api_expr_test, nix_value_call_multi_no_args)
@@ -441,4 +437,105 @@ TEST_F(nix_api_expr_test, nix_value_call_multi_no_args)
assert_ctx_ok();
ASSERT_EQ(3, rInt);
}
// The following is a test case for retryable thunks.
// This is a requirement for the current way in which NixOps4 evaluates its deployment expressions.
// An alternative strategy could be implemented, but unwinding the stack may be a more efficient way to deal with many
// suspensions/resumptions, compared to e.g. using a thread or coroutine stack for each suspended dependency. This test
// models the essential bits of a deployment tool that uses such a strategy.
// State for the retryable primop - simulates deployment resource availability
struct DeploymentResourceState
{
bool vm_created = false;
};
static void primop_load_resource_input(
void * user_data, nix_c_context * context, EvalState * state, nix_value ** args, nix_value * ret)
{
assert(context);
assert(state);
auto * resource_state = static_cast<DeploymentResourceState *>(user_data);
// Get the resource input name argument
std::string input_name;
if (nix_get_string(context, args[0], OBSERVE_STRING(input_name)) != NIX_OK)
return;
// Only handle "vm_id" input - throw for anything else
if (input_name != "vm_id") {
std::string error_msg = "unknown resource input: " + input_name;
nix_set_err_msg(context, NIX_ERR_NIX_ERROR, error_msg.c_str());
return;
}
if (resource_state->vm_created) {
// VM has been created, return the ID
nix_init_string(context, ret, "vm-12345");
} else {
// VM not created yet, fail with dependency error
nix_set_err_msg(context, NIX_ERR_RECOVERABLE, "VM not yet created");
}
}
TEST_F(nix_api_expr_test, nix_expr_thunk_re_evaluation_after_deployment)
{
// This test demonstrates NixOps4's requirement: a thunk calling a primop should be
// re-evaluable when deployment resources become available that were not available initially.
DeploymentResourceState resource_state;
PrimOp * primop = nix_alloc_primop(
ctx,
primop_load_resource_input,
1,
"loadResourceInput",
nullptr,
"load a deployment resource input",
&resource_state);
assert_ctx_ok();
nix_value * primopValue = nix_alloc_value(ctx, state);
assert_ctx_ok();
nix_init_primop(ctx, primopValue, primop);
assert_ctx_ok();
nix_value * inputName = nix_alloc_value(ctx, state);
assert_ctx_ok();
nix_init_string(ctx, inputName, "vm_id");
assert_ctx_ok();
// Create a single thunk by using nix_init_apply instead of nix_value_call
// This creates a lazy application that can be forced multiple times
nix_value * thunk = nix_alloc_value(ctx, state);
assert_ctx_ok();
nix_init_apply(ctx, thunk, primopValue, inputName);
assert_ctx_ok();
// First force: VM not created yet, should fail
nix_value_force(ctx, state, thunk);
ASSERT_EQ(NIX_ERR_NIX_ERROR, nix_err_code(ctx));
ASSERT_THAT(nix_err_msg(nullptr, ctx, nullptr), testing::HasSubstr("VM not yet created"));
// Clear the error context for the next attempt
nix_c_context_free(ctx);
ctx = nix_c_context_create();
// Simulate deployment process: VM gets created
resource_state.vm_created = true;
// Second force of the SAME thunk: this is where the "failed" value issue appears
// With failed value caching, this should fail because the thunk is marked as permanently failed
// Without failed value caching (or with retryable failures), this should succeed
nix_value_force(ctx, state, thunk);
// If we get here without error, the thunk was successfully re-evaluated
assert_ctx_ok();
std::string result;
nix_get_string(ctx, thunk, OBSERVE_STRING(result));
assert_ctx_ok();
ASSERT_STREQ("vm-12345", result.c_str());
}
} // namespace nixC

View File

@@ -1,9 +1,6 @@
#include "nix_api_store.h"
#include "nix_api_store_internal.h"
#include "nix_api_util.h"
#include "nix_api_util_internal.h"
#include "nix_api_expr.h"
#include "nix_api_expr_internal.h"
#include "nix_api_value.h"
#include "nix_api_external.h"
@@ -39,7 +36,7 @@ private:
std::string type_string = "nix-external<MyExternalValueDesc( ";
type_string += std::to_string(obj->_x);
type_string += " )>";
res->str = &*type_string.begin();
nix_set_string_return(res, &*type_string.begin());
}
};

View File

@@ -1,10 +1,7 @@
#include "nix_api_store.h"
#include "nix_api_store_internal.h"
#include "nix_api_util.h"
#include "nix_api_util_internal.h"
#include "nix_api_expr.h"
#include "nix_api_value.h"
#include "nix_api_expr_internal.h"
#include "nix/expr/tests/nix_api_expr.hh"
#include "nix/util/tests/string_callback.hh"
@@ -16,14 +13,6 @@
namespace nixC {
TEST_F(nix_api_expr_test, as_nix_value_ptr)
{
// nix_alloc_value casts nix::Value to nix_value
// It should be obvious from the decl that that works, but if it doesn't,
// the whole implementation would be utterly broken.
ASSERT_EQ(sizeof(nix::Value), sizeof(nix_value));
}
TEST_F(nix_api_expr_test, nix_value_get_int_invalid)
{
ASSERT_EQ(0, nix_get_int(ctx, nullptr));
@@ -320,8 +309,10 @@ TEST_F(nix_api_expr_test, nix_value_init_apply_error)
// Evaluate it
nix_value_force(ctx, state, v);
ASSERT_EQ(ctx->last_err_code, NIX_ERR_NIX_ERROR);
ASSERT_THAT(ctx->last_err.value(), testing::HasSubstr("attempt to call something which is not a function but"));
ASSERT_EQ(nix_err_code(ctx), NIX_ERR_NIX_ERROR);
ASSERT_THAT(
nix_err_msg(nullptr, ctx, nullptr),
testing::HasSubstr("attempt to call something which is not a function but"));
// Clean up
nix_gc_decref(ctx, some_string);
@@ -380,7 +371,9 @@ TEST_F(nix_api_expr_test, nix_value_init_apply_lazy_arg)
// nix_get_attr_byname isn't lazy (it could have been) so it will throw the exception
nix_value * foo = nix_get_attr_byname(ctx, r, state, "foo");
ASSERT_EQ(nullptr, foo);
ASSERT_THAT(ctx->last_err.value(), testing::HasSubstr("error message for test case nix_value_init_apply_lazy_arg"));
ASSERT_THAT(
nix_err_msg(nullptr, ctx, nullptr),
testing::HasSubstr("error message for test case nix_value_init_apply_lazy_arg"));
// Clean up
nix_gc_decref(ctx, f);

View File

@@ -0,0 +1,25 @@
#include "nix_api_store.h"
#include "nix_api_util.h"
#include "nix_api_expr.h"
#include "nix_api_value.h"
#include "nix_api_expr_internal.h"
#include "nix/expr/tests/nix_api_expr.hh"
#include "nix/util/tests/string_callback.hh"
#include <gmock/gmock.h>
#include <cstddef>
#include <cstdlib>
#include <gtest/gtest.h>
namespace nixC {
TEST_F(nix_api_expr_test, as_nix_value_ptr)
{
// nix_alloc_value casts nix::Value to nix_value
// It should be obvious from the decl that that works, but if it doesn't,
// the whole implementation would be utterly broken.
ASSERT_EQ(sizeof(nix::Value), sizeof(nix_value));
}
} // namespace nixC

View File

@@ -35,14 +35,14 @@ TEST_F(ValuePrintingTests, tBool)
TEST_F(ValuePrintingTests, tString)
{
Value vString;
vString.mkString("some-string");
vString.mkStringNoCopy("some-string");
test(vString, "\"some-string\"");
}
TEST_F(ValuePrintingTests, tPath)
{
Value vPath;
vPath.mkString("/foo");
vPath.mkStringNoCopy("/foo");
test(vPath, "\"/foo\"");
}
@@ -61,7 +61,7 @@ TEST_F(ValuePrintingTests, tAttrs)
Value vTwo;
vTwo.mkInt(2);
BindingsBuilder builder(state, state.allocBindings(10));
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.symbols.create("one"), &vOne);
builder.insert(state.symbols.create("two"), &vTwo);
@@ -196,11 +196,11 @@ TEST_F(ValuePrintingTests, depthAttrs)
Value vTwo;
vTwo.mkInt(2);
BindingsBuilder builderEmpty(state, state.allocBindings(0));
BindingsBuilder builderEmpty = state.buildBindings(0);
Value vAttrsEmpty;
vAttrsEmpty.mkAttrs(builderEmpty.finish());
BindingsBuilder builder(state, state.allocBindings(10));
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.symbols.create("one"), &vOne);
builder.insert(state.symbols.create("two"), &vTwo);
builder.insert(state.symbols.create("nested"), &vAttrsEmpty);
@@ -208,7 +208,7 @@ TEST_F(ValuePrintingTests, depthAttrs)
Value vAttrs;
vAttrs.mkAttrs(builder.finish());
BindingsBuilder builder2(state, state.allocBindings(10));
BindingsBuilder builder2 = state.buildBindings(10);
builder2.insert(state.symbols.create("one"), &vOne);
builder2.insert(state.symbols.create("two"), &vTwo);
builder2.insert(state.symbols.create("nested"), &vAttrs);
@@ -233,14 +233,14 @@ TEST_F(ValuePrintingTests, depthList)
Value vTwo;
vTwo.mkInt(2);
BindingsBuilder builder(state, state.allocBindings(10));
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.symbols.create("one"), &vOne);
builder.insert(state.symbols.create("two"), &vTwo);
Value vAttrs;
vAttrs.mkAttrs(builder.finish());
BindingsBuilder builder2(state, state.allocBindings(10));
BindingsBuilder builder2 = state.buildBindings(10);
builder2.insert(state.symbols.create("one"), &vOne);
builder2.insert(state.symbols.create("two"), &vTwo);
builder2.insert(state.symbols.create("nested"), &vAttrs);
@@ -290,12 +290,12 @@ TEST_F(StringPrintingTests, maxLengthTruncation)
TEST_F(ValuePrintingTests, attrsTypeFirst)
{
Value vType;
vType.mkString("puppy");
vType.mkStringNoCopy("puppy");
Value vApple;
vApple.mkString("apple");
vApple.mkStringNoCopy("apple");
BindingsBuilder builder(state, state.allocBindings(10));
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.symbols.create("type"), &vType);
builder.insert(state.symbols.create("apple"), &vApple);
@@ -334,7 +334,7 @@ TEST_F(ValuePrintingTests, ansiColorsBool)
TEST_F(ValuePrintingTests, ansiColorsString)
{
Value v;
v.mkString("puppy");
v.mkStringNoCopy("puppy");
test(v, ANSI_MAGENTA "\"puppy\"" ANSI_NORMAL, PrintOptions{.ansiColors = true});
}
@@ -342,7 +342,7 @@ TEST_F(ValuePrintingTests, ansiColorsString)
TEST_F(ValuePrintingTests, ansiColorsStringElided)
{
Value v;
v.mkString("puppy");
v.mkStringNoCopy("puppy");
test(
v,
@@ -374,7 +374,7 @@ TEST_F(ValuePrintingTests, ansiColorsAttrs)
Value vTwo;
vTwo.mkInt(2);
BindingsBuilder builder(state, state.allocBindings(10));
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.symbols.create("one"), &vOne);
builder.insert(state.symbols.create("two"), &vTwo);
@@ -390,10 +390,10 @@ TEST_F(ValuePrintingTests, ansiColorsAttrs)
TEST_F(ValuePrintingTests, ansiColorsDerivation)
{
Value vDerivation;
vDerivation.mkString("derivation");
vDerivation.mkStringNoCopy("derivation");
BindingsBuilder builder(state, state.allocBindings(10));
builder.insert(state.sType, &vDerivation);
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.s.type, &vDerivation);
Value vAttrs;
vAttrs.mkAttrs(builder.finish());
@@ -413,7 +413,7 @@ TEST_F(ValuePrintingTests, ansiColorsError)
{
Value throw_ = state.getBuiltin("throw");
Value message;
message.mkString("uh oh!");
message.mkStringNoCopy("uh oh!");
Value vError;
vError.mkApp(&throw_, &message);
@@ -430,16 +430,16 @@ TEST_F(ValuePrintingTests, ansiColorsDerivationError)
{
Value throw_ = state.getBuiltin("throw");
Value message;
message.mkString("uh oh!");
message.mkStringNoCopy("uh oh!");
Value vError;
vError.mkApp(&throw_, &message);
Value vDerivation;
vDerivation.mkString("derivation");
vDerivation.mkStringNoCopy("derivation");
BindingsBuilder builder(state, state.allocBindings(10));
builder.insert(state.sType, &vDerivation);
builder.insert(state.sDrvPath, &vError);
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.s.type, &vDerivation);
builder.insert(state.s.drvPath, &vError);
Value vAttrs;
vAttrs.mkAttrs(builder.finish());
@@ -553,12 +553,12 @@ TEST_F(ValuePrintingTests, ansiColorsBlackhole)
TEST_F(ValuePrintingTests, ansiColorsAttrsRepeated)
{
BindingsBuilder emptyBuilder(state, state.allocBindings(1));
BindingsBuilder emptyBuilder = state.buildBindings(1);
Value vEmpty;
vEmpty.mkAttrs(emptyBuilder.finish());
BindingsBuilder builder(state, state.allocBindings(10));
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.symbols.create("a"), &vEmpty);
builder.insert(state.symbols.create("b"), &vEmpty);
@@ -570,7 +570,7 @@ TEST_F(ValuePrintingTests, ansiColorsAttrsRepeated)
TEST_F(ValuePrintingTests, ansiColorsListRepeated)
{
BindingsBuilder emptyBuilder(state, state.allocBindings(1));
BindingsBuilder emptyBuilder = state.buildBindings(1);
Value vEmpty;
vEmpty.mkAttrs(emptyBuilder.finish());
@@ -586,7 +586,7 @@ TEST_F(ValuePrintingTests, ansiColorsListRepeated)
TEST_F(ValuePrintingTests, listRepeated)
{
BindingsBuilder emptyBuilder(state, state.allocBindings(1));
BindingsBuilder emptyBuilder = state.buildBindings(1);
Value vEmpty;
vEmpty.mkAttrs(emptyBuilder.finish());
@@ -609,7 +609,7 @@ TEST_F(ValuePrintingTests, ansiColorsAttrsElided)
Value vTwo;
vTwo.mkInt(2);
BindingsBuilder builder(state, state.allocBindings(10));
BindingsBuilder builder = state.buildBindings(10);
builder.insert(state.symbols.create("one"), &vOne);
builder.insert(state.symbols.create("two"), &vTwo);
@@ -635,8 +635,6 @@ TEST_F(ValuePrintingTests, ansiColorsAttrsElided)
TEST_F(ValuePrintingTests, ansiColorsListElided)
{
BindingsBuilder emptyBuilder(state, state.allocBindings(1));
Value vOne;
vOne.mkInt(1);

View File

@@ -16,19 +16,19 @@ Bindings * EvalState::allocBindings(size_t capacity)
throw Error("attribute set of size %d is too big", capacity);
nrAttrsets++;
nrAttrsInAttrsets += capacity;
return new (allocBytes(sizeof(Bindings) + sizeof(Attr) * capacity)) Bindings((Bindings::size_t) capacity);
return new (allocBytes(sizeof(Bindings) + sizeof(Attr) * capacity)) Bindings();
}
Value & BindingsBuilder::alloc(Symbol name, PosIdx pos)
{
auto value = state.allocValue();
auto value = state.get().allocValue();
bindings->push_back(Attr(name, value, pos));
return *value;
}
Value & BindingsBuilder::alloc(std::string_view name, PosIdx pos)
{
return alloc(state.symbols.create(name), pos);
return alloc(state.get().symbols.create(name), pos);
}
void Bindings::sort()

View File

@@ -330,7 +330,7 @@ AttrCursor::AttrCursor(
AttrKey AttrCursor::getKey()
{
if (!parent)
return {0, root->state.sEpsilon};
return {0, root->state.s.epsilon};
if (!parent->first->cachedValue) {
parent->first->cachedValue = root->db->getAttr(parent->first->getKey());
assert(parent->first->cachedValue);
@@ -702,7 +702,7 @@ bool AttrCursor::isDerivation()
StorePath AttrCursor::forceDerivation()
{
auto aDrvPath = getAttr(root->state.sDrvPath);
auto aDrvPath = getAttr(root->state.s.drvPath);
auto drvPath = root->state.store->parseStorePath(aDrvPath->getString());
drvPath.requireDerivation();
if (!root->state.store->isValidPath(drvPath) && !settings.readOnlyMode) {

View File

@@ -113,5 +113,6 @@ template class EvalErrorBuilder<MissingArgumentError>;
template class EvalErrorBuilder<InfiniteRecursionError>;
template class EvalErrorBuilder<InvalidPathError>;
template class EvalErrorBuilder<IFDError>;
template class EvalErrorBuilder<RecoverableEvalError>;
} // namespace nix

View File

@@ -16,6 +16,7 @@
# endif
# include <gc/gc_allocator.h>
# include <gc/gc_tiny_fl.h> // For GC_GRANULE_BYTES
# include <boost/coroutine2/coroutine.hpp>
# include <boost/coroutine2/protected_fixedsize_stack.hpp>
@@ -23,6 +24,17 @@
#endif
/*
* Ensure that Boehm satisfies our alignment requirements. This is the default configuration [^]
* and this assertion should never break for any platform. Let's assert it just in case.
*
* This alignment is particularly useful to be able to use aligned
* load/store instructions for loading/writing Values.
*
* [^]: https://github.com/bdwgc/bdwgc/blob/54ac18ccbc5a833dd7edaff94a10ab9b65044d61/include/gc/gc_tiny_fl.h#L31-L33
*/
static_assert(sizeof(void *) * 2 == GC_GRANULE_BYTES, "Boehm GC must use GC_GRANULE_WORDS = 2");
namespace nix {
#if NIX_USE_BOEHMGC

View File

@@ -185,7 +185,7 @@ FrameInfo SampleStack::getPrimOpFrameInfo(const PrimOp & primOp, std::span<Value
/* Error context strings don't actually matter, since we ignore all eval errors. */
state.forceAttrs(*args[0], pos, "");
auto attrs = args[0]->attrs();
auto nameAttr = state.getAttr(state.sName, attrs, "");
auto nameAttr = state.getAttr(state.s.name, attrs, "");
auto drvName = std::string(state.forceStringNoCtx(*nameAttr->value, pos, ""));
return DerivationStrictFrameInfo{.callPos = pos, .drvName = std::move(drvName)};
} catch (...) {
@@ -211,7 +211,7 @@ FrameInfo SampleStack::getFrameInfoFromValueAndPos(const Value & v, std::span<Va
/* Resolve primOp eagerly. Must not hold on to a reference to a Value. */
return PrimOpFrameInfo{.expr = v.primOpAppPrimOp(), .callPos = pos};
else if (state.isFunctor(v)) {
const auto functor = v.attrs()->get(state.sFunctor);
const auto functor = v.attrs()->get(state.s.functor);
if (auto pos_ = posCache.lookup(pos); std::holds_alternative<std::monostate>(pos_.origin))
/* HACK: In case callsite position is unresolved. */
return FunctorFrameInfo{.pos = functor->pos};

