mdebray/tvl-depot
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
tvl-depot/tvix/eval/Cargo.toml

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

59 lines
1.7 KiB
TOML
Raw Normal View History

feat(tvix/eval): check in generated project skeleton Change-Id: Iecc8283abb289de71f22076fd88892f6ded99cb3 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6062 Tested-by: BuildkiteCI Reviewed-by: grfn <grfn@gws.fyi>
2022-08-04 15:29:38 +02:00
[package]
name = "tvix-eval"
version = "0.1.0"
edition = "2021"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
chore(tvix/eval): gate REPL-only dependencies behind `repl` feature With this change, it becomes possible to compile tvix-eval to webassembly if the `repl` feature is disabled. Change-Id: Icc0a059964cd0bea2054110c682d50fc5c87ec01 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6446 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2022-09-03 18:40:01 +02:00
[lib]
name = "tvix_eval"
feat(tvix/eval): check in generated project skeleton Change-Id: Iecc8283abb289de71f22076fd88892f6ded99cb3 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6062 Tested-by: BuildkiteCI Reviewed-by: grfn <grfn@gws.fyi>
2022-08-04 15:29:38 +02:00
[dependencies]
refactor(tvix/value): replace static representation with SmolStr The only uses of the static variant were for `"name"` and `"value"`, which are both small enough to fit into a SmolStr. The size of NixString accomodates `String` anyways, so we may as well inline them. Additionally smol_str is already in the dependency graph because rnix uses it, and using it for representations of identifiers is sensible. Change-Id: I9969312256d1657d69128e54c47dc7294a18ce58 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6165 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org> Reviewed-by: grfn <grfn@gws.fyi>
2022-08-11 18:02:19 +02:00
smol_str = "0.1"
chore(tvix/cargo): add dirs dependency This crate provides sensible default locations for directories on different platforms. Change-Id: I0b61cc7f626dc6c8df903ba0f873be24e07d69b5 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6170 Reviewed-by: grfn <grfn@gws.fyi> Tested-by: BuildkiteCI
2022-08-11 23:26:05 +02:00
dirs = "4.0.0"
fix(tvix/eval): use path_clean instead of fs::canonicalize for paths Canonicalisation performs much more functionality than what C++ Nix implements for paths, and causes some undesirable behaviour (e.g. handling non-existant files becomes difficult, but should be possible in literals). Instead, the path_clean crate provides a pure normalisation method. There is an intention to add this to Rust itself: https://github.com/rust-lang/rfcs/issues/2208 Change-Id: I775d238136db0a52cf6b12a68985833c8fb32882 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6186 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2022-08-12 17:52:48 +02:00
path-clean = "0.1"
feat(tvix/eval): implement DisassemblingObserver for compiler This type implements an observer that is called whenever the compiler emits a chunk (after the toplevel, thunks, or lambdas) and prints the output of the disassembler to its internal writer. This replaces half of the uses of the `disassembler` feature, which has been removed from the Cargo configuration. Note that at this commit runtime tracing is not yet implemented as an observer. Change-Id: I7894ca1ba445761aba4ad51d98e4a7b6445f1aea Reviewed-on: https://cl.tvl.fyi/c/depot/+/6449 Reviewed-by: sterni <sternenseemann@systemli.org> Tested-by: BuildkiteCI
2022-09-04 15:56:20 +02:00
tabwriter = "1.2"
refactor(tvix/eval): Upgrade to latest rnix-parser Since the latest published version of rnix-parser on crates.io, the crate has undergone major changes which are only available in the git repository at the moment. This commit updates the compiler to this newer version of rnix. Most notably, the entire AST provided by rnix is now wrapped in the AST type system. As a result of this traversal is much nicer in many places, especially for things like nested attribute selection. There are a handful of smaller features missing for full feature parity with the previous version, especially handling of path literals, but PRs for these already exist in rnix-parser. Change-Id: Icde6d393067976549492b7d89c4cc49e5e575fc7 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6231 Reviewed-by: sterni <sternenseemann@systemli.org> Tested-by: BuildkiteCI
2022-08-16 22:43:45 +02:00
rowan = "*" # pinned by rnix
chore(tvix/eval): add codemap dependency This will be used to track source spans when emitting bytecode. The codemap is a data structure which tracks *all* the source files visited by an evaluation, and makes it possible to represent locations across all of the files using a simple span (i.e. pair of offsets). When reporting errors, this even contains enough information to reconstruct the rnix AST to create fancier reporting in certain cases if desired. Change-Id: I4ae98620b9b150fb5a389bd7f1e12670e3192c62 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6374 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2022-09-01 14:52:48 +02:00
codemap = "0.1.3"
