Commit Graph
960 Commits
Author SHA1 Message Date
Eelco Dolstra 00d9e7e1f4 EvalState::resolveLookupPathPath(): Call resolveSymlinks() before pathExists()
Fixes #12339.
2025-01-24 12:15:07 +01:00
Eelco Dolstra ff9d886f3c Use isAbsolute() 2025-01-14 17:42:26 +01:00
Philipp Otterbein a44e9dd1ea correctly parse strings with null bytes and throw error 2025-01-04 16:14:06 +01:00
Eelco Dolstra 08361f031d EvalState::realiseContext(): Allow access to the entire closure
Fixes #11030.
2024-12-16 16:16:36 +01:00
Philipp Otterbein 3a9d64b8e3 fromJSON/fromTOML: throw if string contains null byte 2024-12-09 22:04:21 +01:00
Sergei Zimmerman 1800853b2a fix(libexpr/eval-inline): get rid of references to nullptr env
When diagnosing infinite recursion references to nullptr `Env` can be formed.
This happens only with `ExprBlackHole` is evaluated, which always leads to
`InfiniteRecursionError`.

UBSAN log for one such case:

```
../src/libexpr/eval-inline.hh:94:31: runtime error: reference binding to null pointer of type 'Env'
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior ../src/libexpr/eval-inline.hh:94:31 in
```
2024-11-20 17:54:58 +03:00
Robert Hensing 32becc87fe Merge pull request #11914 from roberth/evalstate-get-builtins
EvalState::getBuiltins
2024-11-19 19:33:37 +01:00
Robert Hensing 5c258d7e25 refactor: Use EvalState::getBuiltins() 2024-11-19 18:45:53 +01:00
Robert Hensing 8a36d2d8a7 Add EvalState::getBuiltins 2024-11-19 18:23:05 +01:00
Eelco Dolstra af07f33d37 resolveLookupPathPath(): Return a SourcePath instead of a string
Cherry-picked from lazy-trees.
2024-11-19 18:03:31 +01:00
Robert Hensing a58e38dab7 Make EvalState::getBuiltin safe for missing attr 2024-11-19 17:30:58 +01:00
Eelco Dolstra 3e4a83f53b Use range-based for 2024-11-14 16:12:14 +01:00
Eelco Dolstra a2e4a4c238 callFunction: Use std::span
This is a bit safer than having a separate nrArgs argument.
2024-11-12 19:26:39 +01:00
Eelco Dolstra 67a42e6240 Merge remote-tracking branch 'origin/master' into flake-substitution 2024-11-11 13:57:24 +01:00
Robert Hensing a3613f2a3c autoCallFunction: accept const Bindings &
It does not need to mutate the attrs.
`const` is shallow.
Avoids a const_cast in the hercules-ci-cnix-expr bindings.
2024-11-10 20:52:05 +01:00
Eelco Dolstra 4dceca51de Don't allow __final in fetchTree
It's now only allowed in fetchFinalTree, which is not exposed to users
but only to call-flake.nix.
2024-11-08 19:27:54 +01:00
Brian McKenna d711c7e965 Fix compilation of eval under Windows 2024-11-04 19:32:41 +11:00
Philipp Otterbein de96f632f8 std::string_view shall not be null terminated 2024-10-08 02:25:14 +02:00
Philipp Otterbein e21c7895eb MacOS built: add workaround for missing view() member of std::ostringstream 2024-10-07 02:05:53 +02:00
Philipp Otterbein 4c0c8e5428 cleanup: remove superfluous std::string copies 2024-10-06 13:43:29 +02:00
Philipp Otterbein 5cf6b2cb75 eval: remove superfluous strdup 2024-09-30 22:15:04 +02:00
Eelco Dolstra 2f4a7a8301 Add a few more aliases 2024-09-19 21:04:01 +02:00
Eelco Dolstra 589d8f1f2b Move GC-related definitions to eval-gc.hh 2024-09-19 21:04:01 +02:00
Eelco Dolstra b9f78abb7f Alias traceable_allocator to std::allocator when building without GC
This allows us to get rid of a bunch of #ifdefs.
2024-09-19 21:04:01 +02:00
Robert Hensing 88998fae74 Merge pull request #11304 from hercules-ci/repl-doc-functor
`:doc`: support `__functor`
2024-08-26 17:20:37 +02:00
Robert Hensing 77ddcbe12e getDoc: Explain why we partially apply __functor 2024-08-26 16:15:13 +02:00
Robert Hensing 72a4d1f52d Add :doc support for __functor 2024-08-15 13:04:34 +02:00
Robert Hensing 6068e32aa7 refactor: Extract EvalState::addCallDepth 2024-08-15 13:04:34 +02:00
Eelco Dolstra 9f6ee93f48 fetchers::downloadTarball(): Return a cacheable accessor
downloadTarball() is used by `-I foo=<url>` etc. fetchToStore() needs
the accessor to have a fingerprint to enable caching.

Fixes #11271.
2024-08-12 15:48:26 +02:00
Robert Hensing 18485d2d53 Merge pull request #11188 from lf-/jade/kill-int-overflow
Ban integer overflow in the Nix language
2024-08-11 04:24:16 +02:00
John Ericson 733c816d34 Small windows cross fixes (#11230) 2024-07-31 20:04:18 +00:00
Jade Lovelace 7b6622d733 language: cleanly ban integer overflows
This also bans various sneaking of negative numbers from the language
into unsuspecting builtins as was exposed while auditing the
consequences of changing the Nix language integer type to a newtype.

It's unlikely that this change comprehensively ensures correctness when
passing integers out of the Nix language and we should probably add a
checked-narrowing function or something similar, but that's out of scope
for the immediate change.