View File

@@ -1,4 +1,5 @@
#include "nix/expr/eval.hh"
#include "nix/expr/eval-error.hh"
#include "nix/expr/eval-settings.hh"
#include "nix/expr/primops.hh"
#include "nix/expr/print-options.hh"
@@ -22,10 +23,12 @@
#include "nix/fetchers/fetch-to-store.hh"
#include "nix/fetchers/tarball.hh"
#include "nix/fetchers/input-cache.hh"
#include "nix/util/current-process.hh"
#include "parser-tab.hh"
#include <algorithm>
#include <exception>
#include <iostream>
#include <sstream>
#include <cstring>
@@ -38,10 +41,6 @@
#include <nlohmann/json.hpp>
#include <boost/container/small_vector.hpp>
#ifndef _WIN32 // TODO use portable implementation
# include <sys/resource.h>
#endif
#include "nix/util/strings-inline.hh"
using json = nlohmann::json;
@@ -127,6 +126,8 @@ std::string_view showType(ValueType type, bool withArticle)
return WA("a", "float");
case nThunk:
return WA("a", "thunk");
case nFailed:
return WA("a", "failure");
}
unreachable();
}
@@ -203,124 +204,65 @@ EvalState::EvalState(
std::shared_ptr<Store> buildStore)
: fetchSettings{fetchSettings}
, settings{settings}
, sWith(symbols.create("<with>"))
, sOutPath(symbols.create("outPath"))
, sDrvPath(symbols.create("drvPath"))
, sType(symbols.create("type"))
, sMeta(symbols.create("meta"))
, sName(symbols.create("name"))
, sValue(symbols.create("value"))
, sSystem(symbols.create("system"))
, sOverrides(symbols.create("__overrides"))
, sOutputs(symbols.create("outputs"))
, sOutputName(symbols.create("outputName"))
, sIgnoreNulls(symbols.create("__ignoreNulls"))
, sFile(symbols.create("file"))
, sLine(symbols.create("line"))
, sColumn(symbols.create("column"))
, sFunctor(symbols.create("__functor"))
, sToString(symbols.create("__toString"))
, sRight(symbols.create("right"))
, sWrong(symbols.create("wrong"))
, sStructuredAttrs(symbols.create("__structuredAttrs"))
, sJson(symbols.create("__json"))
, sAllowedReferences(symbols.create("allowedReferences"))
, sAllowedRequisites(symbols.create("allowedRequisites"))
, sDisallowedReferences(symbols.create("disallowedReferences"))
, sDisallowedRequisites(symbols.create("disallowedRequisites"))
, sMaxSize(symbols.create("maxSize"))
, sMaxClosureSize(symbols.create("maxClosureSize"))
, sBuilder(symbols.create("builder"))
, sArgs(symbols.create("args"))
, sContentAddressed(symbols.create("__contentAddressed"))
, sImpure(symbols.create("__impure"))
, sOutputHash(symbols.create("outputHash"))
, sOutputHashAlgo(symbols.create("outputHashAlgo"))
, sOutputHashMode(symbols.create("outputHashMode"))
, sRecurseForDerivations(symbols.create("recurseForDerivations"))
, sDescription(symbols.create("description"))
, sSelf(symbols.create("self"))
, sEpsilon(symbols.create(""))
, sStartSet(symbols.create("startSet"))
, sOperator(symbols.create("operator"))
, sKey(symbols.create("key"))
, sPath(symbols.create("path"))
, sPrefix(symbols.create("prefix"))
, sOutputSpecified(symbols.create("outputSpecified"))
, exprSymbols{
.sub = symbols.create("__sub"),
.lessThan = symbols.create("__lessThan"),
.mul = symbols.create("__mul"),
.div = symbols.create("__div"),
.or_ = symbols.create("or"),
.findFile = symbols.create("__findFile"),
.nixPath = symbols.create("__nixPath"),
.body = symbols.create("body"),
}
, symbols(StaticEvalSymbols::staticSymbolTable())
, repair(NoRepair)
, emptyBindings(0)
, storeFS(
makeMountedSourceAccessor(
{
{CanonPath::root, makeEmptySourceAccessor()},
/* In the pure eval case, we can simply require
valid paths. However, in the *impure* eval
case this gets in the way of the union
mechanism, because an invalid access in the
upper layer will *not* be caught by the union
source accessor, but instead abort the entire
lookup.
, emptyBindings(Bindings())
, storeFS(makeMountedSourceAccessor({
{CanonPath::root, makeEmptySourceAccessor()},
/* In the pure eval case, we can simply require
valid paths. However, in the *impure* eval
case this gets in the way of the union
mechanism, because an invalid access in the
upper layer will *not* be caught by the union
source accessor, but instead abort the entire
lookup.
This happens when the store dir in the
ambient file system has a path (e.g. because
another Nix store there), but the relocated
store does not.
This happens when the store dir in the
ambient file system has a path (e.g. because
another Nix store there), but the relocated
store does not.
TODO make the various source accessors doing
access control all throw the same type of
exception, and make union source accessor
catch it, so we don't need to do this hack.
*/
{CanonPath(store->storeDir), store->getFSAccessor(settings.pureEval)},
}))
, rootFS(
({
/* In pure eval mode, we provide a filesystem that only
contains the Nix store.
TODO make the various source accessors doing
access control all throw the same type of
exception, and make union source accessor
catch it, so we don't need to do this hack.
*/
{CanonPath(store->storeDir), store->getFSAccessor(settings.pureEval)},
}))
, rootFS(({
/* In pure eval mode, we provide a filesystem that only
contains the Nix store.
If we have a chroot store and pure eval is not enabled,
use a union accessor to make the chroot store available
at its logical location while still having the
underlying directory available. This is necessary for
instance if we're evaluating a file from the physical
/nix/store while using a chroot store. */
auto accessor = getFSSourceAccessor();
If we have a chroot store and pure eval is not enabled,
use a union accessor to make the chroot store available
at its logical location while still having the
underlying directory available. This is necessary for
instance if we're evaluating a file from the physical
/nix/store while using a chroot store. */
auto accessor = getFSSourceAccessor();
auto realStoreDir = dirOf(store->toRealPath(StorePath::dummy));
if (settings.pureEval || store->storeDir != realStoreDir) {
accessor = settings.pureEval
? storeFS
: makeUnionSourceAccessor({accessor, storeFS});
}
auto realStoreDir = dirOf(store->toRealPath(StorePath::dummy));
if (settings.pureEval || store->storeDir != realStoreDir) {
accessor = settings.pureEval ? storeFS : makeUnionSourceAccessor({accessor, storeFS});
}
/* Apply access control if needed. */
if (settings.restrictEval || settings.pureEval)
accessor = AllowListSourceAccessor::create(accessor, {}, {},
[&settings](const CanonPath & path) -> RestrictedPathError {
auto modeInformation = settings.pureEval
? "in pure evaluation mode (use '--impure' to override)"
: "in restricted mode";
throw RestrictedPathError("access to absolute path '%1%' is forbidden %2%", path, modeInformation);
});
/* Apply access control if needed. */
if (settings.restrictEval || settings.pureEval)
accessor = AllowListSourceAccessor::create(
accessor, {}, {}, [&settings](const CanonPath & path) -> RestrictedPathError {
auto modeInformation = settings.pureEval ? "in pure evaluation mode (use '--impure' to override)"
: "in restricted mode";
throw RestrictedPathError("access to absolute path '%1%' is forbidden %2%", path, modeInformation);
});
accessor;
}))
accessor;
}))
, corepkgsFS(make_ref<MemorySourceAccessor>())
, internalFS(make_ref<MemorySourceAccessor>())
, derivationInternal{corepkgsFS->addFile(
CanonPath("derivation-internal.nix"),
CanonPath("derivation-internal.nix"),
#include "primops/derivation.nix.gen.hh"
)}
)}
, store(store)
, buildStore(buildStore ? buildStore : store)
, inputCache(fetchers::InputCache::create())
@@ -351,10 +293,10 @@ EvalState::EvalState(
vNull.mkNull();
vTrue.mkBool(true);
vFalse.mkBool(false);
vStringRegular.mkString("regular");
vStringDirectory.mkString("directory");
vStringSymlink.mkString("symlink");
vStringUnknown.mkString("unknown");
vStringRegular.mkStringNoCopy("regular");
vStringDirectory.mkStringNoCopy("directory");
vStringSymlink.mkStringNoCopy("symlink");
vStringUnknown.mkStringNoCopy("unknown");
/* Construct the Nix expression search path. */
assert(lookupPath.elements.empty());
@@ -654,7 +596,7 @@ std::optional<EvalState::Doc> EvalState::getDoc(Value & v)
}
if (isFunctor(v)) {
try {
Value & functor = *v.attrs()->find(sFunctor)->value;
Value & functor = *v.attrs()->find(s.functor)->value;
Value * vp[] = {&v};
Value partiallyApplied;
// The first parameter is not user-provided, and may be
@@ -883,7 +825,7 @@ DebugTraceStacker::DebugTraceStacker(EvalState & evalState, DebugTrace t)
void Value::mkString(std::string_view s)
{
mkString(makeImmutableString(s));
mkStringNoCopy(makeImmutableString(s));
}
static const char ** encodeContext(const NixStringContext & context)
@@ -902,12 +844,12 @@ static const char ** encodeContext(const NixStringContext & context)
void Value::mkString(std::string_view s, const NixStringContext & context)
{
mkString(makeImmutableString(s), encodeContext(context));
mkStringNoCopy(makeImmutableString(s), encodeContext(context));
}
void Value::mkStringMove(const char * s, const NixStringContext & context)
{
mkString(s, encodeContext(context));
mkStringNoCopy(s, encodeContext(context));
}
void Value::mkPath(const SourcePath & path)
@@ -978,8 +920,8 @@ void EvalState::mkPos(Value & v, PosIdx p)
auto origin = positions.originOf(p);
if (auto path = std::get_if<SourcePath>(&origin)) {
auto attrs = buildBindings(3);
attrs.alloc(sFile).mkString(path->path.abs());
makePositionThunks(*this, p, attrs.alloc(sLine), attrs.alloc(sColumn));
attrs.alloc(s.file).mkString(path->path.abs());
makePositionThunks(*this, p, attrs.alloc(s.line), attrs.alloc(s.column));
v.mkAttrs(attrs);
} else
v.mkNull();
@@ -1245,7 +1187,7 @@ void ExprAttrs::eval(EvalState & state, Env & env, Value & v)
dynamicEnv = &env2;
Env * inheritEnv = inheritFromExprs ? buildInheritFromEnv(state, env2) : nullptr;
AttrDefs::iterator overrides = attrs.find(state.sOverrides);
AttrDefs::iterator overrides = attrs.find(state.s.overrides);
bool hasOverrides = overrides != attrs.end();
/* The recursive attributes are evaluated in the new
@@ -1277,7 +1219,7 @@ void ExprAttrs::eval(EvalState & state, Env & env, Value & v)
*vOverrides,
[&]() { return vOverrides->determinePos(noPos); },
"while evaluating the `__overrides` attribute");
bindings.grow(state.allocBindings(bindings.capacity() + vOverrides->attrs()->size()));
bindings.grow(state.buildBindings(bindings.capacity() + vOverrides->attrs()->size()));
for (auto & i : *vOverrides->attrs()) {
AttrDefs::iterator j = attrs.find(i.name);
if (j != attrs.end()) {
@@ -1717,7 +1659,7 @@ void EvalState::callFunction(Value & fun, std::span<Value *> args, Value & vRes,
}
}
else if (vCur.type() == nAttrs && (functor = vCur.attrs()->get(sFunctor))) {
else if (vCur.type() == nAttrs && (functor = vCur.attrs()->get(s.functor))) {
/* 'vCur' may be allocated on the stack of the calling
function, but for functors we may keep a reference, so
heap-allocate a copy and use that instead. */
@@ -1779,7 +1721,7 @@ void EvalState::autoCallFunction(const Bindings & args, Value & fun, Value & res
forceValue(fun, pos);
if (fun.type() == nAttrs) {
auto found = fun.attrs()->find(sFunctor);
auto found = fun.attrs()->find(s.functor);
if (found != fun.attrs()->end()) {
Value * v = allocValue();
callFunction(*found->value, fun, *v, pos);
@@ -2122,6 +2064,54 @@ void ExprBlackHole::eval(EvalState & state, [[maybe_unused]] Env & env, Value &
// always force this to be separate, otherwise forceValue may inline it and take
// a massive perf hit
[[gnu::noinline]]
void EvalState::handleEvalExceptionForThunk(Env * env, Expr * expr, Value & v, const PosIdx pos)
{
v.mkThunk(env, expr);
tryFixupBlackHolePos(v, pos);
auto e = std::current_exception();
Value * recovery = nullptr;
try {
std::rethrow_exception(e);
} catch (const RecoverableEvalError & e) {
recovery = allocValue();
} catch (...) {
}
if (recovery) {
recovery->mkThunk(env, expr);
}
v.mkFailed(e, recovery);
}
[[gnu::noinline]]
void EvalState::handleEvalExceptionForApp(Value & v)
{
auto e = std::current_exception();
Value * recovery = nullptr;
try {
std::rethrow_exception(e);
} catch (const RecoverableEvalError & e) {
recovery = allocValue();
} catch (...) {
}
if (recovery) {
*recovery = v;
}
v.mkFailed(e, recovery);
}
[[gnu::noinline]]
void EvalState::handleEvalFailed(Value & v, const PosIdx pos)
{
assert(v.isFailed());
if (auto recoveryValue = v.failed()->recoveryValue) {
v = *recoveryValue;
forceValue(v, pos);
} else {
std::rethrow_exception(v.failed()->ex);
}
}
void EvalState::tryFixupBlackHolePos(Value & v, PosIdx pos)
{
if (!v.isBlackhole())
@@ -2130,7 +2120,8 @@ void EvalState::tryFixupBlackHolePos(Value & v, PosIdx pos)
try {
std::rethrow_exception(e);
} catch (InfiniteRecursionError & e) {
e.atPos(positions[pos]);
if (!e.hasPos())
e.atPos(positions[pos]);
} catch (...) {
}
}
@@ -2241,7 +2232,7 @@ Bindings::const_iterator EvalState::getAttr(Symbol attrSym, const Bindings * att
bool EvalState::isFunctor(const Value & fun) const
{
return fun.type() == nAttrs && fun.attrs()->find(sFunctor) != fun.attrs()->end();
return fun.type() == nAttrs && fun.attrs()->find(s.functor) != fun.attrs()->end();
}
void EvalState::forceFunction(Value & v, const PosIdx pos, std::string_view errorCtx)
@@ -2310,7 +2301,7 @@ bool EvalState::isDerivation(Value & v)
{
if (v.type() != nAttrs)
return false;
auto i = v.attrs()->get(sType);
auto i = v.attrs()->get(s.type);
if (!i)
return false;
forceValue(*i->value, i->pos);
@@ -2322,7 +2313,7 @@ bool EvalState::isDerivation(Value & v)
std::optional<std::string>
EvalState::tryAttrsToString(const PosIdx pos, Value & v, NixStringContext & context, bool coerceMore, bool copyToStore)
{
auto i = v.attrs()->find(sToString);
auto i = v.attrs()->find(s.toString);
if (i != v.attrs()->end()) {
Value v1;
callFunction(*i->value, v, v1, pos);
@@ -2368,7 +2359,7 @@ BackedStringView EvalState::coerceToString(
auto maybeString = tryAttrsToString(pos, v, context, coerceMore, copyToStore);
if (maybeString)
return std::move(*maybeString);
auto i = v.attrs()->find(sOutPath);
auto i = v.attrs()->find(s.outPath);
if (i == v.attrs()->end()) {
error<TypeError>(
"cannot coerce %1% to a string: %2%", showType(v), ValuePrinter(*this, v, errorPrintOptions))
@@ -2475,7 +2466,7 @@ SourcePath EvalState::coerceToPath(const PosIdx pos, Value & v, NixStringContext
/* Similarly, handle __toString where the result may be a path
value. */
if (v.type() == nAttrs) {
auto i = v.attrs()->find(sToString);
auto i = v.attrs()->find(s.toString);
if (i != v.attrs()->end()) {
Value v1;
callFunction(*i->value, v, v1, pos);
@@ -2665,8 +2656,8 @@ void EvalState::assertEqValues(Value & v1, Value & v2, const PosIdx pos, std::st
case nAttrs: {
if (isDerivation(v1) && isDerivation(v2)) {
auto i = v1.attrs()->get(sOutPath);
auto j = v2.attrs()->get(sOutPath);
auto i = v1.attrs()->get(s.outPath);
auto j = v2.attrs()->get(s.outPath);
if (i && j) {
try {
assertEqValues(*i->value, *j->value, pos, errorCtx);
@@ -2755,8 +2746,11 @@ void EvalState::assertEqValues(Value & v1, Value & v2, const PosIdx pos, std::st
}
return;
case nThunk: // Must not be left by forceValue
assert(false);
// Cannot be returned by forceValue().
case nThunk:
case nFailed:
unreachable();
default: // Note that we pass compiler flags that should make `default:` unreachable.
// Also note that this probably ran after `eqValues`, which implements
// the same logic more efficiently (without having to unwind stacks),
@@ -2819,8 +2813,8 @@ bool EvalState::eqValues(Value & v1, Value & v2, const PosIdx pos, std::string_v
/* If both sets denote a derivation (type = "derivation"),
then compare their outPaths. */
if (isDerivation(v1) && isDerivation(v2)) {
auto i = v1.attrs()->get(sOutPath);
auto j = v2.attrs()->get(sOutPath);
auto i = v1.attrs()->get(s.outPath);
auto j = v2.attrs()->get(s.outPath);
if (i && j)
return eqValues(*i->value, *j->value, pos, errorCtx);
}
@@ -2848,8 +2842,11 @@ bool EvalState::eqValues(Value & v1, Value & v2, const PosIdx pos, std::string_v
// !!!
return v1.fpoint() == v2.fpoint();
case nThunk: // Must not be left by forceValue
assert(false);
// Cannot be returned by forceValue().
case nThunk:
case nFailed:
unreachable();
default: // Note that we pass compiler flags that should make `default:` unreachable.
error<EvalError>("eqValues: cannot compare %1% with %2%", showType(v1), showType(v2))
.withTrace(pos, errorCtx)
@@ -2889,11 +2886,8 @@ void EvalState::maybePrintStats()
void EvalState::printStatistics()
{
#ifndef _WIN32 // TODO use portable implementation
struct rusage buf;
getrusage(RUSAGE_SELF, &buf);
float cpuTime = buf.ru_utime.tv_sec + ((float) buf.ru_utime.tv_usec / 1000000);
#endif
std::chrono::microseconds cpuTimeDuration = getCpuUserTime();
float cpuTime = std::chrono::duration_cast<std::chrono::duration<float>>(cpuTimeDuration).count();
uint64_t bEnvs = nrEnvs * sizeof(Env) + nrValuesInEnvs * sizeof(Value *);
uint64_t bLists = nrListElems * sizeof(Value *);
@@ -2915,18 +2909,12 @@ void EvalState::printStatistics()
if (outPath != "-")
fs.open(outPath, std::fstream::out);
json topObj = json::object();
#ifndef _WIN32 // TODO implement
topObj["cpuTime"] = cpuTime;
#endif
topObj["time"] = {
#ifndef _WIN32 // TODO implement
{"cpu", cpuTime},
#endif
#if NIX_USE_BOEHMGC
{GC_is_incremental_mode() ? "gcNonIncremental" : "gc", gcFullOnlyTime},
# ifndef _WIN32 // TODO implement
{GC_is_incremental_mode() ? "gcNonIncrementalFraction" : "gcFraction", gcFullOnlyTime / cpuTime},
# endif
#endif
};
topObj["envs"] = {
@@ -3196,8 +3184,7 @@ Expr * EvalState::parse(
docComments = &it->second;
}
auto result = parseExprFromBuf(
text, length, origin, basePath, symbols, settings, positions, *docComments, rootFS, exprSymbols);
auto result = parseExprFromBuf(text, length, origin, basePath, symbols, settings, positions, *docComments, rootFS);
result->bindVars(*this, staticEnv);

View File

@@ -45,7 +45,7 @@ PackageInfo::PackageInfo(EvalState & state, ref<Store> store, const std::string
std::string PackageInfo::queryName() const
{
if (name == "" && attrs) {
auto i = attrs->find(state->sName);
auto i = attrs->find(state->s.name);
if (i == attrs->end())
state->error<TypeError>("derivation name missing").debugThrow();
name = state->forceStringNoCtx(*i->value, noPos, "while evaluating the 'name' attribute of a derivation");
@@ -56,7 +56,7 @@ std::string PackageInfo::queryName() const
std::string PackageInfo::querySystem() const
{
if (system == "" && attrs) {
auto i = attrs->find(state->sSystem);
auto i = attrs->find(state->s.system);
system =
i == attrs->end()
? "unknown"
@@ -68,7 +68,7 @@ std::string PackageInfo::querySystem() const
std::optional<StorePath> PackageInfo::queryDrvPath() const
{
if (!drvPath && attrs) {
if (auto i = attrs->get(state->sDrvPath)) {
if (auto i = attrs->get(state->s.drvPath)) {
NixStringContext context;
auto found = state->coerceToStorePath(
i->pos, *i->value, context, "while evaluating the 'drvPath' attribute of a derivation");
@@ -95,7 +95,7 @@ StorePath PackageInfo::requireDrvPath() const
StorePath PackageInfo::queryOutPath() const
{
if (!outPath && attrs) {
auto i = attrs->find(state->sOutPath);
auto i = attrs->find(state->s.outPath);
NixStringContext context;
if (i != attrs->end())
outPath = state->coerceToStorePath(
@@ -111,7 +111,7 @@ PackageInfo::Outputs PackageInfo::queryOutputs(bool withPaths, bool onlyOutputsT
if (outputs.empty()) {
/* Get the outputs list. */
const Attr * i;
if (attrs && (i = attrs->get(state->sOutputs))) {
if (attrs && (i = attrs->get(state->s.outputs))) {
state->forceList(*i->value, i->pos, "while evaluating the 'outputs' attribute of a derivation");
/* For each output... */
@@ -127,7 +127,7 @@ PackageInfo::Outputs PackageInfo::queryOutputs(bool withPaths, bool onlyOutputsT
state->forceAttrs(*out->value, i->pos, "while evaluating an output of a derivation");
/* And evaluate its outPath attribute. */
auto outPath = out->value->attrs()->get(state->sOutPath);
auto outPath = out->value->attrs()->get(state->s.outPath);
if (!outPath)
continue; // FIXME: throw error?
NixStringContext context;
@@ -146,7 +146,7 @@ PackageInfo::Outputs PackageInfo::queryOutputs(bool withPaths, bool onlyOutputsT
return outputs;
const Attr * i;
if (attrs && (i = attrs->get(state->sOutputSpecified))
if (attrs && (i = attrs->get(state->s.outputSpecified))
&& state->forceBool(*i->value, i->pos, "while evaluating the 'outputSpecified' attribute of a derivation")) {
Outputs result;
auto out = outputs.find(queryOutputName());
@@ -181,7 +181,7 @@ PackageInfo::Outputs PackageInfo::queryOutputs(bool withPaths, bool onlyOutputsT
std::string PackageInfo::queryOutputName() const
{
if (outputName == "" && attrs) {
auto i = attrs->get(state->sOutputName);
auto i = attrs->get(state->s.outputName);
outputName =
i ? state->forceStringNoCtx(*i->value, noPos, "while evaluating the output name of a derivation") : "";
}
@@ -194,7 +194,7 @@ const Bindings * PackageInfo::getMeta()
return meta;
if (!attrs)
return 0;
auto a = attrs->get(state->sMeta);
auto a = attrs->get(state->s.meta);
if (!a)
return 0;
state->forceAttrs(*a->value, a->pos, "while evaluating the 'meta' attribute of a derivation");
@@ -221,7 +221,7 @@ bool PackageInfo::checkMeta(Value & v)
return false;
return true;
} else if (v.type() == nAttrs) {
if (v.attrs()->get(state->sOutPath))
if (v.attrs()->get(state->s.outPath))
return false;
for (auto & i : *v.attrs())
if (!checkMeta(*i.value))
@@ -411,7 +411,7 @@ static void getDerivations(
should we recurse into it? => Only if it has a
`recurseForDerivations = true' attribute. */
if (i->value->type() == nAttrs) {
auto j = i->value->attrs()->get(state.sRecurseForDerivations);
auto j = i->value->attrs()->get(state.s.recurseForDerivations);
if (j
&& state.forceBool(
*j->value, j->pos, "while evaluating the attribute `recurseForDerivations`"))

View File

@@ -5,6 +5,7 @@
#include "nix/expr/symbol-table.hh"
#include <algorithm>
#include <functional>
namespace nix {
@@ -54,16 +55,14 @@ public:
PosIdx pos;
private:
size_t size_, capacity_;
size_t size_ = 0;
Attr attrs[0];
Bindings(size_t capacity)
: size_(0)
, capacity_(capacity)
{
}
Bindings(const Bindings & bindings) = delete;
Bindings() = default;
Bindings(const Bindings &) = delete;
Bindings(Bindings &&) = delete;
Bindings & operator=(const Bindings &) = delete;
Bindings & operator=(Bindings &&) = delete;
public:
size_t size() const
@@ -82,7 +81,6 @@ public:
void push_back(const Attr & attr)
{
assert(size_ < capacity_);
attrs[size_++] = attr;
}
@@ -136,11 +134,6 @@ public:
void sort();
size_t capacity() const
{
return capacity_;
}
/**
* Returns the attributes in lexicographically sorted order.
*/
@@ -165,22 +158,29 @@ public:
* order at the end. The only way to consume a BindingsBuilder is to
* call finish(), which sorts the bindings.
*/
class BindingsBuilder
class BindingsBuilder final
{
Bindings * bindings;
public:
// needed by std::back_inserter
using value_type = Attr;
using size_type = Bindings::size_t;
EvalState & state;
private:
Bindings * bindings;
Bindings::size_t capacity_;
BindingsBuilder(EvalState & state, Bindings * bindings)
friend class EvalState;
BindingsBuilder(EvalState & state, Bindings * bindings, size_type capacity)
: bindings(bindings)
, capacity_(capacity)
, state(state)
{
}
public:
std::reference_wrapper<EvalState> state;
void insert(Symbol name, Value * value, PosIdx pos = noPos)
{
insert(Attr(name, value, pos));
@@ -193,6 +193,7 @@ public:
void push_back(const Attr & attr)
{
assert(bindings->size() < capacity_);
bindings->push_back(attr);
}
@@ -211,16 +212,16 @@ public:
return bindings;
}
size_t capacity()
size_t capacity() const noexcept
{
return bindings->capacity();
return capacity_;
}
void grow(Bindings * newBindings)
void grow(BindingsBuilder newBindings)
{
for (auto & i : *bindings)
newBindings->push_back(i);
bindings = newBindings;
newBindings.push_back(i);
std::swap(*this, newBindings);
}
friend struct ExprAttrs;

View File

@@ -56,6 +56,14 @@ MakeError(MissingArgumentError, EvalError);
MakeError(InfiniteRecursionError, EvalError);
MakeError(IFDError, EvalBaseError);
/**
* An evaluation error which should be retried instead of rethrown
*
* A RecoverableEvalError is not an EvalError, because we shouldn't cache it in the eval cache, as it should be retried
* anyway.
*/
MakeError(RecoverableEvalError, EvalBaseError);
struct InvalidPathError : public EvalError
{
public:

View File

@@ -5,6 +5,7 @@
#include "nix/expr/eval.hh"
#include "nix/expr/eval-error.hh"
#include "nix/expr/eval-settings.hh"
#include <exception>
namespace nix {
@@ -97,12 +98,19 @@ void EvalState::forceValue(Value & v, const PosIdx pos)
else
ExprBlackHole::throwInfiniteRecursionError(*this, v);
} catch (...) {
v.mkThunk(env, expr);
tryFixupBlackHolePos(v, pos);
handleEvalExceptionForThunk(env, expr, v, pos);
throw;
}
} else if (v.isApp())
callFunction(*v.app().left, *v.app().right, v, pos);
} else if (v.isApp()) {
try {
callFunction(*v.app().left, *v.app().right, v, pos);
} catch (...) {
handleEvalExceptionForApp(v);
throw;
}
} else if (v.isFailed()) {
handleEvalFailed(v, pos);
}
}
[[gnu::always_inline]]

View File

@@ -213,23 +213,100 @@ struct DebugTrace
}
};
struct StaticEvalSymbols
{
Symbol with, outPath, drvPath, type, meta, name, value, system, overrides, outputs, outputName, ignoreNulls, file,
line, column, functor, toString, right, wrong, structuredAttrs, json, allowedReferences, allowedRequisites,
disallowedReferences, disallowedRequisites, maxSize, maxClosureSize, builder, args, contentAddressed, impure,
outputHash, outputHashAlgo, outputHashMode, recurseForDerivations, description, self, epsilon, startSet,
operator_, key, path, prefix, outputSpecified;
Expr::AstSymbols exprSymbols;
static constexpr auto preallocate()
{
StaticSymbolTable alloc;
StaticEvalSymbols staticSymbols = {
.with = alloc.create("<with>"),
.outPath = alloc.create("outPath"),
.drvPath = alloc.create("drvPath"),
.type = alloc.create("type"),
.meta = alloc.create("meta"),
.name = alloc.create("name"),
.value = alloc.create("value"),
.system = alloc.create("system"),
.overrides = alloc.create("__overrides"),
.outputs = alloc.create("outputs"),
.outputName = alloc.create("outputName"),
.ignoreNulls = alloc.create("__ignoreNulls"),
.file = alloc.create("file"),
.line = alloc.create("line"),
.column = alloc.create("column"),
.functor = alloc.create("__functor"),
.toString = alloc.create("__toString"),
.right = alloc.create("right"),
.wrong = alloc.create("wrong"),
.structuredAttrs = alloc.create("__structuredAttrs"),
.json = alloc.create("__json"),
.allowedReferences = alloc.create("allowedReferences"),
.allowedRequisites = alloc.create("allowedRequisites"),
.disallowedReferences = alloc.create("disallowedReferences"),
.disallowedRequisites = alloc.create("disallowedRequisites"),
.maxSize = alloc.create("maxSize"),
.maxClosureSize = alloc.create("maxClosureSize"),
.builder = alloc.create("builder"),
.args = alloc.create("args"),
.contentAddressed = alloc.create("__contentAddressed"),
.impure = alloc.create("__impure"),
.outputHash = alloc.create("outputHash"),
.outputHashAlgo = alloc.create("outputHashAlgo"),
.outputHashMode = alloc.create("outputHashMode"),
.recurseForDerivations = alloc.create("recurseForDerivations"),
.description = alloc.create("description"),
.self = alloc.create("self"),
.epsilon = alloc.create(""),
.startSet = alloc.create("startSet"),
.operator_ = alloc.create("operator"),
.key = alloc.create("key"),
.path = alloc.create("path"),
.prefix = alloc.create("prefix"),
.outputSpecified = alloc.create("outputSpecified"),
.exprSymbols = {
.sub = alloc.create("__sub"),
.lessThan = alloc.create("__lessThan"),
.mul = alloc.create("__mul"),
.div = alloc.create("__div"),
.or_ = alloc.create("or"),
.findFile = alloc.create("__findFile"),
.nixPath = alloc.create("__nixPath"),
.body = alloc.create("body"),
}};
return std::pair{staticSymbols, alloc};
}
static consteval StaticEvalSymbols create()
{
return preallocate().first;
}
static constexpr StaticSymbolTable staticSymbolTable()
{
return preallocate().second;
}
};
class EvalState : public std::enable_shared_from_this<EvalState>
{
public:
static constexpr StaticEvalSymbols s = StaticEvalSymbols::create();
const fetchers::Settings & fetchSettings;
const EvalSettings & settings;
SymbolTable symbols;
PosTable positions;
const Symbol sWith, sOutPath, sDrvPath, sType, sMeta, sName, sValue, sSystem, sOverrides, sOutputs, sOutputName,
sIgnoreNulls, sFile, sLine, sColumn, sFunctor, sToString, sRight, sWrong, sStructuredAttrs, sJson,
sAllowedReferences, sAllowedRequisites, sDisallowedReferences, sDisallowedRequisites, sMaxSize, sMaxClosureSize,
sBuilder, sArgs, sContentAddressed, sImpure, sOutputHash, sOutputHashAlgo, sOutputHashMode,
sRecurseForDerivations, sDescription, sSelf, sEpsilon, sStartSet, sOperator, sKey, sPath, sPrefix,
sOutputSpecified;
const Expr::AstSymbols exprSymbols;
/**
* If set, force copying files to the Nix store even if they
* already exist there.
@@ -533,8 +610,28 @@ public:
*/
inline void forceValue(Value & v, const PosIdx pos);
private:
/**
* Internal support function for forceValue
*
* This code is factored out so that it's not in the heavily inlined hot path.
*/
void handleEvalExceptionForThunk(Env * env, Expr * expr, Value & v, const PosIdx pos);
/**
* Internal support function for forceValue
*
* This code is factored out so that it's not in the heavily inlined hot path.
*/
void handleEvalExceptionForApp(Value & v);
void handleEvalFailed(Value & v, PosIdx pos);
void tryFixupBlackHolePos(Value & v, PosIdx pos);
public:
/**
* Force a value, then recursively force list elements and
* attributes.
@@ -802,7 +899,7 @@ public:
BindingsBuilder buildBindings(size_t capacity)
{
return BindingsBuilder(*this, allocBindings(capacity));
return BindingsBuilder(*this, allocBindings(capacity), capacity);
}
ListBuilder buildList(size_t size)

View File

@@ -158,7 +158,7 @@ struct ExprString : Expr
ExprString(std::string && s)
: s(std::move(s))
{
v.mkString(this->s.data());
v.mkStringNoCopy(this->s.data());
};
Value * maybeThunk(EvalState & state, Env & env) override;
@@ -595,12 +595,17 @@ struct ExprOpNot : Expr
{ \
return pos; \
} \
};
}
MakeBinOp(ExprOpEq, "==") MakeBinOp(ExprOpNEq, "!=") MakeBinOp(ExprOpAnd, "&&") MakeBinOp(ExprOpOr, "||")
MakeBinOp(ExprOpImpl, "->") MakeBinOp(ExprOpUpdate, "//") MakeBinOp(ExprOpConcatLists, "++")
MakeBinOp(ExprOpEq, "==");
MakeBinOp(ExprOpNEq, "!=");
MakeBinOp(ExprOpAnd, "&&");
MakeBinOp(ExprOpOr, "||");
MakeBinOp(ExprOpImpl, "->");
MakeBinOp(ExprOpUpdate, "//");
MakeBinOp(ExprOpConcatLists, "++");
struct ExprConcatStrings : Expr
struct ExprConcatStrings : Expr
{
PosIdx pos;
bool forceString;

View File

@@ -88,7 +88,7 @@ struct ParserState
SourcePath basePath;
PosTable::Origin origin;
const ref<SourceAccessor> rootFS;
const Expr::AstSymbols & s;
static constexpr Expr::AstSymbols s = StaticEvalSymbols::create().exprSymbols;
const EvalSettings & settings;
void dupAttr(const AttrPath & attrPath, const PosIdx pos, const PosIdx prevPos);

View File

@@ -28,6 +28,8 @@ public:
}
};
class StaticSymbolTable;
/**
* Symbols have the property that they can be compared efficiently
* (using an equality test), because the symbol table stores only one
@@ -37,36 +39,38 @@ class Symbol
{
friend class SymbolStr;
friend class SymbolTable;
friend class StaticSymbolTable;
private:
uint32_t id;
explicit Symbol(uint32_t id) noexcept
explicit constexpr Symbol(uint32_t id) noexcept
: id(id)
{
}
public:
Symbol() noexcept
constexpr Symbol() noexcept
: id(0)
{
}
[[gnu::always_inline]]
explicit operator bool() const noexcept
constexpr explicit operator bool() const noexcept
{
return id > 0;
}
auto operator<=>(const Symbol other) const noexcept
/**
* The ID is a private implementation detail that should generally not be observed. However, we expose here just for
* sake of `switch...case`, which needs to dispatch on numbers. */
[[gnu::always_inline]]
constexpr uint32_t getId() const noexcept
{
return id <=> other.id;
return id;
}
bool operator==(const Symbol other) const noexcept
{
return id == other.id;
}
constexpr auto operator<=>(const Symbol & other) const noexcept = default;
friend class std::hash<Symbol>;
};
@@ -118,12 +122,12 @@ public:
// for multi-threaded implementations: lock store and allocator here
const auto & [v, idx] = key.store.add(SymbolValue{});
if (size == 0) {
v.mkString("", nullptr);
v.mkStringNoCopy("", nullptr);
} else {
auto s = key.alloc.allocate(size + 1);
memcpy(s, key.s.data(), size);
s[size] = '\0';
v.mkString(s, nullptr);
v.mkStringNoCopy(s, nullptr);
}
v.size_ = size;
v.idx = idx;
@@ -210,6 +214,39 @@ public:
};
};
class SymbolTable;
/**
* Convenience class to statically assign symbol identifiers at compile-time.
*/
class StaticSymbolTable
{
static constexpr std::size_t maxSize = 1024;
struct StaticSymbolInfo
{
std::string_view str;
Symbol sym;
};
std::array<StaticSymbolInfo, maxSize> symbols;
std::size_t size = 0;
public:
constexpr StaticSymbolTable() = default;
constexpr Symbol create(std::string_view str)
{
/* No need to check bounds because out of bounds access is
a compilation error. */
auto sym = Symbol(size + 1); //< +1 because Symbol with id = 0 is reserved
symbols[size++] = {str, sym};
return sym;
}
void copyIntoSymbolTable(SymbolTable & symtab) const;
};
/**
* Symbol table used by the parser and evaluator to represent and look
* up identifiers and attributes efficiently.
@@ -232,6 +269,10 @@ private:
boost::unordered_flat_set<SymbolStr, SymbolStr::Hash, SymbolStr::Equal> symbols{SymbolStr::chunkSize};
public:
SymbolTable(const StaticSymbolTable & staticSymtab)
{
staticSymtab.copyIntoSymbolTable(*this);
}
/**
* Converts a string into a symbol.
@@ -276,6 +317,16 @@ public:
}
};
inline void StaticSymbolTable::copyIntoSymbolTable(SymbolTable & symtab) const
{
for (std::size_t i = 0; i < size; ++i) {
auto [str, staticSym] = symbols[i];
auto sym = symtab.create(str);
if (sym != staticSym) [[unlikely]]
unreachable();
}
}
} // namespace nix
template<>

View File

@@ -2,6 +2,7 @@
///@file
#include <cassert>
#include <exception>
#include <span>
#include <type_traits>
#include <concepts>
@@ -35,6 +36,7 @@ typedef enum {
tBool,
tNull,
tFloat,
tFailed,
tExternal,
tPrimOp,
tAttrs,
@@ -57,6 +59,7 @@ typedef enum {
*/
typedef enum {
nThunk,
nFailed,
nInt,
nFloat,
nBool,
@@ -265,6 +268,30 @@ struct ValueBase
size_t size;
Value * const * elems;
};
/**
Representation of an evaluation that previously failed.
`Value` references `Failed` by packed pointer, and its `new` is GC managed.
@see gc_cleanup std::exception_ptr
*/
class Failed : public gc_cleanup
{
public:
std::exception_ptr ex;
/**
* Optional value for recovering `RecoverableEvalError`
* Must be set iff `ex` is an instance of `RecoverableEvalError`.
*/
Value * recoveryValue;
Failed(std::exception_ptr ex, Value * recoveryValue)
: ex(ex)
, recoveryValue(recoveryValue)
{
}
};
};
template<typename T>
@@ -291,6 +318,7 @@ struct PayloadTypeToInternalType
MACRO(PrimOp *, primOp, tPrimOp) \
MACRO(ValueBase::PrimOpApplicationThunk, primOpApp, tPrimOpApp) \
MACRO(ExternalValueBase *, external, tExternal) \
MACRO(ValueBase::Failed *, failed, tFailed) \
MACRO(NixFloat, fpoint, tFloat)
#define NIX_VALUE_PAYLOAD_TYPE(T, FIELD_NAME, DISCRIMINATOR) \
@@ -369,7 +397,7 @@ namespace detail {
/* Whether to use a specialization of ValueStorage that does bitpacking into
alignment niches. */
template<std::size_t ptrSize>
inline constexpr bool useBitPackedValueStorage = (ptrSize == 8) && (__STDCPP_DEFAULT_NEW_ALIGNMENT__ >= 8);
inline constexpr bool useBitPackedValueStorage = (ptrSize == 8) && (__STDCPP_DEFAULT_NEW_ALIGNMENT__ >= 16);
} // namespace detail
@@ -378,7 +406,8 @@ inline constexpr bool useBitPackedValueStorage = (ptrSize == 8) && (__STDCPP_DEF
* Packs discriminator bits into the pointer alignment niches.
*/
template<std::size_t ptrSize>
class ValueStorage<ptrSize, std::enable_if_t<detail::useBitPackedValueStorage<ptrSize>>> : public detail::ValueBase
class alignas(16) ValueStorage<ptrSize, std::enable_if_t<detail::useBitPackedValueStorage<ptrSize>>>
: public detail::ValueBase
{
/* Needs a dependent type name in order for member functions (and
* potentially ill-formed bit casts) to be SFINAE'd out.
@@ -594,6 +623,11 @@ protected:
path.path = std::bit_cast<const char *>(payload[1]);
}
void getStorage(Failed *& failed) const noexcept
{
failed = std::bit_cast<Failed *>(payload[1]);
}
void setStorage(NixInt integer) noexcept
{
setSingleDWordPayload<tInt>(integer.value);
@@ -643,6 +677,11 @@ protected:
{
setUntaggablePayload<pdPath>(path.accessor, path.path);
}
void setStorage(Failed * failed) noexcept
{
setSingleDWordPayload<tFailed>(std::bit_cast<PackedPointer>(failed));
}
};
/**
@@ -865,12 +904,12 @@ public:
inline bool isThunk() const
{
return isa<tThunk>();
};
}
inline bool isApp() const
{
return isa<tApp>();
};
}
inline bool isBlackhole() const;
@@ -878,17 +917,22 @@ public:
inline bool isLambda() const
{
return isa<tLambda>();
};
}
inline bool isPrimOp() const
{
return isa<tPrimOp>();
};
}
inline bool isPrimOpApp() const
{
return isa<tPrimOpApp>();
};
}
inline bool isFailed() const
{
return isa<tFailed>();
}
/**
* Returns the normal type of a Value. This only returns nThunk if
@@ -925,6 +969,8 @@ public:
return nExternal;
case tFloat:
return nFloat;
case tFailed:
return nFailed;
case tThunk:
case tApp:
return nThunk;
@@ -960,7 +1006,7 @@ public:
setStorage(b);
}
inline void mkString(const char * s, const char ** context = 0) noexcept
void mkStringNoCopy(const char * s, const char ** context = 0) noexcept
{
setStorage(StringWithContext{.c_str = s, .context = context});
}
@@ -972,7 +1018,6 @@ public:
void mkStringMove(const char * s, const NixStringContext & context);
void mkPath(const SourcePath & path);
void mkPath(std::string_view path);
inline void mkPath(SourceAccessor * accessor, const char * path) noexcept
{
@@ -993,12 +1038,20 @@ public:
void mkList(const ListBuilder & builder) noexcept
{
if (builder.size == 1)
switch (builder.size) {
case 0:
setStorage(List{.size = 0, .elems = nullptr});
break;
case 1:
setStorage(std::array<Value *, 2>{builder.inlineElems[0], nullptr});
else if (builder.size == 2)
break;
case 2:
setStorage(std::array<Value *, 2>{builder.inlineElems[0], builder.inlineElems[1]});
else
break;
default:
setStorage(List{.size = builder.size, .elems = builder.elems});
break;
}
}
inline void mkThunk(Env * e, Expr * ex) noexcept
@@ -1040,6 +1093,11 @@ public:
setStorage(n);
}
inline void mkFailed(std::exception_ptr e, Value * recovery) noexcept
{
setStorage(new Value::Failed(e, recovery));
}
bool isList() const noexcept
{
return isa<tListSmall, tListN>();
@@ -1143,6 +1201,11 @@ public:
{
return getStorage<Path>().accessor;
}
Failed * failed() const noexcept
{
return getStorage<Failed *>();
}
};
extern ExprBlackHole eBlackHole;