During the development of this I found a few fun facts about the
language:
- You could overflow integers by converting from unsigned JSON values.
- You could overflow unsigned integers by converting negative numbers
  into them when going into Nix config, into fetchTree, and into flake
  inputs.

  The flake inputs and Nix config cannot actually be tested properly
  since they both ban thunks, however, we put in checks anyway because
  it's possible these could somehow be used to do such shenanigans some
  other way.

Note that Lix has banned Nix language integer overflows since the very
first public beta, but threw a SIGILL about them because we run with
-fsanitize=signed-overflow -fsanitize-undefined-trap-on-error in
production builds. Since the Nix language uses signed integers, overflow
was simply undefined behaviour, and since we defined that to trap, it
did.

Trapping on it was a bad UX, but we didn't even entirely notice
that we had done this at all until it was reported as a bug a couple of
months later (which is, to be fair, that flag working as intended), and
it's got enough production time that, aside from code that is IMHO buggy
(and which is, in any case, not in nixpkgs) such as
https://git.lix.systems/lix-project/lix/issues/445, we don't think
anyone doing anything reasonable actually depends on wrapping overflow.

Even for weird use cases such as doing funny bit crimes, it doesn't make
sense IMO to have wrapping behaviour, since two's complement arithmetic
overflow behaviour is so *aggressively* not what you want for *any* kind
of mathematics/algorithms. The Nix language exists for package
management, a domain where bit crimes are already only dubiously in
scope to begin with, and it makes a lot more sense for that domain for
the integers to never lose precision, either by throwing errors if they
would, or by being arbitrary-precision.

Fixes: https://github.com/NixOS/nix/issues/10968
Original-CL: https://gerrit.lix.systems/c/lix/+/1596

Change-Id: I51f253840c4af2ea5422b8a420aa5fafbf8fae75
2024-07-30 18:13:05 -07:00
Robert Hensing 861bd102a6 Merge pull request #11167 from NixOS/repl-test-rejiggle
Fix repl test for `buildReadlineNoMarkdown`
2024-07-27 00:55:57 +02:00
e062021314 fix NIX_PATH for real (#11079)
* fix NIX_PATH overriding

- test restricted evaluation
- test precedence for setting the search path

Co-authored-by: Robert Hensing <robert@roberthensing.nl>
Co-authored-by: John Ericson <git@JohnEricson.me>
2024-07-24 21:17:15 +00:00
Robert Hensing 3172e88af5 Make abort() call sites log first 2024-07-24 16:52:04 +02:00
Eelco Dolstra d08bb025e1 Merge pull request #11043 from hercules-ci/assert-eq
`assert`: Report why values aren't equal
2024-07-22 17:34:28 +02:00
Robert Hensing b16861d82e libexpr: Track and show GC time and cycle number 2024-07-22 12:56:01 +02:00
Eelco Dolstra 9c6678da0e Merge pull request #11092 from DeterminateSystems/hash-SourcePath
Use std::unordered_map for the EvalState caches
2024-07-16 11:06:43 +02:00
Robert Hensing 03d33703ef Revert "Use HintFmt for doc comments"
Unfortunately these don't render correctly, because they go into the
markdown renderer, instead of the terminal.

```
nix-repl> :doc lib.version
Attribute '[35;1mversion[0m'

    … defined at [35;1m/home/user/h/nixpkgs/lib/default.nix:73:40[0m
```

We could switch that to go direct to the terminal, but then we should
do the same for the primops, to get a consistent look.

Reverting for now.

This reverts commit 3413e0338cbee1c7734d5cb614b5325e51815cde.
2024-07-15 19:56:40 +02:00
Robert Hensing ce31a0457f Use HintFmt for doc comments 2024-07-15 19:56:40 +02:00
Robert Hensing cef11b23e8 Add missing .sh in _NIX_TEST_ACCEPT=1 message 2024-07-15 19:56:40 +02:00
Robert Hensing d4f576b0b2 nix repl: Render docs for attributes 2024-07-15 19:56:40 +02:00
Robert Hensing 491b9cf415 Refactor: extract DocComment::getInnerText(PosTable) 2024-07-15 19:56:40 +02:00
Robert Hensing 7fae378835 Track doc comments and render them in :doc 2024-07-15 19:56:40 +02:00
Robert Hensing 7e604f716c concatStrings: Give compiler access to definition for inlining
... at call sites that are may be in the hot path.

I do not know how clever the compiler gets at these sites.
My primary concern is to not regress performance and I am confident
that this achieves it the easy way.
2024-07-14 12:20:45 +02:00
John Ericson bc83b9dc1f Remove comparator.hh and switch to <=> in a bunch of places
Known behavior changes:

- `MemorySourceAccessor`'s comparison operators no longer forget to
  compare the `SourceAccessor` base class.

Progress on #10832

What remains for that issue is hopefully much easier!
2024-07-12 14:54:18 -04:00
Eelco Dolstra cdc23b67a6 Provide std::hash<SourcePath> 2024-07-12 17:54:27 +02:00
John Ericson 337a5a23b7 Merge pull request #11089 from NixOS/warnings-includes
Fix warnings and optimize includes
2024-07-12 10:29:26 -04:00
Robert Hensing 8df041cbc6 Solve unused header warnings reported by clangd 2024-07-12 15:37:54 +02:00
John Ericson 3fc77f281e No global settings in libnixfetchers and libnixflake
Progress on #5638

There are still a global fetcher and eval settings, but they are pushed
down into `libnixcmd`, which is a lot less bad a place for this sort of
thing.

Continuing process pioneered in
52bfccf8d8.
2024-07-12 08:50:28 -04:00