View File

@@ -40,7 +40,10 @@ endforeach
boost = dependency(
'boost',
modules : [ 'container', 'context' ],
modules : [
'container',
'context',
],
include_type : 'system',
)
# boost is a public dependency, but not a pkg-config dependency unfortunately, so we
@@ -71,6 +74,12 @@ toml11 = dependency(
method : 'cmake',
include_type : 'system',
)
configdata_priv.set(
'HAVE_TOML11_4',
toml11.version().version_compare('>= 4.0.0').to_int(),
)
deps_other += toml11
config_priv_h = configure_file(

View File

@@ -68,8 +68,7 @@ Expr * parseExprFromBuf(
const EvalSettings & settings,
PosTable & positions,
DocCommentMap & docComments,
const ref<SourceAccessor> rootFS,
const Expr::AstSymbols & astSymbols);
const ref<SourceAccessor> rootFS);
}
@@ -542,8 +541,7 @@ Expr * parseExprFromBuf(
const EvalSettings & settings,
PosTable & positions,
DocCommentMap & docComments,
const ref<SourceAccessor> rootFS,
const Expr::AstSymbols & astSymbols)
const ref<SourceAccessor> rootFS)
{
yyscan_t scanner;
LexerState lexerState {
@@ -558,7 +556,6 @@ Expr * parseExprFromBuf(
.basePath = basePath,
.origin = lexerState.origin,
.rootFS = rootFS,
.s = astSymbols,
.settings = settings,
};

View File

@@ -214,20 +214,20 @@ void derivationToValue(
auto path2 = path.path.abs();
Derivation drv = state.store->readDerivation(storePath);
auto attrs = state.buildBindings(3 + drv.outputs.size());
attrs.alloc(state.sDrvPath)
attrs.alloc(state.s.drvPath)
.mkString(
path2,
{
NixStringContextElem::DrvDeep{.drvPath = storePath},
});
attrs.alloc(state.sName).mkString(drv.env["name"]);
attrs.alloc(state.s.name).mkString(drv.env["name"]);
auto list = state.buildList(drv.outputs.size());
for (const auto & [i, o] : enumerate(drv.outputs)) {
mkOutputString(state, attrs, storePath, o);
(list[i] = state.allocValue())->mkString(o.first);
}
attrs.alloc(state.sOutputs).mkList(list);
attrs.alloc(state.s.outputs).mkList(list);
auto w = state.allocValue();
w->mkAttrs(attrs);
@@ -483,42 +483,41 @@ void prim_exec(EvalState & state, const PosIdx pos, Value ** args, Value & v)
static void prim_typeOf(EvalState & state, const PosIdx pos, Value ** args, Value & v)
{
state.forceValue(*args[0], pos);
std::string t;
switch (args[0]->type()) {
case nInt:
t = "int";
v.mkStringNoCopy("int");
break;
case nBool:
t = "bool";
v.mkStringNoCopy("bool");
break;
case nString:
t = "string";
v.mkStringNoCopy("string");
break;
case nPath:
t = "path";
v.mkStringNoCopy("path");
break;
case nNull:
t = "null";
v.mkStringNoCopy("null");
break;
case nAttrs:
t = "set";
v.mkStringNoCopy("set");
break;
case nList:
t = "list";
v.mkStringNoCopy("list");
break;
case nFunction:
t = "lambda";
v.mkStringNoCopy("lambda");
break;
case nExternal:
t = args[0]->external()->typeOf();
v.mkString(args[0]->external()->typeOf());
break;
case nFloat:
t = "float";
v.mkStringNoCopy("float");
break;
case nThunk:
case nFailed:
unreachable();
}
v.mkString(t);
}
static RegisterPrimOp primop_typeOf({
@@ -731,7 +730,7 @@ static void prim_genericClosure(EvalState & state, const PosIdx pos, Value ** ar
/* Get the start set. */
auto startSet = state.getAttr(
state.sStartSet, args[0]->attrs(), "in the attrset passed as argument to builtins.genericClosure");
state.s.startSet, args[0]->attrs(), "in the attrset passed as argument to builtins.genericClosure");
state.forceList(
*startSet->value,
@@ -749,7 +748,7 @@ static void prim_genericClosure(EvalState & state, const PosIdx pos, Value ** ar
/* Get the operator. */
auto op = state.getAttr(
state.sOperator, args[0]->attrs(), "in the attrset passed as argument to builtins.genericClosure");
state.s.operator_, args[0]->attrs(), "in the attrset passed as argument to builtins.genericClosure");
state.forceFunction(
*op->value, noPos, "while evaluating the 'operator' attribute passed as argument to builtins.genericClosure");
@@ -771,7 +770,7 @@ static void prim_genericClosure(EvalState & state, const PosIdx pos, Value ** ar
"while evaluating one of the elements generated by (or initially passed to) builtins.genericClosure");
auto key = state.getAttr(
state.sKey,
state.s.key,
e->attrs(),
"in one of the attrsets generated by (or initially passed to) builtins.genericClosure");
state.forceValue(*key->value, noPos);
@@ -1076,11 +1075,11 @@ static void prim_tryEval(EvalState & state, const PosIdx pos, Value ** args, Val
try {
state.forceValue(*args[0], pos);
attrs.insert(state.sValue, args[0]);
attrs.insert(state.s.value, args[0]);
attrs.insert(state.symbols.create("success"), &state.vTrue);
} catch (AssertionError & e) {
// `value = false;` is unfortunate but removing it is a breaking change.
attrs.insert(state.sValue, &state.vFalse);
attrs.insert(state.s.value, &state.vFalse);
attrs.insert(state.symbols.create("success"), &state.vFalse);
}
@@ -1292,7 +1291,8 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value **
auto attrs = args[0]->attrs();
/* Figure out the name first (for stack backtraces). */
auto nameAttr = state.getAttr(state.sName, attrs, "in the attrset passed as argument to builtins.derivationStrict");
auto nameAttr =
state.getAttr(state.s.name, attrs, "in the attrset passed as argument to builtins.derivationStrict");
std::string_view drvName;
try {
@@ -1366,7 +1366,7 @@ static void derivationStrictInternal(EvalState & state, std::string_view drvName
using nlohmann::json;
std::optional<StructuredAttrs> jsonObject;
auto pos = v.determinePos(noPos);
auto attr = attrs->find(state.sStructuredAttrs);
auto attr = attrs->find(state.s.structuredAttrs);
if (attr != attrs->end()
&& state.forceBool(
*attr->value,
@@ -1377,7 +1377,7 @@ static void derivationStrictInternal(EvalState & state, std::string_view drvName
/* Check whether null attributes should be ignored. */
bool ignoreNulls = false;
attr = attrs->find(state.sIgnoreNulls);
attr = attrs->find(state.s.ignoreNulls);
if (attr != attrs->end())
ignoreNulls = state.forceBool(
*attr->value,
@@ -1401,7 +1401,7 @@ static void derivationStrictInternal(EvalState & state, std::string_view drvName
outputs.insert("out");
for (auto & i : attrs->lexicographicOrder(state.symbols)) {
if (i->name == state.sIgnoreNulls)
if (i->name == state.s.ignoreNulls)
continue;
auto key = state.symbols[i->name];
vomit("processing attribute '%1%'", key);
@@ -1453,19 +1453,22 @@ static void derivationStrictInternal(EvalState & state, std::string_view drvName
continue;
}
if (i->name == state.sContentAddressed && state.forceBool(*i->value, pos, context_below)) {
contentAddressed = true;
experimentalFeatureSettings.require(Xp::CaDerivations);
}
else if (i->name == state.sImpure && state.forceBool(*i->value, pos, context_below)) {
isImpure = true;
experimentalFeatureSettings.require(Xp::ImpureDerivations);
}
switch (i->name.getId()) {
case EvalState::s.contentAddressed.getId():
if (state.forceBool(*i->value, pos, context_below)) {
contentAddressed = true;
experimentalFeatureSettings.require(Xp::CaDerivations);
}
break;
case EvalState::s.impure.getId():
if (state.forceBool(*i->value, pos, context_below)) {
isImpure = true;
experimentalFeatureSettings.require(Xp::ImpureDerivations);
}
break;
/* The `args' attribute is special: it supplies the
command-line arguments to the builder. */
else if (i->name == state.sArgs) {
case EvalState::s.args.getId():
state.forceList(*i->value, pos, context_below);
for (auto elem : i->value->listView()) {
auto s = state
@@ -1474,86 +1477,116 @@ static void derivationStrictInternal(EvalState & state, std::string_view drvName
.toOwned();
drv.args.push_back(s);
}
}
break;
/* All other attributes are passed to the builder through
the environment. */
else {
default:
if (jsonObject) {
if (i->name == state.sStructuredAttrs)
if (i->name == state.s.structuredAttrs)
continue;
jsonObject->structuredAttrs.emplace(key, printValueAsJSON(state, true, *i->value, pos, context));
if (i->name == state.sBuilder)
switch (i->name.getId()) {
case EvalState::s.builder.getId():
drv.builder = state.forceString(*i->value, context, pos, context_below);
else if (i->name == state.sSystem)
break;
case EvalState::s.system.getId():
drv.platform = state.forceStringNoCtx(*i->value, pos, context_below);
else if (i->name == state.sOutputHash)
break;
case EvalState::s.outputHash.getId():
outputHash = state.forceStringNoCtx(*i->value, pos, context_below);
else if (i->name == state.sOutputHashAlgo)
break;
case EvalState::s.outputHashAlgo.getId():
outputHashAlgo = parseHashAlgoOpt(state.forceStringNoCtx(*i->value, pos, context_below));
else if (i->name == state.sOutputHashMode)
break;
case EvalState::s.outputHashMode.getId():
handleHashMode(state.forceStringNoCtx(*i->value, pos, context_below));
else if (i->name == state.sOutputs) {
/* Require outputs to be a list of strings. */
break;
case EvalState::s.outputs.getId(): {
/* Require 'outputs' to be a list of strings. */
state.forceList(*i->value, pos, context_below);
Strings ss;
for (auto elem : i->value->listView())
ss.emplace_back(state.forceStringNoCtx(*elem, pos, context_below));
handleOutputs(ss);
break;
}
default:
break;
}
if (i->name == state.sAllowedReferences)
switch (i->name.getId()) {
case EvalState::s.allowedReferences.getId():
warn(
"In a derivation named '%s', 'structuredAttrs' disables the effect of the derivation attribute 'allowedReferences'; use 'outputChecks.<output>.allowedReferences' instead",
drvName);
if (i->name == state.sAllowedRequisites)
break;
case EvalState::s.allowedRequisites.getId():
warn(
"In a derivation named '%s', 'structuredAttrs' disables the effect of the derivation attribute 'allowedRequisites'; use 'outputChecks.<output>.allowedRequisites' instead",
drvName);
if (i->name == state.sDisallowedReferences)
break;
case EvalState::s.disallowedReferences.getId():
warn(
"In a derivation named '%s', 'structuredAttrs' disables the effect of the derivation attribute 'disallowedReferences'; use 'outputChecks.<output>.disallowedReferences' instead",
drvName);
if (i->name == state.sDisallowedRequisites)
break;
case EvalState::s.disallowedRequisites.getId():
warn(
"In a derivation named '%s', 'structuredAttrs' disables the effect of the derivation attribute 'disallowedRequisites'; use 'outputChecks.<output>.disallowedRequisites' instead",
drvName);
if (i->name == state.sMaxSize)
break;
case EvalState::s.maxSize.getId():
warn(
"In a derivation named '%s', 'structuredAttrs' disables the effect of the derivation attribute 'maxSize'; use 'outputChecks.<output>.maxSize' instead",
drvName);
if (i->name == state.sMaxClosureSize)
break;
case EvalState::s.maxClosureSize.getId():
warn(
"In a derivation named '%s', 'structuredAttrs' disables the effect of the derivation attribute 'maxClosureSize'; use 'outputChecks.<output>.maxClosureSize' instead",
drvName);
break;
default:
break;
}
} else {
auto s = state.coerceToString(pos, *i->value, context, context_below, true).toOwned();
if (i->name == state.sJson) {
if (i->name == state.s.json) {
warn(
"In derivation '%s': setting structured attributes via '__json' is deprecated, and may be disallowed in future versions of Nix. Set '__structuredAttrs = true' instead.",
drvName);
drv.structuredAttrs = StructuredAttrs::parse(s);
} else {
drv.env.emplace(key, s);
if (i->name == state.sBuilder)
switch (i->name.getId()) {
case EvalState::s.builder.getId():
drv.builder = std::move(s);
else if (i->name == state.sSystem)
break;
case EvalState::s.system.getId():
drv.platform = std::move(s);
else if (i->name == state.sOutputHash)
break;
case EvalState::s.outputHash.getId():
outputHash = std::move(s);
else if (i->name == state.sOutputHashAlgo)
break;
case EvalState::s.outputHashAlgo.getId():
outputHashAlgo = parseHashAlgoOpt(s);
else if (i->name == state.sOutputHashMode)
break;
case EvalState::s.outputHashMode.getId():
handleHashMode(s);
else if (i->name == state.sOutputs)
break;
case EvalState::s.outputs.getId():
handleOutputs(tokenizeString<Strings>(s));
break;
default:
break;
}
}
}
break;
}
} catch (Error & e) {
@@ -1722,7 +1755,7 @@ static void derivationStrictInternal(EvalState & state, std::string_view drvName
}
auto result = state.buildBindings(1 + drv.outputs.size());
result.alloc(state.sDrvPath)
result.alloc(state.s.drvPath)
.mkString(
drvPathS,
{
@@ -2006,14 +2039,14 @@ static void prim_findFile(EvalState & state, const PosIdx pos, Value ** args, Va
state.forceAttrs(*v2, pos, "while evaluating an element of the list passed to builtins.findFile");
std::string prefix;
auto i = v2->attrs()->find(state.sPrefix);
auto i = v2->attrs()->find(state.s.prefix);
if (i != v2->attrs()->end())
prefix = state.forceStringNoCtx(
*i->value,
pos,
"while evaluating the `prefix` attribute of an element of the list passed to builtins.findFile");
i = state.getAttr(state.sPath, v2->attrs(), "in an element of the __nixPath");
i = state.getAttr(state.s.path, v2->attrs(), "in an element of the __nixPath");
NixStringContext context;
auto path =
@@ -2786,7 +2819,7 @@ static void prim_path(EvalState & state, const PosIdx pos, Value ** args, Value
if (n == "path")
path.emplace(state.coerceToPath(
attr.pos, *attr.value, context, "while evaluating the 'path' attribute passed to 'builtins.path'"));
else if (attr.name == state.sName)
else if (attr.name == state.s.name)
name = state.forceStringNoCtx(
*attr.value, attr.pos, "while evaluating the `name` attribute passed to builtins.path");
else if (n == "filter")
@@ -3105,7 +3138,7 @@ static void prim_listToAttrs(EvalState & state, const PosIdx pos, Value ** args,
for (const auto & [n, v2] : enumerate(listView)) {
state.forceAttrs(*v2, pos, "while evaluating an element of the list passed to builtins.listToAttrs");
auto j = state.getAttr(state.sName, v2->attrs(), "in a {name=...; value=...;} pair");
auto j = state.getAttr(state.s.name, v2->attrs(), "in a {name=...; value=...;} pair");
auto name = state.forceStringNoCtx(
*j->value,
@@ -3132,7 +3165,7 @@ static void prim_listToAttrs(EvalState & state, const PosIdx pos, Value ** args,
// Note that .value is actually a Value * *; see earlier comments
Value * v2 = *std::bit_cast<ElemPtr>(attr.value);
auto j = state.getAttr(state.sValue, v2->attrs(), "in a {name=...; value=...;} pair");
auto j = state.getAttr(state.s.value, v2->attrs(), "in a {name=...; value=...;} pair");
prev = attr.name;
bindings.push_back({prev, j->value, j->pos});
}
@@ -3948,13 +3981,13 @@ static void prim_partition(EvalState & state, const PosIdx pos, Value ** args, V
auto rlist = state.buildList(rsize);
if (rsize)
memcpy(rlist.elems, right.data(), sizeof(Value *) * rsize);
attrs.alloc(state.sRight).mkList(rlist);
attrs.alloc(state.s.right).mkList(rlist);
auto wsize = wrong.size();
auto wlist = state.buildList(wsize);
if (wsize)
memcpy(wlist.elems, wrong.data(), sizeof(Value *) * wsize);
attrs.alloc(state.sWrong).mkList(wlist);
attrs.alloc(state.s.wrong).mkList(wlist);
v.mkAttrs(attrs);
}
@@ -4348,7 +4381,7 @@ static void prim_substring(EvalState & state, const PosIdx pos, Value ** args, V
if (len == 0) {
state.forceValue(*args[2], pos);
if (args[2]->type() == nString) {
v.mkString("", args[2]->context());
v.mkStringNoCopy("", args[2]->context());
return;
}
}
@@ -4873,7 +4906,7 @@ static void prim_parseDrvName(EvalState & state, const PosIdx pos, Value ** args
state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.parseDrvName");
DrvName parsed(name);
auto attrs = state.buildBindings(2);
attrs.alloc(state.sName).mkString(parsed.name);
attrs.alloc(state.s.name).mkString(parsed.name);
attrs.alloc("version").mkString(parsed.version);
v.mkAttrs(attrs);
}

View File

@@ -219,7 +219,7 @@ static void prim_getContext(EvalState & state, const PosIdx pos, Value ** args,
auto list = state.buildList(info.second.outputs.size());
for (const auto & [i, output] : enumerate(info.second.outputs))
(list[i] = state.allocValue())->mkString(output);
infoAttrs.alloc(state.sOutputs).mkList(list);
infoAttrs.alloc(state.s.outputs).mkList(list);
}
attrs.alloc(state.store->printStorePath(info.first)).mkAttrs(infoAttrs);
}
@@ -300,7 +300,7 @@ static void prim_appendContext(EvalState & state, const PosIdx pos, Value ** arg
}
}
if (auto attr = i.value->attrs()->get(state.sOutputs)) {
if (auto attr = i.value->attrs()->get(state.s.outputs)) {
state.forceList(*attr->value, attr->pos, "while evaluating the `outputs` attribute of a string context");
if (attr->value->listSize() && !isDerivation(name)) {
state

View File

@@ -185,7 +185,7 @@ static void prim_fetchClosure(EvalState & state, const PosIdx pos, Value ** args
{.msg = HintFmt("attribute '%s' is missing in call to 'fetchClosure'", "fromStore"),
.pos = state.positions[pos]});
auto parsedURL = parseURL(*fromStoreUrl);
auto parsedURL = parseURL(*fromStoreUrl, /*lenient=*/true);
if (parsedURL.scheme != "http" && parsedURL.scheme != "https"
&& !(getEnv("_NIX_IN_TEST").has_value() && parsedURL.scheme == "file"))

View File

@@ -84,7 +84,7 @@ static void prim_fetchMercurial(EvalState & state, const PosIdx pos, Value ** ar
auto [storePath, input2] = input.fetchToStore(state.store);
auto attrs2 = state.buildBindings(8);
state.mkStorePathString(storePath, attrs2.alloc(state.sOutPath));
state.mkStorePathString(storePath, attrs2.alloc(state.s.outPath));
if (input2.getRef())
attrs2.alloc("branch").mkString(*input2.getRef());
// Backward compatibility: set 'rev' to

View File

@@ -29,7 +29,7 @@ void emitTreeAttrs(
{
auto attrs = state.buildBindings(100);
state.mkStorePathString(storePath, attrs.alloc(state.sOutPath));
state.mkStorePathString(storePath, attrs.alloc(state.s.outPath));
// FIXME: support arbitrary input attributes.
@@ -95,7 +95,7 @@ static void fetchTree(
fetchers::Attrs attrs;
if (auto aType = args[0]->attrs()->get(state.sType)) {
if (auto aType = args[0]->attrs()->get(state.s.type)) {
if (type)
state.error<EvalError>("unexpected argument 'type'").atPos(pos).debugThrow();
type = state.forceStringNoCtx(
@@ -106,14 +106,14 @@ static void fetchTree(
attrs.emplace("type", type.value());
for (auto & attr : *args[0]->attrs()) {
if (attr.name == state.sType)
if (attr.name == state.s.type)
continue;
state.forceValue(*attr.value, attr.pos);
if (attr.value->type() == nPath || attr.value->type() == nString) {
auto s = state.coerceToString(attr.pos, *attr.value, context, "", false, false).toOwned();
attrs.emplace(
state.symbols[attr.name],
params.isFetchGit && state.symbols[attr.name] == "url" ? fixGitURL(s) : s);
params.isFetchGit && state.symbols[attr.name] == "url" ? fixGitURL(s).to_string() : s);
} else if (attr.value->type() == nBool)
attrs.emplace(state.symbols[attr.name], Explicit<bool>{attr.value->boolean()});
else if (attr.value->type() == nInt) {
@@ -175,7 +175,7 @@ static void fetchTree(
if (params.isFetchGit) {
fetchers::Attrs attrs;
attrs.emplace("type", "git");
attrs.emplace("url", fixGitURL(url));
attrs.emplace("url", fixGitURL(url).to_string());
if (!attrs.contains("exportIgnore")
&& (!attrs.contains("submodules") || !*fetchers::maybeGetBoolAttr(attrs, "submodules"))) {
attrs.emplace("exportIgnore", Explicit<bool>{true});

View File

@@ -1,75 +1,142 @@
#include "nix/expr/primops.hh"
#include "nix/expr/eval-inline.hh"
#include "expr-config-private.hh"
#include <sstream>
#include <toml.hpp>
namespace nix {
#if HAVE_TOML11_4
/**
* This is what toml11 < 4.0 did when choosing the subsecond precision.
* TOML 1.0.0 spec doesn't define how sub-millisecond ranges should be handled and calls it
* implementation defined behavior. For a lack of a better choice we stick with what older versions
* of toml11 did [1].
*
* [1]: https://github.com/ToruNiina/toml11/blob/dcfe39a783a94e8d52c885e5883a6fbb21529019/toml/datetime.hpp#L282
*/
static size_t normalizeSubsecondPrecision(toml::local_time lt)
{
auto millis = lt.millisecond;
auto micros = lt.microsecond;
auto nanos = lt.nanosecond;
if (millis != 0 || micros != 0 || nanos != 0) {
if (micros != 0 || nanos != 0) {
if (nanos != 0)
return 9;
return 6;
}
return 3;
}
return 0;
}
/**
* Normalize date/time formats to serialize to the same strings as versions prior to toml11 4.0.
*
* Several things to consider:
*
* 1. Sub-millisecond range is represented the same way as in toml11 versions prior to 4.0. Precisioun is rounded
* towards the next multiple of 3 or capped at 9 digits.
* 2. Seconds must be specified. This may become optional in (yet unreleased) TOML 1.1.0, but 1.0.0 defined local time
* in terms of RFC3339 [1].
* 3. date-time separator (`t`, `T` or space ` `) is canonicalized to an upper T. This is compliant with RFC3339
* [1] 5.6:
* > Applications that generate this format SHOULD use upper case letters.
*
* [1]: https://datatracker.ietf.org/doc/html/rfc3339#section-5.6
*/
static void normalizeDatetimeFormat(toml::value & t)
{
if (t.is_local_datetime()) {
auto & ldt = t.as_local_datetime();
t.as_local_datetime_fmt() = {
.delimiter = toml::datetime_delimiter_kind::upper_T,
// https://datatracker.ietf.org/doc/html/rfc3339#section-5.6
.has_seconds = true, // Mandated by TOML 1.0.0
.subsecond_precision = normalizeSubsecondPrecision(ldt.time),
};
return;
}
if (t.is_offset_datetime()) {
auto & odt = t.as_offset_datetime();
t.as_offset_datetime_fmt() = {
.delimiter = toml::datetime_delimiter_kind::upper_T,
// https://datatracker.ietf.org/doc/html/rfc3339#section-5.6
.has_seconds = true, // Mandated by TOML 1.0.0
.subsecond_precision = normalizeSubsecondPrecision(odt.time),
};
return;
}
if (t.is_local_time()) {
auto & lt = t.as_local_time();
t.as_local_time_fmt() = {
.has_seconds = true, // Mandated by TOML 1.0.0
.subsecond_precision = normalizeSubsecondPrecision(lt),
};
return;
}
}
#endif
static void prim_fromTOML(EvalState & state, const PosIdx pos, Value ** args, Value & val)
{
auto toml = state.forceStringNoCtx(*args[0], pos, "while evaluating the argument passed to builtins.fromTOML");
std::istringstream tomlStream(std::string{toml});
std::function<void(Value &, toml::value)> visit;
visit = [&](Value & v, toml::value t) {
auto visit = [&](auto & self, Value & v, toml::value t) -> void {
switch (t.type()) {
case toml::value_t::table: {
auto table = toml::get<toml::table>(t);
size_t size = 0;
for (auto & i : table) {
(void) i;
size++;
}
auto attrs = state.buildBindings(size);
auto attrs = state.buildBindings(table.size());
for (auto & elem : table) {
forceNoNullByte(elem.first);
visit(attrs.alloc(elem.first), elem.second);
self(self, attrs.alloc(elem.first), elem.second);
}
v.mkAttrs(attrs);
} break;
;
case toml::value_t::array: {
auto array = toml::get<std::vector<toml::value>>(t);
auto list = state.buildList(array.size());
for (const auto & [n, v] : enumerate(list))
visit(*(v = state.allocValue()), array[n]);
self(self, *(v = state.allocValue()), array[n]);
v.mkList(list);
} break;
;
case toml::value_t::boolean:
v.mkBool(toml::get<bool>(t));
break;
;
case toml::value_t::integer:
v.mkInt(toml::get<int64_t>(t));
break;
;
case toml::value_t::floating:
v.mkFloat(toml::get<NixFloat>(t));
break;
;
case toml::value_t::string: {
auto s = toml::get<std::string_view>(t);
forceNoNullByte(s);
v.mkString(s);
} break;
;
case toml::value_t::local_datetime:
case toml::value_t::offset_datetime:
case toml::value_t::local_date:
case toml::value_t::local_time: {
if (experimentalFeatureSettings.isEnabled(Xp::ParseTomlTimestamps)) {
#if HAVE_TOML11_4
normalizeDatetimeFormat(t);
#endif
auto attrs = state.buildBindings(2);
attrs.alloc("_type").mkString("timestamp");
attrs.alloc("_type").mkStringNoCopy("timestamp");
std::ostringstream s;
s << t;
auto str = toView(s);
@@ -80,16 +147,24 @@ static void prim_fromTOML(EvalState & state, const PosIdx pos, Value ** args, Va
throw std::runtime_error("Dates and times are not supported");
}
} break;
;
case toml::value_t::empty:
v.mkNull();
break;
;
}
};
try {
visit(val, toml::parse(tomlStream, "fromTOML" /* the "filename" */));
visit(
visit,
val,
toml::parse(
tomlStream,
"fromTOML" /* the "filename" */
#if HAVE_TOML11_4
,
toml::spec::v(1, 0, 0) // Be explicit that we are parsing TOML 1.0.0 without extensions
#endif
));
} catch (std::exception & e) { // TODO: toml::syntax_error
state.error<EvalError>("while parsing TOML: %s", e.what()).atPos(pos).debugThrow();
}

View File

@@ -75,6 +75,9 @@ void printAmbiguous(
str << "«potential infinite recursion»";
}
break;
case nFailed:
str << "«failed»";
break;
case nFunction:
if (v.isLambda()) {
str << "<LAMBDA>";

View File

@@ -272,7 +272,7 @@ private:
void printDerivation(Value & v)
{
std::optional<StorePath> storePath;
if (auto i = v.attrs()->get(state.sDrvPath)) {
if (auto i = v.attrs()->get(state.s.drvPath)) {
NixStringContext context;
storePath =
state.coerceToStorePath(i->pos, *i->value, context, "while evaluating the drvPath of a derivation");
@@ -508,6 +508,11 @@ private:
}
}
void printFailed(Value & v)
{
output << "«failed»";
}
void printExternal(Value & v)
{
v.external()->print(output);
@@ -583,6 +588,10 @@ private:
printThunk(v);
break;
case nFailed:
printFailed(v);
break;
case nExternal:
printExternal(v);
break;

View File

@@ -53,7 +53,7 @@ json printValueAsJSON(
out = *maybeString;
break;
}
if (auto i = v.attrs()->get(state.sOutPath))
if (auto i = v.attrs()->get(state.s.outPath))
return printValueAsJSON(state, strict, *i->value, i->pos, context, copyToStore);
else {
out = json::object();
@@ -96,6 +96,7 @@ json printValueAsJSON(
break;
case nThunk:
case nFailed:
case nFunction:
state.error<TypeError>("cannot convert %1% to JSON", showType(v)).atPos(v.determinePos(pos)).debugThrow();
}

View File

@@ -98,14 +98,14 @@ static void printValueAsXML(
XMLAttrs xmlAttrs;
Path drvPath;
if (auto a = v.attrs()->get(state.sDrvPath)) {
if (auto a = v.attrs()->get(state.s.drvPath)) {
if (strict)
state.forceValue(*a->value, a->pos);
if (a->value->type() == nString)
xmlAttrs["drvPath"] = drvPath = a->value->c_str();
}
if (auto a = v.attrs()->get(state.sOutPath)) {
if (auto a = v.attrs()->get(state.s.outPath)) {
if (strict)
state.forceValue(*a->value, a->pos);
if (a->value->type() == nString)
@@ -170,6 +170,11 @@ static void printValueAsXML(
case nThunk:
doc.writeEmptyElement("unevaluated");
break;
case nFailed:
doc.writeEmptyElement("failed");
break;
}
}

View File

@@ -2,6 +2,8 @@
#include "nix_api_fetchers_internal.hh"
#include "nix_api_util_internal.h"
extern "C" {
nix_fetchers_settings * nix_fetchers_settings_new(nix_c_context * context)
{
try {
@@ -17,3 +19,5 @@ void nix_fetchers_settings_free(nix_fetchers_settings * settings)
{
delete settings;
}
} // extern "C"

View File

@@ -25,7 +25,7 @@ static void downloadToSink(
std::string sha256Expected,
size_t sizeExpected)
{
FileTransferRequest request(url);
FileTransferRequest request(parseURL(url));
Headers headers;
if (authHeader.has_value())
headers.push_back({"Authorization", *authHeader});
@@ -69,7 +69,8 @@ static LfsApiInfo getLfsApi(const ParsedURL & url)
args.push_back("--");
args.push_back("git-lfs-authenticate");
args.push_back(url.path);
// FIXME %2F encode slashes? Does this command take/accept percent encoding?
args.push_back(url.renderPath(/*encode=*/false));
args.push_back("download");
auto [status, output] = runProgram({.program = "ssh", .args = args});
@@ -179,7 +180,7 @@ Fetch::Fetch(git_repository * repo, git_oid rev)
const auto remoteUrl = lfs::getLfsEndpointUrl(repo);
this->url = nix::parseURL(nix::fixGitURL(remoteUrl)).canonicalise();
this->url = nix::fixGitURL(remoteUrl).canonicalise();
}
bool Fetch::shouldFetch(const CanonPath & path) const
@@ -207,7 +208,7 @@ std::vector<nlohmann::json> Fetch::fetchUrls(const std::vector<Pointer> & pointe
auto api = lfs::getLfsApi(this->url);
auto url = api.endpoint + "/objects/batch";
const auto & authHeader = api.authHeader;
FileTransferRequest request(url);
FileTransferRequest request(parseURL(url));
request.post = true;
Headers headers;
if (authHeader.has_value())

View File

@@ -568,23 +568,34 @@ struct GitRepoImpl : GitRepo, std::enable_shared_from_this<GitRepoImpl>
void verifyCommit(const Hash & rev, const std::vector<fetchers::PublicKey> & publicKeys) override
{
// Map of SSH key types to their internal OpenSSH representations
static const std::unordered_map<std::string_view, std::string_view> keyTypeMap = {
{"ssh-dsa", "ssh-dsa"},
{"ssh-ecdsa", "ssh-ecdsa"},
{"ssh-ecdsa-sk", "sk-ecdsa-sha2-nistp256@openssh.com"},
{"ssh-ed25519", "ssh-ed25519"},
{"ssh-ed25519-sk", "sk-ssh-ed25519@openssh.com"},
{"ssh-rsa", "ssh-rsa"}};
// Create ad-hoc allowedSignersFile and populate it with publicKeys
auto allowedSignersFile = createTempFile().second;
std::string allowedSigners;
for (const fetchers::PublicKey & k : publicKeys) {
if (k.type != "ssh-dsa" && k.type != "ssh-ecdsa" && k.type != "ssh-ecdsa-sk" && k.type != "ssh-ed25519"
&& k.type != "ssh-ed25519-sk" && k.type != "ssh-rsa")
auto it = keyTypeMap.find(k.type);
if (it == keyTypeMap.end()) {
std::string supportedTypes;
for (const auto & [type, _] : keyTypeMap) {
supportedTypes += fmt(" %s\n", type);
}
throw Error(
"Unknown key type '%s'.\n"
"Please use one of\n"
"- ssh-dsa\n"
" ssh-ecdsa\n"
" ssh-ecdsa-sk\n"
" ssh-ed25519\n"
" ssh-ed25519-sk\n"
" ssh-rsa",
k.type);
allowedSigners += "* " + k.type + " " + k.key + "\n";
"Invalid SSH key type '%s' in publicKeys.\n"
"Please use one of:\n%s",
k.type,
supportedTypes);
}
allowedSigners += fmt("* %s %s\n", it->second, k.key);
}
writeFile(allowedSignersFile, allowedSigners);

View File

@@ -163,8 +163,8 @@ struct GitInputScheme : InputScheme
{
std::optional<Input> inputFromURL(const Settings & settings, const ParsedURL & url, bool requireTree) const override
{
if (url.scheme != "git" && url.scheme != "git+http" && url.scheme != "git+https" && url.scheme != "git+ssh"
&& url.scheme != "git+file")
auto parsedScheme = parseUrlScheme(url.scheme);
if (parsedScheme.application != "git")
return {};
auto url2(url);
@@ -233,9 +233,7 @@ struct GitInputScheme : InputScheme
Input input{settings};
input.attrs = attrs;
auto url = fixGitURL(getStrAttr(attrs, "url"));
parseURL(url);
input.attrs["url"] = url;
input.attrs["url"] = fixGitURL(getStrAttr(attrs, "url")).to_string();
getShallowAttr(input);
getSubmodulesAttr(input);
getAllRefsAttr(input);
@@ -464,8 +462,8 @@ struct GitInputScheme : InputScheme
// Why are we checking for bare repository?
// well if it's a bare repository we want to force a git fetch rather than copying the folder
bool isBareRepository = url.scheme == "file" && pathExists(url.path) && !pathExists(url.path + "/.git");
//
auto isBareRepository = [](PathView path) { return pathExists(path) && !pathExists(path + "/.git"); };
// FIXME: here we turn a possibly relative path into an absolute path.
// This allows relative git flake inputs to be resolved against the
// **current working directory** (as in POSIX), which tends to work out
@@ -474,8 +472,10 @@ struct GitInputScheme : InputScheme
//
// See: https://discourse.nixos.org/t/57783 and #9708
//
if (url.scheme == "file" && !forceHttp && !isBareRepository) {
if (!isAbsolute(url.path)) {
if (url.scheme == "file" && !forceHttp && !isBareRepository(renderUrlPathEnsureLegal(url.path))) {
auto path = renderUrlPathEnsureLegal(url.path);
if (!isAbsolute(path)) {
warn(
"Fetching Git repository '%s', which uses a path relative to the current directory. "
"This is not supported and will stop working in a future release. "
@@ -485,10 +485,10 @@ struct GitInputScheme : InputScheme
// If we don't check here for the path existence, then we can give libgit2 any directory
// and it will initialize them as git directories.
if (!pathExists(url.path)) {
throw Error("The path '%s' does not exist.", url.path);
if (!pathExists(path)) {
throw Error("The path '%s' does not exist.", path);
}
repoInfo.location = std::filesystem::absolute(url.path);
repoInfo.location = std::filesystem::absolute(path);
} else {
if (url.scheme == "file")
/* Query parameters are meaningless for file://, but

View File

@@ -19,7 +19,7 @@ namespace nix::fetchers {
struct DownloadUrl
{
std::string url;
ParsedURL url;
Headers headers;
};
@@ -38,7 +38,8 @@ struct GitArchiveInputScheme : InputScheme
if (url.scheme != schemeName())
return {};
auto path = tokenizeString<std::vector<std::string>>(url.path, "/");
/* This ignores empty path segments for back-compat. Older versions used a tokenizeString here. */
auto path = url.pathSegments(/*skipEmpty=*/true) | std::ranges::to<std::vector<std::string>>();
std::optional<Hash> rev;
std::optional<std::string> ref;
@@ -139,12 +140,12 @@ struct GitArchiveInputScheme : InputScheme
auto repo = getStrAttr(input.attrs, "repo");
auto ref = input.getRef();
auto rev = input.getRev();
auto path = owner + "/" + repo;
std::vector<std::string> path{owner, repo};
assert(!(ref && rev));
if (ref)
path += "/" + *ref;
path.push_back(*ref);
if (rev)
path += "/" + rev->to_string(HashFormat::Base16, false);
path.push_back(rev->to_string(HashFormat::Base16, false));
auto url = ParsedURL{
.scheme = std::string{schemeName()},
.path = path,
@@ -420,7 +421,7 @@ struct GitHubInputScheme : GitArchiveInputScheme
const auto url =
fmt(urlFmt, host, getOwner(input), getRepo(input), input.getRev()->to_string(HashFormat::Base16, false));
return DownloadUrl{url, headers};
return DownloadUrl{parseURL(url), headers};
}
void clone(const Input & input, const Path & destDir) const override
@@ -500,7 +501,7 @@ struct GitLabInputScheme : GitArchiveInputScheme
input.getRev()->to_string(HashFormat::Base16, false));
Headers headers = makeHeadersWithAuthTokens(*input.settings, host, input);
return DownloadUrl{url, headers};
return DownloadUrl{parseURL(url), headers};
}
void clone(const Input & input, const Path & destDir) const override
@@ -592,7 +593,7 @@ struct SourceHutInputScheme : GitArchiveInputScheme
input.getRev()->to_string(HashFormat::Base16, false));
Headers headers = makeHeadersWithAuthTokens(*input.settings, host, input);
return DownloadUrl{url, headers};
return DownloadUrl{parseURL(url), headers};
}
void clone(const Input & input, const Path & destDir) const override

View File

@@ -11,6 +11,7 @@ struct InputCache
ref<SourceAccessor> accessor;
Input resolvedInput;
Input lockedInput;
Attrs extraAttrs;
};
CachedResult getAccessor(ref<Store> store, const Input & originalInput, UseRegistries useRegistries);
@@ -19,6 +20,7 @@ struct InputCache
{
Input lockedInput;
ref<SourceAccessor> accessor;
Attrs extraAttrs;
};
virtual std::optional<CachedInput> lookup(const Input & originalInput) const = 0;

View File

@@ -14,7 +14,8 @@ struct IndirectInputScheme : InputScheme
if (url.scheme != "flake")
return {};
auto path = tokenizeString<std::vector<std::string>>(url.path, "/");
/* This ignores empty path segments for back-compat. Older versions used a tokenizeString here. */
auto path = url.pathSegments(/*skipEmpty=*/true) | std::ranges::to<std::vector<std::string>>();
std::optional<Hash> rev;
std::optional<std::string> ref;
@@ -82,16 +83,15 @@ struct IndirectInputScheme : InputScheme
ParsedURL toURL(const Input & input) const override
{
ParsedURL url;
url.scheme = "flake";
url.path = getStrAttr(input.attrs, "id");
ParsedURL url{
.scheme = "flake",
.path = {getStrAttr(input.attrs, "id")},
};
if (auto ref = input.getRef()) {
url.path += '/';
url.path += *ref;
url.path.push_back(*ref);
};
if (auto rev = input.getRev()) {
url.path += '/';
url.path += rev->gitRev();
url.path.push_back(rev->gitRev());
};
return url;
}

View File

@@ -22,7 +22,8 @@ InputCache::getAccessor(ref<Store> store, const Input & originalInput, UseRegist
fetched = lookup(resolvedInput);
if (!fetched) {
auto [accessor, lockedInput] = resolvedInput.getAccessor(store);
fetched.emplace(CachedInput{.lockedInput = lockedInput, .accessor = accessor});
fetched.emplace(
CachedInput{.lockedInput = lockedInput, .accessor = accessor, .extraAttrs = extraAttrs});
}
upsert(resolvedInput, *fetched);
} else {
@@ -36,7 +37,7 @@ InputCache::getAccessor(ref<Store> store, const Input & originalInput, UseRegist
debug("got tree '%s' from '%s'", fetched->accessor, fetched->lockedInput.to_string());
return {fetched->accessor, resolvedInput, fetched->lockedInput};
return {fetched->accessor, resolvedInput, fetched->lockedInput, fetched->extraAttrs};
}
struct InputCacheImpl : InputCache

View File

@@ -120,7 +120,7 @@ struct MercurialInputScheme : InputScheme
{
auto url = parseURL(getStrAttr(input.attrs, "url"));
if (url.scheme == "file" && !input.getRef() && !input.getRev())
return url.path;
return renderUrlPathEnsureLegal(url.path);
return {};
}
@@ -152,7 +152,7 @@ struct MercurialInputScheme : InputScheme
{
auto url = parseURL(getStrAttr(input.attrs, "url"));
bool isLocal = url.scheme == "file";
return {isLocal, isLocal ? url.path : url.to_string()};
return {isLocal, isLocal ? renderUrlPathEnsureLegal(url.path) : url.to_string()};
}
StorePath fetchToStore(ref<Store> store, Input & input) const

View File

@@ -20,7 +20,7 @@ struct PathInputScheme : InputScheme
Input input{settings};
input.attrs.insert_or_assign("type", "path");
input.attrs.insert_or_assign("path", url.path);
input.attrs.insert_or_assign("path", renderUrlPathEnsureLegal(url.path));
for (auto & [name, value] : url.query)
if (name == "rev" || name == "narHash")
@@ -74,7 +74,7 @@ struct PathInputScheme : InputScheme
query.erase("__final");
return ParsedURL{
.scheme = "path",
.path = getStrAttr(input.attrs, "path"),
.path = splitString<std::vector<std::string>>(getStrAttr(input.attrs, "path"), "/"),
.query = query,
};
}

View File

@@ -43,7 +43,7 @@ DownloadFileResult downloadFile(
if (cached && !cached->expired)
return useCached();
FileTransferRequest request(url);
FileTransferRequest request(ValidURL{url});
request.headers = headers;
if (cached)
request.expectedETag = getStrAttr(cached->value, "etag");
@@ -107,20 +107,20 @@ DownloadFileResult downloadFile(
}
static DownloadTarballResult downloadTarball_(
const Settings & settings, const std::string & url, const Headers & headers, const std::string & displayPrefix)
const Settings & settings, const std::string & urlS, const Headers & headers, const std::string & displayPrefix)
{
ValidURL url = urlS;
// Some friendly error messages for common mistakes.
// Namely lets catch when the url is a local file path, but
// it is not in fact a tarball.
if (url.rfind("file://", 0) == 0) {
// Remove "file://" prefix to get the local file path
std::string localPath = url.substr(7);
if (!std::filesystem::exists(localPath)) {
if (url.scheme() == "file") {
std::filesystem::path localPath = renderUrlPathEnsureLegal(url.path());
if (!exists(localPath)) {
throw Error("tarball '%s' does not exist.", localPath);
}
if (std::filesystem::is_directory(localPath)) {
if (std::filesystem::exists(localPath + "/.git")) {
if (is_directory(localPath)) {
if (exists(localPath / ".git")) {
throw Error(
"tarball '%s' is a git repository, not a tarball. Please use `git+file` as the scheme.", localPath);
}
@@ -128,7 +128,7 @@ static DownloadTarballResult downloadTarball_(
}
}
Cache::Key cacheKey{"tarball", {{"url", url}}};
Cache::Key cacheKey{"tarball", {{"url", urlS}}};
auto cached = settings.getCache()->lookupExpired(cacheKey);
@@ -166,7 +166,7 @@ static DownloadTarballResult downloadTarball_(
/* Note: if the download is cached, `importTarball()` will receive
no data, which causes it to import an empty tarball. */
auto archive = hasSuffix(toLower(parseURL(url).path), ".zip") ? ({
auto archive = !url.path().empty() && hasSuffix(toLower(url.path().back()), ".zip") ? ({
/* In streaming mode, libarchive doesn't handle
symlinks in zip files correctly (#10649). So write
the entire file to disk so libarchive can access it
@@ -180,7 +180,7 @@ static DownloadTarballResult downloadTarball_(
}
TarArchive{path};
})
: TarArchive{*source};
: TarArchive{*source};
auto tarballCache = getTarballCache();
auto parseSink = tarballCache->getFileSystemObjectSink();
auto lastModified = unpackTarfileToSink(archive, *parseSink);
@@ -234,8 +234,11 @@ struct CurlInputScheme : InputScheme
{
const StringSet transportUrlSchemes = {"file", "http", "https"};
bool hasTarballExtension(std::string_view path) const
bool hasTarballExtension(const ParsedURL & url) const
{
if (url.path.empty())
return false;
const auto & path = url.path.back();
return hasSuffix(path, ".zip") || hasSuffix(path, ".tar") || hasSuffix(path, ".tgz")
|| hasSuffix(path, ".tar.gz") || hasSuffix(path, ".tar.xz") || hasSuffix(path, ".tar.bz2")
|| hasSuffix(path, ".tar.zst");
@@ -336,7 +339,7 @@ struct FileInputScheme : CurlInputScheme
auto parsedUrlScheme = parseUrlScheme(url.scheme);
return transportUrlSchemes.count(std::string(parsedUrlScheme.transport))
&& (parsedUrlScheme.application ? parsedUrlScheme.application.value() == schemeName()
: (!requireTree && !hasTarballExtension(url.path)));
: (!requireTree && !hasTarballExtension(url)));
}
std::pair<ref<SourceAccessor>, Input> getAccessor(ref<Store> store, const Input & _input) const override
@@ -373,7 +376,7 @@ struct TarballInputScheme : CurlInputScheme
return transportUrlSchemes.count(std::string(parsedUrlScheme.transport))
&& (parsedUrlScheme.application ? parsedUrlScheme.application.value() == schemeName()
: (requireTree || hasTarballExtension(url.path)));
: (requireTree || hasTarballExtension(url)));
}
std::pair<ref<SourceAccessor>, Input> getAccessor(ref<Store> store, const Input & _input) const override

View File

@@ -10,6 +10,8 @@
#include "nix/flake/flake.hh"
extern "C" {
nix_flake_settings * nix_flake_settings_new(nix_c_context * context)
{
nix_clear_err(context);
@@ -203,3 +205,5 @@ nix_value * nix_locked_flake_get_output_attrs(
}
NIXC_CATCH_ERRS_NULL
}
} // extern "C"

View File

@@ -2,6 +2,7 @@
#include "nix/fetchers/fetch-settings.hh"
#include "nix/flake/flakeref.hh"
#include "nix/fetchers/attrs.hh"
namespace nix {
@@ -90,6 +91,158 @@ TEST(parseFlakeRef, GitArchiveInput)
}
}
struct InputFromURLTestCase
{
std::string url;
fetchers::Attrs attrs;
std::string description;
std::string expectedUrl = url;
};
class InputFromURLTest : public ::testing::WithParamInterface<InputFromURLTestCase>, public ::testing::Test
{};
TEST_P(InputFromURLTest, attrsAreCorrectAndRoundTrips)
{
experimentalFeatureSettings.experimentalFeatures.get().insert(Xp::Flakes);
fetchers::Settings fetchSettings;
const auto & testCase = GetParam();
auto flakeref = parseFlakeRef(fetchSettings, testCase.url);
EXPECT_EQ(flakeref.toAttrs(), testCase.attrs);
EXPECT_EQ(flakeref.to_string(), testCase.expectedUrl);
auto input = fetchers::Input::fromURL(fetchSettings, flakeref.to_string());
EXPECT_EQ(input.toURLString(), testCase.expectedUrl);
EXPECT_EQ(input.toAttrs(), testCase.attrs);
// Round-trip check.
auto input2 = fetchers::Input::fromURL(fetchSettings, input.toURLString());
EXPECT_EQ(input, input2);
EXPECT_EQ(input.toURLString(), input2.toURLString());
}
using fetchers::Attr;
INSTANTIATE_TEST_SUITE_P(
InputFromURL,
InputFromURLTest,
::testing::Values(
InputFromURLTestCase{
.url = "flake:nixpkgs",
.attrs =
{
{"id", Attr("nixpkgs")},
{"type", Attr("indirect")},
},
.description = "basic_indirect",
},
InputFromURLTestCase{
.url = "flake:nixpkgs/branch",
.attrs =
{
{"id", Attr("nixpkgs")},
{"type", Attr("indirect")},
{"ref", Attr("branch")},
},
.description = "basic_indirect_branch",
},
InputFromURLTestCase{
.url = "nixpkgs/branch",
.attrs =
{
{"id", Attr("nixpkgs")},
{"type", Attr("indirect")},
{"ref", Attr("branch")},
},
.description = "flake_id_ref_branch",
.expectedUrl = "flake:nixpkgs/branch",
},
InputFromURLTestCase{
.url = "nixpkgs/branch/2aae6c35c94fcfb415dbe95f408b9ce91ee846ed",
.attrs =
{
{"id", Attr("nixpkgs")},
{"type", Attr("indirect")},
{"ref", Attr("branch")},
{"rev", Attr("2aae6c35c94fcfb415dbe95f408b9ce91ee846ed")},
},
.description = "flake_id_ref_branch_trailing_slash",
.expectedUrl = "flake:nixpkgs/branch/2aae6c35c94fcfb415dbe95f408b9ce91ee846ed",
},
// The following tests are for back-compat with lax parsers in older versions
// that used `tokenizeString` for splitting path segments, which ignores empty
// strings.
InputFromURLTestCase{
.url = "nixpkgs/branch////",
.attrs =
{
{"id", Attr("nixpkgs")},
{"type", Attr("indirect")},
{"ref", Attr("branch")},
},
.description = "flake_id_ref_branch_ignore_empty_trailing_segments",
.expectedUrl = "flake:nixpkgs/branch",
},
InputFromURLTestCase{
.url = "nixpkgs/branch///2aae6c35c94fcfb415dbe95f408b9ce91ee846ed///",
.attrs =
{
{"id", Attr("nixpkgs")},
{"type", Attr("indirect")},
{"ref", Attr("branch")},
{"rev", Attr("2aae6c35c94fcfb415dbe95f408b9ce91ee846ed")},
},
.description = "flake_id_ref_branch_ignore_empty_segments_ref_rev",
.expectedUrl = "flake:nixpkgs/branch/2aae6c35c94fcfb415dbe95f408b9ce91ee846ed",
},
InputFromURLTestCase{
// Note that this is different from above because the "flake id" shorthand
// doesn't allow this.
.url = "flake:/nixpkgs///branch////",
.attrs =
{
{"id", Attr("nixpkgs")},
{"type", Attr("indirect")},
{"ref", Attr("branch")},
},
.description = "indirect_branch_empty_segments_everywhere",
.expectedUrl = "flake:nixpkgs/branch",
},
InputFromURLTestCase{
// TODO: Technically this has an empty authority, but it's ignored
// for now. Yes, this is what all versions going back to at least
// 2.18 did and yes, this should not be allowed.
.url = "github://////owner%42/////repo%41///branch%43////",
.attrs =
{
{"type", Attr("github")},
{"owner", Attr("ownerB")},
{"repo", Attr("repoA")},
{"ref", Attr("branchC")},
},
.description = "github_ref_slashes_in_path_everywhere",
.expectedUrl = "github:ownerB/repoA/branchC",
},
InputFromURLTestCase{
// FIXME: Subgroups in gitlab URLs are busted. This double-encoding
// behavior exists since 2.18. See issue #9161 and PR #8845.
.url = "gitlab:/owner%252Fsubgroup/////repo%41///branch%43////",
.attrs =
{
{"type", Attr("gitlab")},
{"owner", Attr("owner%2Fsubgroup")},
{"repo", Attr("repoA")},
{"ref", Attr("branchC")},
},
.description = "gitlab_ref_slashes_in_path_everywhere_with_pct_encoding",
.expectedUrl = "gitlab:owner%252Fsubgroup/repoA/branchC",
}),
[](const ::testing::TestParamInfo<InputFromURLTestCase> & info) { return info.param.description; });
TEST(to_string, doesntReencodeUrl)
{
fetchers::Settings fetchSettings;

View File

@@ -13,8 +13,9 @@ TEST(getNameFromURL, getNameFromURL)
ASSERT_EQ(getNameFromURL(parseURL("path:~/repos/nixpkgs#packages.x86_64-linux.Hello")), "Hello");
ASSERT_EQ(getNameFromURL(parseURL("path:.#nonStandardAttr.mylaptop")), "mylaptop");
ASSERT_EQ(getNameFromURL(parseURL("path:./repos/myflake#nonStandardAttr.mylaptop")), "mylaptop");
ASSERT_EQ(getNameFromURL(parseURL("path:./nixpkgs#packages.x86_64-linux.complex^bin,man")), "complex");
ASSERT_EQ(getNameFromURL(parseURL("path:./myproj#packages.x86_64-linux.default^*")), "myproj");
ASSERT_EQ(
getNameFromURL(parseURL("path:./nixpkgs#packages.x86_64-linux.complex^bin,man", /*lenient=*/true)), "complex");
ASSERT_EQ(getNameFromURL(parseURL("path:./myproj#packages.x86_64-linux.default^*", /*lenient=*/true)), "myproj");
ASSERT_EQ(getNameFromURL(parseURL("path:./myproj#defaultPackage.x86_64-linux")), "myproj");
ASSERT_EQ(getNameFromURL(parseURL("github:NixOS/nixpkgs#packages.x86_64-linux.hello")), "hello");
@@ -80,6 +81,6 @@ TEST(getNameFromURL, getNameFromURL)
ASSERT_EQ(getNameFromURL(parseURL("path:.")), std::nullopt);
ASSERT_EQ(getNameFromURL(parseURL("file:.#")), std::nullopt);
ASSERT_EQ(getNameFromURL(parseURL("path:.#packages.x86_64-linux.default")), std::nullopt);
ASSERT_EQ(getNameFromURL(parseURL("path:.#packages.x86_64-linux.default^*")), std::nullopt);
ASSERT_EQ(getNameFromURL(parseURL("path:.#packages.x86_64-linux.default^*", /*lenient=*/true)), std::nullopt);
}
} // namespace nix

View File

@@ -232,7 +232,7 @@ static Flake readFlake(
.path = flakePath,
};
if (auto description = vInfo.attrs()->get(state.sDescription)) {
if (auto description = vInfo.attrs()->get(state.s.description)) {
expectType(state, nString, *description->value, description->pos);
flake.description = description->value->c_str();
}
@@ -253,7 +253,7 @@ static Flake readFlake(
if (outputs->value->isLambda() && outputs->value->lambda().fun->hasFormals()) {
for (auto & formal : outputs->value->lambda().fun->formals->formals) {
if (formal.name != state.sSelf)
if (formal.name != state.s.self)
flake.inputs.emplace(
state.symbols[formal.name],
FlakeInput{.ref = parseFlakeRef(state.fetchSettings, std::string(state.symbols[formal.name]))});
@@ -305,7 +305,8 @@ static Flake readFlake(
}
for (auto & attr : *vInfo.attrs()) {
if (attr.name != state.sDescription && attr.name != sInputs && attr.name != sOutputs && attr.name != sNixConfig)
if (attr.name != state.s.description && attr.name != sInputs && attr.name != sOutputs
&& attr.name != sNixConfig)
throw Error(
"flake '%s' has an unsupported attribute '%s', at %s",
resolvedRef,
@@ -340,8 +341,9 @@ static Flake getFlake(
// Fetch a lazy tree first.
auto cachedInput = state.inputCache->getAccessor(state.store, originalRef.input, useRegistries);
auto resolvedRef = FlakeRef(std::move(cachedInput.resolvedInput), originalRef.subdir);
auto lockedRef = FlakeRef(std::move(cachedInput.lockedInput), originalRef.subdir);
auto subdir = fetchers::maybeGetStrAttr(cachedInput.extraAttrs, "dir").value_or(originalRef.subdir);
auto resolvedRef = FlakeRef(std::move(cachedInput.resolvedInput), subdir);
auto lockedRef = FlakeRef(std::move(cachedInput.lockedInput), subdir);
// Parse/eval flake.nix to get at the input.self attributes.
auto flake = readFlake(state, originalRef, resolvedRef, lockedRef, {cachedInput.accessor}, lockRootAttrPath);

View File

@@ -82,7 +82,7 @@ std::pair<FlakeRef, std::string> parsePathFlakeRefWithFragment(
auto succeeds = std::regex_match(url, match, pathFlakeRegex);
assert(succeeds);
auto path = match[1].str();
auto query = decodeQuery(match[3]);
auto query = decodeQuery(match[3].str(), /*lenient=*/true);
auto fragment = percentDecode(match[5].str());
if (baseDir) {
@@ -143,7 +143,7 @@ std::pair<FlakeRef, std::string> parsePathFlakeRefWithFragment(
auto parsedURL = ParsedURL{
.scheme = "git+file",
.authority = ParsedURL::Authority{},
.path = flakeRoot,
.path = splitString<std::vector<std::string>>(flakeRoot, "/"),
.query = query,
.fragment = fragment,
};
@@ -172,7 +172,13 @@ std::pair<FlakeRef, std::string> parsePathFlakeRefWithFragment(
return fromParsedURL(
fetchSettings,
{.scheme = "path", .authority = ParsedURL::Authority{}, .path = path, .query = query, .fragment = fragment},
{
.scheme = "path",
.authority = ParsedURL::Authority{},
.path = splitString<std::vector<std::string>>(path, "/"),
.query = query,
.fragment = fragment,
},
isFlake);
}
@@ -192,8 +198,8 @@ parseFlakeIdRef(const fetchers::Settings & fetchSettings, const std::string & ur
if (std::regex_match(url, match, flakeRegex)) {
auto parsedURL = ParsedURL{
.scheme = "flake",
.authority = ParsedURL::Authority{},
.path = match[1],
.authority = std::nullopt,
.path = splitString<std::vector<std::string>>(match[1].str(), "/"),
};
return std::make_pair(
@@ -210,9 +216,13 @@ std::optional<std::pair<FlakeRef, std::string>> parseURLFlakeRef(
bool isFlake)
{
try {
auto parsed = parseURL(url);
if (baseDir && (parsed.scheme == "path" || parsed.scheme == "git+file") && !isAbsolute(parsed.path))
parsed.path = absPath(parsed.path, *baseDir);
auto parsed = parseURL(url, /*lenient=*/true);
if (baseDir && (parsed.scheme == "path" || parsed.scheme == "git+file")) {
/* Here we know that the path must not contain encoded '/' or NUL bytes. */
auto path = renderUrlPathEnsureLegal(parsed.path);
if (!isAbsolute(path))
parsed.path = splitString<std::vector<std::string>>(absPath(path, *baseDir), "/");
}
return fromParsedURL(fetchSettings, std::move(parsed), isFlake);
} catch (BadURL &) {
return std::nullopt;
@@ -289,7 +299,7 @@ FlakeRef FlakeRef::canonicalize() const
filtering the `dir` query parameter from the URL. */
if (auto url = fetchers::maybeGetStrAttr(flakeRef.input.attrs, "url")) {
try {
auto parsed = parseURL(*url);
auto parsed = parseURL(*url, /*lenient=*/true);
if (auto dir2 = get(parsed.query, "dir")) {
if (flakeRef.subdir != "" && flakeRef.subdir == *dir2)
parsed.query.erase("dir");

View File

@@ -27,16 +27,21 @@ std::optional<std::string> getNameFromURL(const ParsedURL & url)
return match.str(2);
}
/* This is not right, because special chars like slashes within the
path fragments should be percent encoded, but I don't think any
of the regexes above care. */
auto path = concatStringsSep("/", url.path);
/* If this is a github/gitlab/sourcehut flake, use the repo name */
if (std::regex_match(url.scheme, gitProviderRegex) && std::regex_match(url.path, match, secondPathSegmentRegex))
if (std::regex_match(url.scheme, gitProviderRegex) && std::regex_match(path, match, secondPathSegmentRegex))
return match.str(1);
/* If it is a regular git flake, use the directory name */
if (std::regex_match(url.scheme, gitSchemeRegex) && std::regex_match(url.path, match, lastPathSegmentRegex))
if (std::regex_match(url.scheme, gitSchemeRegex) && std::regex_match(path, match, lastPathSegmentRegex))
return match.str(1);
/* If there is no fragment, take the last element of the path */
if (std::regex_match(url.path, match, lastPathSegmentRegex))
if (std::regex_match(path, match, lastPathSegmentRegex))
return match.str(1);
/* If even that didn't work, the URL does not contain enough info to determine a useful name */

View File

@@ -5,6 +5,8 @@
#include "nix/main/plugin.hh"
extern "C" {
nix_err nix_init_plugins(nix_c_context * context)
{
if (context)
@@ -14,3 +16,5 @@ nix_err nix_init_plugins(nix_c_context * context)
}
NIXC_CATCH_ERRS
}
} // extern "C"

View File

@@ -10,6 +10,8 @@
#include "nix/store/globals.hh"
extern "C" {
nix_err nix_libstore_init(nix_c_context * context)
{
if (context)
@@ -91,7 +93,7 @@ nix_store_get_version(nix_c_context * context, Store * store, nix_get_string_cal
NIXC_CATCH_ERRS
}
bool nix_store_is_valid_path(nix_c_context * context, Store * store, StorePath * path)
bool nix_store_is_valid_path(nix_c_context * context, Store * store, const StorePath * path)
{
if (context)
context->last_err_code = NIX_OK;
@@ -129,7 +131,7 @@ nix_err nix_store_realise(
Store * store,
StorePath * path,
void * userdata,
void (*callback)(void * userdata, const char *, const char *))
void (*callback)(void * userdata, const char *, const StorePath *))
{
if (context)
context->last_err_code = NIX_OK;
@@ -144,8 +146,8 @@ nix_err nix_store_realise(
if (callback) {
for (const auto & result : results) {
for (const auto & [outputName, realisation] : result.builtOutputs) {
auto op = store->ptr->printStorePath(realisation.outPath);
callback(userdata, outputName.c_str(), op.c_str());
StorePath p{realisation.outPath};
callback(userdata, outputName.c_str(), &p);
}
}
}
@@ -180,3 +182,5 @@ nix_err nix_store_copy_closure(nix_c_context * context, Store * srcStore, Store
}
NIXC_CATCH_ERRS
}
} // extern "C"

View File

@@ -148,7 +148,7 @@ void nix_store_path_free(StorePath * p);
* @param[in] path Path to check
* @return true or false, error info in context
*/
bool nix_store_is_valid_path(nix_c_context * context, Store * store, StorePath * path);
bool nix_store_is_valid_path(nix_c_context * context, Store * store, const StorePath * path);
/**
* @brief Get the physical location of a store path
@@ -190,7 +190,7 @@ nix_err nix_store_realise(
Store * store,
StorePath * path,
void * userdata,
void (*callback)(void * userdata, const char * outname, const char * out));
void (*callback)(void * userdata, const char * outname, const StorePath * out));
/**
* @brief get the version of a nix store.

View File

@@ -2,6 +2,8 @@
#define NIX_API_STORE_INTERNAL_H
#include "nix/store/store-api.hh"
extern "C" {
struct Store
{
nix::ref<nix::Store> ptr;
@@ -12,4 +14,6 @@ struct StorePath
nix::StorePath path;
};
} // extern "C"
#endif

View File

@@ -0,0 +1 @@
daemon

View File

@@ -0,0 +1 @@
local

View File

@@ -0,0 +1 @@
ssh://::1

View File

@@ -0,0 +1 @@
ssh://userinfo@fea5:23e1:3916:fc24:cb52:2837:2ecb:ea8e

View File

@@ -0,0 +1 @@
ssh://userinfo@fea5:23e1:3916:fc24:cb52:2837:2ecb:ea8e?a=b&c=d

View File

@@ -28,12 +28,33 @@ static void BM_ParseRealDerivationFile(benchmark::State & state, const std::stri
state.SetBytesProcessed(state.iterations() * content.size());
}
// Benchmark unparsing real derivation files
static void BM_UnparseRealDerivationFile(benchmark::State & state, const std::string & filename)
{
// Read the file once
std::ifstream file(filename);
std::stringstream buffer;
buffer << file.rdbuf();
std::string content = buffer.str();
auto store = openStore("dummy://");
ExperimentalFeatureSettings xpSettings;
auto drv = parseDerivation(*store, std::string(content), "test", xpSettings);
for (auto _ : state) {
auto unparsed = drv.unparse(*store, /*maskOutputs=*/false);
benchmark::DoNotOptimize(unparsed);
assert(unparsed.size() == content.size());
}
state.SetBytesProcessed(state.iterations() * content.size());
}
// Register benchmarks for actual test derivation files if they exist
BENCHMARK_CAPTURE(
BM_ParseRealDerivationFile,
hello,
getEnvNonEmpty("_NIX_TEST_UNIT_DATA").value_or(NIX_UNIT_TEST_DATA) + "/derivation/hello.drv");
BM_ParseRealDerivationFile, hello, getEnvNonEmpty("_NIX_TEST_UNIT_DATA").value() + "/derivation/hello.drv");
BENCHMARK_CAPTURE(
BM_ParseRealDerivationFile,
firefox,
getEnvNonEmpty("_NIX_TEST_UNIT_DATA").value_or(NIX_UNIT_TEST_DATA) + "/derivation/firefox.drv");
BM_ParseRealDerivationFile, firefox, getEnvNonEmpty("_NIX_TEST_UNIT_DATA").value() + "/derivation/firefox.drv");
BENCHMARK_CAPTURE(
BM_UnparseRealDerivationFile, hello, getEnvNonEmpty("_NIX_TEST_UNIT_DATA").value() + "/derivation/hello.drv");
BENCHMARK_CAPTURE(
BM_UnparseRealDerivationFile, firefox, getEnvNonEmpty("_NIX_TEST_UNIT_DATA").value() + "/derivation/firefox.drv");

View File

@@ -33,4 +33,10 @@ TEST(LocalStore, constructConfig_rootPath)
EXPECT_EQ(config.rootDir.get(), std::optional{"/foo/bar"});
}
TEST(LocalStore, constructConfig_to_string)
{
LocalStoreConfig config{"local", "", {}};
EXPECT_EQ(config.getReference().to_string(), "local");
}
} // namespace nix

View File

@@ -130,10 +130,13 @@ if get_option('benchmarks')
link_args : linker_export_flags,
install : true,
cpp_pch : do_pch ? [ 'pch/precompiled-headers.hh' ] : [],
cpp_args : [
'-DNIX_UNIT_TEST_DATA="' + meson.current_source_dir() + '/data"',
],
)
benchmark('nix-store-benchmarks', benchmark_exe)
benchmark(
'nix-store-benchmarks',
benchmark_exe,
env : {
'_NIX_TEST_UNIT_DATA' : meson.current_source_dir() / 'data',
},
)
endif

View File

@@ -1,7 +1,5 @@
#include "nix_api_util.h"
#include "nix_api_util_internal.h"
#include "nix_api_store.h"
#include "nix_api_store_internal.h"
#include "nix/store/tests/nix_api_store.hh"
#include "nix/util/tests/string_callback.hh"
@@ -65,7 +63,7 @@ TEST_F(nix_api_store_test, nix_store_get_storedir)
TEST_F(nix_api_store_test, InvalidPathFails)
{
nix_store_parse_path(ctx, store, "invalid-path");
ASSERT_EQ(ctx->last_err_code, NIX_ERR_NIX_ERROR);
ASSERT_EQ(nix_err_code(ctx), NIX_ERR_NIX_ERROR);
}
TEST_F(nix_api_store_test, ReturnsValidStorePath)
@@ -80,7 +78,7 @@ TEST_F(nix_api_store_test, ReturnsValidStorePath)
TEST_F(nix_api_store_test, SetsLastErrCodeToNixOk)
{
StorePath * path = nix_store_parse_path(ctx, store, (nixStoreDir + PATH_SUFFIX).c_str());
ASSERT_EQ(ctx->last_err_code, NIX_OK);
ASSERT_EQ(nix_err_code(ctx), NIX_OK);
nix_store_path_free(path);
}
@@ -103,7 +101,7 @@ TEST_F(nix_api_util_context, nix_store_open_dummy)
{
nix_libstore_init(ctx);
Store * store = nix_store_open(ctx, "dummy://", nullptr);
ASSERT_EQ(NIX_OK, ctx->last_err_code);
ASSERT_EQ(NIX_OK, nix_err_code(ctx));
ASSERT_STREQ("dummy://", store->ptr->config.getReference().render(/*withParams=*/true).c_str());
std::string str;
@@ -117,7 +115,7 @@ TEST_F(nix_api_util_context, nix_store_open_invalid)
{
nix_libstore_init(ctx);
Store * store = nix_store_open(ctx, "invalid://", nullptr);
ASSERT_EQ(NIX_ERR_NIX_ERROR, ctx->last_err_code);
ASSERT_EQ(NIX_ERR_NIX_ERROR, nix_err_code(ctx));
ASSERT_EQ(nullptr, store);
nix_store_free(store);
}

View File

@@ -21,7 +21,7 @@ class ParsedS3URLTest : public ::testing::WithParamInterface<ParsedS3URLTestCase
TEST_P(ParsedS3URLTest, parseS3URLSuccessfully)
{
const auto & testCase = GetParam();
auto parsed = ParsedS3URL::parse(testCase.url);
auto parsed = ParsedS3URL::parse(parseURL(testCase.url));
ASSERT_EQ(parsed, testCase.expected);
}
@@ -33,51 +33,57 @@ INSTANTIATE_TEST_SUITE_P(
"s3://my-bucket/my-key.txt",
{
.bucket = "my-bucket",
.key = "my-key.txt",
.key = {"my-key.txt"},
},
"basic_s3_bucket"},
"basic_s3_bucket",
},
ParsedS3URLTestCase{
"s3://prod-cache/nix/store/abc123.nar.xz?region=eu-west-1",
{
.bucket = "prod-cache",
.key = "nix/store/abc123.nar.xz",
.key = {"nix", "store", "abc123.nar.xz"},
.region = "eu-west-1",
},
"with_region"},
"with_region",
},
ParsedS3URLTestCase{
"s3://bucket/key?region=us-west-2&profile=prod&endpoint=custom.s3.com&scheme=https&region=us-east-1",
{
.bucket = "bucket",
.key = "key",
.key = {"key"},
.profile = "prod",
.region = "us-west-2", //< using the first parameter (decodeQuery ignores dupicates)
.scheme = "https",
.endpoint = ParsedURL::Authority{.host = "custom.s3.com"},
},
"complex"},
"complex",
},
ParsedS3URLTestCase{
"s3://cache/file.txt?profile=production&region=ap-southeast-2",
{
.bucket = "cache",
.key = "file.txt",
.key = {"file.txt"},
.profile = "production",
.region = "ap-southeast-2",
},
"with_profile_and_region"},
"with_profile_and_region",
},
ParsedS3URLTestCase{
"s3://bucket/key?endpoint=https://minio.local&scheme=http",
{
.bucket = "bucket",
.key = "key",
.key = {"key"},
/* TODO: Figure out what AWS SDK is doing when both endpointOverride and scheme are set. */
.scheme = "http",
.endpoint =
ParsedURL{
.scheme = "https",
.authority = ParsedURL::Authority{.host = "minio.local"},
.path = {""},
},
},
"with_absolute_endpoint_uri"}),
"with_absolute_endpoint_uri",
}),
[](const ::testing::TestParamInfo<ParsedS3URLTestCase> & info) { return info.param.description; });
TEST(InvalidParsedS3URLTest, parseS3URLErrors)
@@ -86,11 +92,114 @@ TEST(InvalidParsedS3URLTest, parseS3URLErrors)
testing::HasSubstrIgnoreANSIMatcher("error: URI has a missing or invalid bucket name"));
/* Empty bucket (authority) */
ASSERT_THAT([]() { ParsedS3URL::parse("s3:///key"); }, invalidBucketMatcher);
ASSERT_THAT([]() { ParsedS3URL::parse(parseURL("s3:///key")); }, invalidBucketMatcher);
/* Invalid bucket name */
ASSERT_THAT([]() { ParsedS3URL::parse("s3://127.0.0.1"); }, invalidBucketMatcher);
ASSERT_THAT([]() { ParsedS3URL::parse(parseURL("s3://127.0.0.1")); }, invalidBucketMatcher);
}
// Parameterized test for s3ToHttpsUrl conversion
struct S3ToHttpsConversionTestCase
{
ParsedS3URL input;
ParsedURL expected;
std::string expectedRendered;
std::string description;
};
class S3ToHttpsConversionTest : public ::testing::WithParamInterface<S3ToHttpsConversionTestCase>,
public ::testing::Test
{};
TEST_P(S3ToHttpsConversionTest, ConvertsCorrectly)
{
const auto & testCase = GetParam();
auto result = testCase.input.toHttpsUrl();
EXPECT_EQ(result, testCase.expected) << "Failed for: " << testCase.description;
EXPECT_EQ(result.to_string(), testCase.expectedRendered);
}
INSTANTIATE_TEST_SUITE_P(
S3ToHttpsConversion,
S3ToHttpsConversionTest,
::testing::Values(
S3ToHttpsConversionTestCase{
ParsedS3URL{
.bucket = "my-bucket",
.key = {"my-key.txt"},
},
ParsedURL{
.scheme = "https",
.authority = ParsedURL::Authority{.host = "s3.us-east-1.amazonaws.com"},
.path = {"", "my-bucket", "my-key.txt"},
},
"https://s3.us-east-1.amazonaws.com/my-bucket/my-key.txt",
"basic_s3_default_region",
},
S3ToHttpsConversionTestCase{
ParsedS3URL{
.bucket = "prod-cache",
.key = {"nix", "store", "abc123.nar.xz"},
.region = "eu-west-1",
},
ParsedURL{
.scheme = "https",
.authority = ParsedURL::Authority{.host = "s3.eu-west-1.amazonaws.com"},
.path = {"", "prod-cache", "nix", "store", "abc123.nar.xz"},
},
"https://s3.eu-west-1.amazonaws.com/prod-cache/nix/store/abc123.nar.xz",
"with_eu_west_1_region",
},
S3ToHttpsConversionTestCase{
ParsedS3URL{
.bucket = "bucket",
.key = {"key"},
.scheme = "http",
.endpoint = ParsedURL::Authority{.host = "custom.s3.com"},
},
ParsedURL{
.scheme = "http",
.authority = ParsedURL::Authority{.host = "custom.s3.com"},
.path = {"", "bucket", "key"},
},
"http://custom.s3.com/bucket/key",
"custom_endpoint_authority",
},
S3ToHttpsConversionTestCase{
ParsedS3URL{
.bucket = "bucket",
.key = {"key"},
.endpoint =
ParsedURL{
.scheme = "http",
.authority = ParsedURL::Authority{.host = "server", .port = 9000},
.path = {""},
},
},
ParsedURL{
.scheme = "http",
.authority = ParsedURL::Authority{.host = "server", .port = 9000},
.path = {"", "bucket", "key"},
},
"http://server:9000/bucket/key",
"custom_endpoint_with_port",
},
S3ToHttpsConversionTestCase{
ParsedS3URL{
.bucket = "bucket",
.key = {"path", "to", "file.txt"},
.region = "ap-southeast-2",
.scheme = "https",
},
ParsedURL{
.scheme = "https",
.authority = ParsedURL::Authority{.host = "s3.ap-southeast-2.amazonaws.com"},
.path = {"", "bucket", "path", "to", "file.txt"},
},
"https://s3.ap-southeast-2.amazonaws.com/bucket/path/to/file.txt",
"complex_path_and_region",
}),
[](const ::testing::TestParamInfo<S3ToHttpsConversionTestCase> & info) { return info.param.description; });
} // namespace nix
#endif

View File

@@ -107,6 +107,13 @@ URI_TEST_READ(local_shorthand_1, localExample_1)
URI_TEST_READ(local_shorthand_2, localExample_2)
URI_TEST(
local_shorthand_3,
(StoreReference{
.variant = StoreReference::Local{},
.params = {},
}))
static StoreReference unixExample{
.variant =
StoreReference::Specified{
@@ -134,4 +141,46 @@ URI_TEST(
.params = {},
}))
URI_TEST(
daemon_shorthand,
(StoreReference{
.variant = StoreReference::Daemon{},
.params = {},
}))
static StoreReference sshLoopbackIPv6{
.variant =
StoreReference::Specified{
.scheme = "ssh",
.authority = "[::1]",
},
};
URI_TEST_READ(ssh_unbracketed_ipv6_1, sshLoopbackIPv6)
static StoreReference sshIPv6AuthorityWithUserinfo{
.variant =
StoreReference::Specified{
.scheme = "ssh",
.authority = "userinfo@[fea5:23e1:3916:fc24:cb52:2837:2ecb:ea8e]",
},
};
URI_TEST_READ(ssh_unbracketed_ipv6_2, sshIPv6AuthorityWithUserinfo)
static StoreReference sshIPv6AuthorityWithUserinfoAndParams{
.variant =
StoreReference::Specified{
.scheme = "ssh",
.authority = "userinfo@[fea5:23e1:3916:fc24:cb52:2837:2ecb:ea8e]",
},
.params =
{
{"a", "b"},
{"c", "d"},
},
};
URI_TEST_READ(ssh_unbracketed_ipv6_3, sshIPv6AuthorityWithUserinfoAndParams)
} // namespace nix

View File

@@ -16,4 +16,10 @@ TEST(UDSRemoteStore, constructConfigWrongScheme)
EXPECT_THROW(UDSRemoteStoreConfig("http", "/tmp/socket", {}), UsageError);
}
TEST(UDSRemoteStore, constructConfig_to_string)
{
UDSRemoteStoreConfig config{"unix", "", {}};
EXPECT_EQ(config.getReference().to_string(), "daemon");
}
} // namespace nix

View File

@@ -125,11 +125,8 @@ void BinaryCacheStore::writeNarInfo(ref<NarInfo> narInfo)
upsertFile(narInfoFile, narInfo->to_string(*this), "text/x-nix-narinfo");
{
auto state_(state.lock());
state_->pathInfoCache.upsert(
std::string(narInfo->path.to_string()), PathInfoCacheValue{.value = std::shared_ptr<NarInfo>(narInfo)});
}
pathInfoCache->lock()->upsert(
std::string(narInfo->path.to_string()), PathInfoCacheValue{.value = std::shared_ptr<NarInfo>(narInfo)});
if (diskCache)
diskCache->upsertNarInfo(

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,190 @@
#include <queue>
#include "nix/store/store-api.hh"
#include "nix/store/build-result.hh"
#include "derivation-check.hh"
namespace nix {
void checkOutputs(
Store & store,
const StorePath & drvPath,
const decltype(Derivation::outputs) & drvOutputs,
const decltype(DerivationOptions::outputChecks) & outputChecks,
const std::map<std::string, ValidPathInfo> & outputs)
{
std::map<Path, const ValidPathInfo &> outputsByPath;
for (auto & output : outputs)
outputsByPath.emplace(store.printStorePath(output.second.path), output.second);
for (auto & [outputName, info] : outputs) {
auto * outputSpec = get(drvOutputs, outputName);
assert(outputSpec);
if (const auto * dof = std::get_if<DerivationOutput::CAFixed>(&outputSpec->raw)) {
auto & wanted = dof->ca.hash;
/* Check wanted hash */
assert(info.ca);
auto & got = info.ca->hash;
if (wanted != got) {
/* Throw an error after registering the path as
valid. */
throw BuildError(
BuildResult::HashMismatch,
"hash mismatch in fixed-output derivation '%s':\n specified: %s\n got: %s",
store.printStorePath(drvPath),
wanted.to_string(HashFormat::SRI, true),
got.to_string(HashFormat::SRI, true));
}
if (!info.references.empty()) {
auto numViolations = info.references.size();
throw BuildError(
BuildResult::HashMismatch,
"fixed-output derivations must not reference store paths: '%s' references %d distinct paths, e.g. '%s'",
store.printStorePath(drvPath),
numViolations,
store.printStorePath(*info.references.begin()));
}
}
/* Compute the closure and closure size of some output. This
is slightly tricky because some of its references (namely
other outputs) may not be valid yet. */
auto getClosure = [&](const StorePath & path) {
uint64_t closureSize = 0;
StorePathSet pathsDone;
std::queue<StorePath> pathsLeft;
pathsLeft.push(path);
while (!pathsLeft.empty()) {
auto path = pathsLeft.front();
pathsLeft.pop();
if (!pathsDone.insert(path).second)
continue;
auto i = outputsByPath.find(store.printStorePath(path));
if (i != outputsByPath.end()) {
closureSize += i->second.narSize;
for (auto & ref : i->second.references)
pathsLeft.push(ref);
} else {
auto info = store.queryPathInfo(path);
closureSize += info->narSize;
for (auto & ref : info->references)
pathsLeft.push(ref);
}
}
return std::make_pair(std::move(pathsDone), closureSize);
};
auto applyChecks = [&](const DerivationOptions::OutputChecks & checks) {
if (checks.maxSize && info.narSize > *checks.maxSize)
throw BuildError(
BuildResult::OutputRejected,
"path '%s' is too large at %d bytes; limit is %d bytes",
store.printStorePath(info.path),
info.narSize,
*checks.maxSize);
if (checks.maxClosureSize) {
uint64_t closureSize = getClosure(info.path).second;
if (closureSize > *checks.maxClosureSize)
throw BuildError(
BuildResult::OutputRejected,
"closure of path '%s' is too large at %d bytes; limit is %d bytes",
store.printStorePath(info.path),
closureSize,
*checks.maxClosureSize);
}
auto checkRefs = [&](const StringSet & value, bool allowed, bool recursive) {
/* Parse a list of reference specifiers. Each element must
either be a store path, or the symbolic name of the output
of the derivation (such as `out'). */
StorePathSet spec;
for (auto & i : value) {
if (store.isStorePath(i))
spec.insert(store.parseStorePath(i));
else if (auto output = get(outputs, i))
spec.insert(output->path);
else {
std::string outputsListing =
concatMapStringsSep(", ", outputs, [](auto & o) { return o.first; });
throw BuildError(
BuildResult::OutputRejected,
"derivation '%s' output check for '%s' contains an illegal reference specifier '%s',"
" expected store path or output name (one of [%s])",
store.printStorePath(drvPath),
outputName,
i,
outputsListing);
}
}
auto used = recursive ? getClosure(info.path).first : info.references;
if (recursive && checks.ignoreSelfRefs)
used.erase(info.path);
StorePathSet badPaths;
for (auto & i : used)
if (allowed) {
if (!spec.count(i))
badPaths.insert(i);
} else {
if (spec.count(i))
badPaths.insert(i);
}
if (!badPaths.empty()) {
std::string badPathsStr;
for (auto & i : badPaths) {
badPathsStr += "\n ";
badPathsStr += store.printStorePath(i);
}
throw BuildError(
BuildResult::OutputRejected,
"output '%s' is not allowed to refer to the following paths:%s",
store.printStorePath(info.path),
badPathsStr);
}
};
/* Mandatory check: absent whitelist, and present but empty
whitelist mean very different things. */
if (auto & refs = checks.allowedReferences) {
checkRefs(*refs, true, false);
}
if (auto & refs = checks.allowedRequisites) {
checkRefs(*refs, true, true);
}
/* Optimization: don't need to do anything when
disallowed and empty set. */
if (!checks.disallowedReferences.empty()) {
checkRefs(checks.disallowedReferences, false, false);
}
if (!checks.disallowedRequisites.empty()) {
checkRefs(checks.disallowedRequisites, false, true);
}
};
std::visit(
overloaded{
[&](const DerivationOptions::OutputChecks & checks) { applyChecks(checks); },
[&](const std::map<std::string, DerivationOptions::OutputChecks> & checksPerOutput) {
if (auto outputChecks = get(checksPerOutput, outputName))
applyChecks(*outputChecks);
},
},
outputChecks);
}
}
} // namespace nix

View File

@@ -0,0 +1,27 @@
#pragma once
///@file
#include "nix/store/derivations.hh"
#include "nix/store/derivation-options.hh"
#include "nix/store/path-info.hh"
namespace nix {
/**
* Check that outputs meets the requirements specified by the
* 'outputChecks' attribute (or the legacy
* '{allowed,disallowed}{References,Requisites}' attributes).
*
* The outputs may not be valid yet, hence outputs needs to contain all
* needed info like the NAR size. However, the external (not other
* output) references of the output must be valid, so we can compute the
* closure size.
*/
void checkOutputs(
Store & store,
const StorePath & drvPath,
const decltype(Derivation::outputs) & drvOutputs,
const decltype(DerivationOptions::outputChecks) & drvOptions,
const std::map<std::string, ValidPathInfo> & outputs);
} // namespace nix

View File

@@ -0,0 +1,59 @@
#include "nix/store/build/derivation-env-desugar.hh"
#include "nix/store/store-api.hh"
#include "nix/store/derivations.hh"
#include "nix/store/derivation-options.hh"
namespace nix {
std::string & DesugaredEnv::atFileEnvPair(std::string_view name, std::string fileName)
{
auto & ret = extraFiles[fileName];
variables.insert_or_assign(
std::string{name},
EnvEntry{
.prependBuildDirectory = true,
.value = std::move(fileName),
});
return ret;
}
DesugaredEnv DesugaredEnv::create(
Store & store, const Derivation & drv, const DerivationOptions & drvOptions, const StorePathSet & inputPaths)
{
DesugaredEnv res;
if (drv.structuredAttrs) {
auto json = drv.structuredAttrs->prepareStructuredAttrs(store, drvOptions, inputPaths, drv.outputs);
res.atFileEnvPair("NIX_ATTRS_SH_FILE", ".attrs.sh") = StructuredAttrs::writeShell(json);
res.atFileEnvPair("NIX_ATTRS_JSON_FILE", ".attrs.json") = json.dump();
} else {
/* In non-structured mode, set all bindings either directory in the
environment or via a file, as specified by
`DerivationOptions::passAsFile`. */
for (auto & [envName, envValue] : drv.env) {
if (!drvOptions.passAsFile.contains(envName)) {
res.variables.insert_or_assign(
envName,
EnvEntry{
.value = envValue,
});
} else {
res.atFileEnvPair(
envName + "Path",
".attr-" + hashString(HashAlgorithm::SHA256, envName).to_string(HashFormat::Nix32, false)) =
envValue;
}
}
/* Handle exportReferencesGraph(), if set. */
for (auto & [fileName, storePaths] : drvOptions.getParsedExportReferencesGraph(store)) {
/* Write closure info to <fileName>. */
res.extraFiles.insert_or_assign(
fileName, store.makeValidityRegistration(store.exportReferences(storePaths, inputPaths), false, false));
}
}
return res;
}
} // namespace nix

View File

@@ -94,7 +94,7 @@ Goal::Co DerivationGoal::haveDerivation()
/* If they are all valid, then we're done. */
if (checkResult && checkResult->second == PathStatus::Valid && buildMode == bmNormal) {
co_return done(BuildResult::AlreadyValid, checkResult->first);
co_return doneSuccess(BuildResult::AlreadyValid, checkResult->first);
}
Goals waitees;
@@ -122,12 +122,10 @@ Goal::Co DerivationGoal::haveDerivation()
assert(!drv->type().isImpure());
if (nrFailed > 0 && nrFailed > nrNoSubstituters && !settings.tryFallback) {
co_return done(
co_return doneFailure(BuildError(
BuildResult::TransientFailure,
{},
Error(
"some substitutes for the outputs of derivation '%s' failed (usually happens due to networking issues); try '--fallback' to build derivation from source ",
worker.store.printStorePath(drvPath)));
"some substitutes for the outputs of derivation '%s' failed (usually happens due to networking issues); try '--fallback' to build derivation from source ",
worker.store.printStorePath(drvPath)));
}
nrFailed = nrNoSubstituters = 0;
@@ -137,7 +135,7 @@ Goal::Co DerivationGoal::haveDerivation()
bool allValid = checkResult && checkResult->second == PathStatus::Valid;
if (buildMode == bmNormal && allValid) {
co_return done(BuildResult::Substituted, checkResult->first);
co_return doneSuccess(BuildResult::Substituted, checkResult->first);
}
if (buildMode == bmRepair && allValid) {
co_return repairClosure();
@@ -281,7 +279,7 @@ Goal::Co DerivationGoal::repairClosure()
"some paths in the output closure of derivation '%s' could not be repaired",
worker.store.printStorePath(drvPath));
}
co_return done(BuildResult::AlreadyValid, assertPathValidity());
co_return doneSuccess(BuildResult::AlreadyValid, assertPathValidity());
}
std::optional<std::pair<Realisation, PathStatus>> DerivationGoal::checkPathValidity()
@@ -339,12 +337,27 @@ Realisation DerivationGoal::assertPathValidity()
return checkResult->first;
}
Goal::Done
DerivationGoal::done(BuildResult::Status status, std::optional<Realisation> builtOutput, std::optional<Error> ex)
Goal::Done DerivationGoal::doneSuccess(BuildResult::Status status, Realisation builtOutput)
{
buildResult.status = status;
if (ex)
buildResult.errorMsg = fmt("%s", Uncolored(ex->info().msg));
assert(buildResult.success());
mcExpectedBuilds.reset();
buildResult.builtOutputs = {{wantedOutput, std::move(builtOutput)}};
if (status == BuildResult::Built)
worker.doneBuilds++;
worker.updateProgress();
return amDone(ecSuccess, std::nullopt);
}
Goal::Done DerivationGoal::doneFailure(BuildError ex)
{
buildResult.status = ex.status;
buildResult.errorMsg = fmt("%s", Uncolored(ex.info().msg));
if (buildResult.status == BuildResult::TimedOut)
worker.timedOut = true;
if (buildResult.status == BuildResult::PermanentFailure)
@@ -352,26 +365,12 @@ DerivationGoal::done(BuildResult::Status status, std::optional<Realisation> buil
mcExpectedBuilds.reset();
if (buildResult.success()) {
assert(builtOutput);
buildResult.builtOutputs = {{wantedOutput, std::move(*builtOutput)}};
if (status == BuildResult::Built)
worker.doneBuilds++;
} else {
if (status != BuildResult::DependencyFailed)
worker.failedBuilds++;
}
if (ex.status != BuildResult::DependencyFailed)
worker.failedBuilds++;
worker.updateProgress();
auto traceBuiltOutputsFile = getEnv("_NIX_TRACE_BUILT_OUTPUTS").value_or("");
if (traceBuiltOutputsFile != "") {
std::fstream fs;
fs.open(traceBuiltOutputsFile, std::fstream::out);
fs << worker.store.printStorePath(drvPath) << "\t" << buildResult.toString() << std::endl;
}
return amDone(buildResult.success() ? ecSuccess : ecFailed, std::move(ex));
return amDone(ecFailed, {std::move(ex)});
}
} // namespace nix

View File

@@ -37,7 +37,7 @@ static void builtinFetchurl(const BuiltinBuilderContext & ctx)
auto fetch = [&](const std::string & url) {
auto source = sinkToSource([&](Sink & sink) {
FileTransferRequest request(url);
FileTransferRequest request(ValidURL{url});
request.decompress = false;
auto decompressor = makeDecompressionSink(unpack && hasSuffix(mainUrl, ".xz") ? "xz" : "none", sink);

View File

@@ -546,47 +546,22 @@ static void performOp(
break;
}
case WorkerProto::Op::ExportPath: {
auto path = store->parseStorePath(readString(conn.from));
readInt(conn.from); // obsolete
logger->startWork();
TunnelSink sink(conn.to);
store->exportPath(path, sink);
logger->stopWork();
conn.to << 1;
break;
}
case WorkerProto::Op::ImportPaths: {
logger->startWork();
TunnelSource source(conn.from, conn.to);
auto paths = store->importPaths(source, trusted ? NoCheckSigs : CheckSigs);
logger->stopWork();
Strings paths2;
for (auto & i : paths)
paths2.push_back(store->printStorePath(i));
conn.to << paths2;
break;
}
case WorkerProto::Op::BuildPaths: {
auto drvs = WorkerProto::Serialise<DerivedPaths>::read(*store, rconn);
BuildMode mode = bmNormal;
if (GET_PROTOCOL_MINOR(conn.protoVersion) >= 15) {
mode = WorkerProto::Serialise<BuildMode>::read(*store, rconn);
mode = WorkerProto::Serialise<BuildMode>::read(*store, rconn);
/* Repairing is not atomic, so disallowed for "untrusted"
clients.
/* Repairing is not atomic, so disallowed for "untrusted"
clients.
FIXME: layer violation in this message: the daemon code (i.e.
this file) knows whether a client/connection is trusted, but it
does not how how the client was authenticated. The mechanism
need not be getting the UID of the other end of a Unix Domain
Socket.
*/
if (mode == bmRepair && !trusted)
throw Error("repairing is not allowed because you are not in 'trusted-users'");
}
FIXME: layer violation in this message: the daemon code (i.e.
this file) knows whether a client/connection is trusted, but it
does not how how the client was authenticated. The mechanism
need not be getting the UID of the other end of a Unix Domain
Socket.
*/
if (mode == bmRepair && !trusted)
throw Error("repairing is not allowed because you are not in 'trusted-users'");
logger->startWork();
store->buildPaths(drvs, mode);
logger->stopWork();
@@ -805,13 +780,11 @@ static void performOp(
clientSettings.buildCores = readInt(conn.from);
clientSettings.useSubstitutes = readInt(conn.from);
if (GET_PROTOCOL_MINOR(conn.protoVersion) >= 12) {
unsigned int n = readInt(conn.from);
for (unsigned int i = 0; i < n; i++) {
auto name = readString(conn.from);
auto value = readString(conn.from);
clientSettings.overrides.emplace(name, value);
}
unsigned int n = readInt(conn.from);
for (unsigned int i = 0; i < n; i++) {
auto name = readString(conn.from);
auto value = readString(conn.from);
clientSettings.overrides.emplace(name, value);
}
logger->startWork();
@@ -876,19 +849,12 @@ static void performOp(
auto path = store->parseStorePath(readString(conn.from));
std::shared_ptr<const ValidPathInfo> info;
logger->startWork();
try {
info = store->queryPathInfo(path);
} catch (InvalidPath &) {
if (GET_PROTOCOL_MINOR(conn.protoVersion) < 17)
throw;
}
info = store->queryPathInfo(path);
logger->stopWork();
if (info) {
if (GET_PROTOCOL_MINOR(conn.protoVersion) >= 17)
conn.to << 1;
conn.to << 1;
WorkerProto::write(*store, wconn, static_cast<const UnkeyedValidPathInfo &>(*info));
} else {
assert(GET_PROTOCOL_MINOR(conn.protoVersion) >= 17);
conn.to << 0;
}
break;
@@ -1063,7 +1029,7 @@ void processConnection(ref<Store> store, FdSource && from, FdSink && to, Trusted
auto [protoVersion, features] =
WorkerProto::BasicServerConnection::handshake(to, from, PROTOCOL_VERSION, WorkerProto::allFeatures);
if (protoVersion < 0x10a)
if (protoVersion < 256 + 18)
throw Error("the Nix client version is too old");
WorkerProto::BasicServerConnection conn;

View File

@@ -265,7 +265,8 @@ DerivationOptions::getParsedExportReferencesGraph(const StoreDirConfig & store)
StorePathSet storePaths;
for (auto & storePathS : ss) {
if (!store.isInStore(storePathS))
throw BuildError("'exportReferencesGraph' contains a non-store path '%1%'", storePathS);
throw BuildError(
BuildResult::InputRejected, "'exportReferencesGraph' contains a non-store path '%1%'", storePathS);
storePaths.insert(store.toStorePath(storePathS).first);
}
res.insert_or_assign(fileName, storePaths);

View File

@@ -498,28 +498,33 @@ Derivation parseDerivation(
*/
static void printString(std::string & res, std::string_view s)
{
boost::container::small_vector<char, 64 * 1024> buffer;
buffer.reserve(s.size() * 2 + 2);
char * buf = buffer.data();
char * p = buf;
*p++ = '"';
for (auto c : s)
if (c == '\"' || c == '\\') {
*p++ = '\\';
*p++ = c;
} else if (c == '\n') {
*p++ = '\\';
*p++ = 'n';
} else if (c == '\r') {
*p++ = '\\';
*p++ = 'r';
} else if (c == '\t') {
*p++ = '\\';
*p++ = 't';
} else
*p++ = c;
*p++ = '"';
res.append(buf, p - buf);
res.reserve(res.size() + s.size() * 2 + 2);
res += '"';
static constexpr auto chunkSize = 1024;
std::array<char, 2 * chunkSize + 2> buffer;
while (!s.empty()) {
auto chunk = s.substr(0, /*n=*/chunkSize);
s.remove_prefix(chunk.size());
char * buf = buffer.data();
char * p = buf;
for (auto c : chunk)
if (c == '\"' || c == '\\') {
*p++ = '\\';
*p++ = c;
} else if (c == '\n') {
*p++ = '\\';
*p++ = 'n';
} else if (c == '\r') {
*p++ = '\\';
*p++ = 'r';
} else if (c == '\t') {
*p++ = '\\';
*p++ = 't';
} else
*p++ = c;
res.append(buf, p - buf);
}
res += '"';
}
static void printUnquotedString(std::string & res, std::string_view s)

View File

@@ -100,7 +100,7 @@ struct curlFileTransfer : public FileTransfer
lvlTalkative,
actFileTransfer,
fmt("%sing '%s'", request.verb(), request.uri),
{request.uri},
{request.uri.to_string()},
request.parentAct)
, callback(std::move(callback))
, finalSink([this](std::string_view data) {
@@ -121,7 +121,7 @@ struct curlFileTransfer : public FileTransfer
this->result.data.append(data);
})
{
result.urls.push_back(request.uri);
result.urls.push_back(request.uri.to_string());
requestHeaders = curl_slist_append(requestHeaders, "Accept-Encoding: zstd, br, gzip, deflate, bzip2, xz");
if (!request.expectedETag.empty())
@@ -350,7 +350,7 @@ struct curlFileTransfer : public FileTransfer
curl_easy_setopt(req, CURLOPT_DEBUGFUNCTION, TransferItem::debugCallback);
}
curl_easy_setopt(req, CURLOPT_URL, request.uri.c_str());
curl_easy_setopt(req, CURLOPT_URL, request.uri.to_string().c_str());
curl_easy_setopt(req, CURLOPT_FOLLOWLOCATION, 1L);
curl_easy_setopt(req, CURLOPT_MAXREDIRS, 10);
curl_easy_setopt(req, CURLOPT_NOSIGNAL, 1);
@@ -784,8 +784,8 @@ struct curlFileTransfer : public FileTransfer
void enqueueItem(std::shared_ptr<TransferItem> item)
{
if (item->request.data && !hasPrefix(item->request.uri, "http://") && !hasPrefix(item->request.uri, "https://"))
throw nix::Error("uploading to '%s' is not supported", item->request.uri);
if (item->request.data && item->request.uri.scheme() != "http" && item->request.uri.scheme() != "https")
throw nix::Error("uploading to '%s' is not supported", item->request.uri.to_string());
{
auto state(state_.lock());
@@ -801,11 +801,11 @@ struct curlFileTransfer : public FileTransfer
void enqueueFileTransfer(const FileTransferRequest & request, Callback<FileTransferResult> callback) override
{
/* Ugly hack to support s3:// URIs. */
if (hasPrefix(request.uri, "s3://")) {
if (request.uri.scheme() == "s3") {
// FIXME: do this on a worker thread
try {
#if NIX_WITH_S3_SUPPORT
auto parsed = ParsedS3URL::parse(request.uri);
auto parsed = ParsedS3URL::parse(request.uri.parsed());
std::string profile = parsed.profile.value_or("");
std::string region = parsed.region.value_or(Aws::Region::US_EAST_1);
@@ -815,15 +815,16 @@ struct curlFileTransfer : public FileTransfer
S3Helper s3Helper(profile, region, scheme, endpoint);
// FIXME: implement ETag
auto s3Res = s3Helper.getObject(parsed.bucket, parsed.key);
auto s3Res = s3Helper.getObject(parsed.bucket, encodeUrlPath(parsed.key));
FileTransferResult res;
if (!s3Res.data)
throw FileTransferError(NotFound, {}, "S3 object '%s' does not exist", request.uri);
res.data = std::move(*s3Res.data);
res.urls.push_back(request.uri);
res.urls.push_back(request.uri.to_string());
callback(std::move(res));
#else
throw nix::Error("cannot download '%s' because Nix is not built with S3 support", request.uri);
throw nix::Error(
"cannot download '%s' because Nix is not built with S3 support", request.uri.to_string());
#endif
} catch (...) {
callback.rethrow();

View File

@@ -931,7 +931,7 @@ void LocalStore::autoGC(bool sync)
std::shared_future<void> future;
{
auto state(_state.lock());
auto state(_state->lock());
if (state->gcRunning) {
future = state->gcFuture;
@@ -964,7 +964,7 @@ void LocalStore::autoGC(bool sync)
/* Wake up any threads waiting for the auto-GC to finish. */
Finally wakeup([&]() {
auto state(_state.lock());
auto state(_state->lock());
state->gcRunning = false;
state->lastGCCheck = std::chrono::steady_clock::now();
promise.set_value();
@@ -979,7 +979,7 @@ void LocalStore::autoGC(bool sync)
collectGarbage(options, results);
_state.lock()->availAfterGC = getAvail();
_state->lock()->availAfterGC = getAvail();
} catch (...) {
// FIXME: we could propagate the exception to the

View File

@@ -27,7 +27,7 @@ HttpBinaryCacheStoreConfig::HttpBinaryCacheStoreConfig(
+ (!_cacheUri.empty() ? _cacheUri
: throw UsageError("`%s` Store requires a non-empty authority in Store URL", scheme))))
{
while (!cacheUri.path.empty() && cacheUri.path.back() == '/')
while (!cacheUri.path.empty() && cacheUri.path.back() == "")
cacheUri.path.pop_back();
}
@@ -37,7 +37,7 @@ StoreReference HttpBinaryCacheStoreConfig::getReference() const
.variant =
StoreReference::Specified{
.scheme = cacheUri.scheme,
.authority = (cacheUri.authority ? cacheUri.authority->to_string() : "") + cacheUri.path,
.authority = cacheUri.renderAuthorityAndPath(),
},
.params = cacheUri.query,
};
@@ -154,22 +154,17 @@ protected:
FileTransferRequest makeRequest(const std::string & path)
{
/* FIXME path is not a path, but a full relative or absolute
/* Otherwise the last path fragment will get discarded. */
auto cacheUriWithTrailingSlash = config->cacheUri;
if (!cacheUriWithTrailingSlash.path.empty())
cacheUriWithTrailingSlash.path.push_back("");
/* path is not a path, but a full relative or absolute
URL, e.g. we've seen in the wild NARINFO files have a URL
field which is
`nar/15f99rdaf26k39knmzry4xd0d97wp6yfpnfk1z9avakis7ipb9yg.nar?hash=zphkqn2wg8mnvbkixnl2aadkbn0rcnfj`
(note the query param) and that gets passed here.
What should actually happen is that we have two parsed URLs
(if we support relative URLs), and then we combined them with
a URL `operator/` which would be like
`std::filesystem::path`'s equivalent operator, which properly
combines the the URLs, whether the right is relative or
absolute. */
return FileTransferRequest(
hasPrefix(path, "https://") || hasPrefix(path, "http://") || hasPrefix(path, "file://")
? path
: config->cacheUri.to_string() + "/" + path);
(note the query param) and that gets passed here. */
return FileTransferRequest(parseURLRelative(path, cacheUriWithTrailingSlash));
}
void getFile(const std::string & path, Sink & sink) override

Some files were not shown because too many files have changed in this diff Show More