chore(tvix/eval): add dependency on codemap-diagnostic This is a crate for source-span based error reporting. Since all of our spans are already codemap spans, it is a good starting point. We have to figure out quite a bit of logic for neat error printing; later on if we want fancier presentation we might want to look at one of the other libraries in this space like miette. Change-Id: I4e28886af1ed199b7112d9dbf063c9f29b612bf1 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6531 Autosubmit: tazjin <tazjin@tvl.su> Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org> Reviewed-by: grfn <grfn@gws.fyi>
2022-09-11 23:34:25 +02:00
codemap-diagnostic = "0.1.1"
chore(tvix/eval): Add deps on proptest+test_strategy It's proptest time! Add deps on the proptest and test_strategy crates, gated behind a new (default-enabled) "arbitrary" feature flag so that they don't affect dependencies of things like tvixbolt that depend on tvix. These are going in dependencies, not dev-dependencies, so that we can impl Arbitrary for stuff outside of test modules (which will be important for integration suites which want to run proptests) Change-Id: I1613bd3ea9a835e22986ad4e59700e8736007963 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6624 Autosubmit: grfn <grfn@gws.fyi> Tested-by: BuildkiteCI Reviewed-by: tazjin <tazjin@tvl.su>
2022-09-17 19:52:02 +02:00
proptest = { version = "1.0.0", default_features = false, features = ["std", "alloc", "break-dead-code", "tempfile"], optional = true }
test-strategy = { version = "0.2.1", optional = true }
feat(tvix/eval): add an AST pretty-printing module This implements serde::Serialize for the rnix AST through a wrapper type, and exposes a function for serialising the AST into a (pretty-printed JSON) string representation. This can be used to debug issues with the AST, and to display an AST reprsentation in tools like tvixbolt. Serialize is implemented manually because we don't own any of the structs and the way to traverse them is not easily derived automatically, and this is quite verbose. We might be able to condense it a little bit, but at the same time it's also fairly straightforward. Change-Id: I922df43cfc25636f3c8baee7944c75ade516055c Reviewed-on: https://cl.tvl.fyi/c/depot/+/6943 Autosubmit: tazjin <tazjin@tvl.su> Tested-by: BuildkiteCI Reviewed-by: Adam Joseph <adam@westernsemico.com> Reviewed-by: tazjin <tazjin@tvl.su>
2022-10-13 14:23:45 +02:00
serde = "1.0"
serde_json = "1.0"
feat(tvix/eval): Initial impl of builtins.match Implement an *initial* version of builtins.match, using the rust `regex` crate for regular expressions. The rust regex crate definitely has different semantics than nix's regular expressions - but we'd like to see how far we can get before the incompatibility starts to matter. This consciously leaves out any sort of memo for compiled regular expressions (which upstream nix also has) for the sake of expediency - in the future we should implement that so we don't have to compile the same regular expression multiple times. Change-Id: I5b718635831ec83397940e417a9047c4342b6fa1 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6989 Tested-by: BuildkiteCI Reviewed-by: Adam Joseph <adam@westernsemico.com> Reviewed-by: tazjin <tazjin@tvl.su>
2022-10-13 00:45:52 +02:00
regex = "1.6.0"
feat(tvix/eval): Add a proc-macro for defining builtins Add a single new proc macro to a new proc-macro crate, `tvix-eval-proc-macros` for defining an inline module containing nix builtins, and automatically generating a function within that module which returns a list of those builtins as `tvix_eval::value::Builtin`. Change-Id: Ie4afae438914d2af93d15637151a49b4c68aa352 Reviewed-on: https://cl.tvl.fyi/c/depot/+/7198 Reviewed-by: tazjin <tazjin@tvl.su> Reviewed-by: Adam Joseph <adam@westernsemico.com> Tested-by: BuildkiteCI
2022-11-06 00:50:31 +01:00
builtin-macros = { path = "./builtin-macros", package = "tvix-eval-builtin-macros" }
feat(tvix/eval): use backtrace-on-stack-overflow crate The backtrace-on-stack-overflow create provides best-effort stack traces when a stack overflow happens. Since it's running on the (usually tiny) signal alternate stack this isn't easy. This is guarded by a new `backtrace_overflow` feature flag and never enabled (even if that feature is selected) for release builds. This is strictly for debugging; there's crazy unsafe voodoo in there. https://lib.rs/crates/backtrace-on-stack-overflow Example output: ``` Stack Overflow: 0: backtrace_on_stack_overflow::handle_sigsegv at /home/amjoseph/.cargo/registry/src/github.com-1ecc6299db9ec823/backtrace-on-stack-overflow-0.2.0/src/lib.rs:93:40 1: <unknown> 2: __rust_probestack 3: tvix_eval::vm::VM::run_op at src/vm.rs:399 4: tvix_eval::vm::VM::run at src/vm.rs:388:23 5: tvix_eval::vm::VM::enter_frame at src/vm.rs:360:22 6: tvix_eval::value::thunk::Thunk::force at src/value/thunk.rs:116:25 7: tvix_eval::vm::VM::run_op at src/vm.rs:801:37 8: tvix_eval::vm::VM::run at src/vm.rs:388:23 9: tvix_eval::vm::VM::enter_frame at src/vm.rs:360:22 10: tvix_eval::value::thunk::Thunk::force at src/value/thunk.rs:116:25 ... ``` Signed-off-by: Adam Joseph <adam@westernsemico.com> Change-Id: I1d8a2017f836be7bf91a2223e7adacb86fa1dbb2 Reviewed-on: https://cl.tvl.fyi/c/depot/+/7354 Tested-by: BuildkiteCI Reviewed-by: tazjin <tazjin@tvl.su>
2022-11-23 04:42:06 +01:00
backtrace-on-stack-overflow = { version = "0.2.0", optional = true }
fix(tvix/eval): use rnix-parser from crates.io A few weeks ago, oberblastmeister did a release to crates.io so we can stop importing it via GitHub. Change-Id: I9d5fa5cd281685779c71b12fed45ed201a1db17e Reviewed-on: https://cl.tvl.fyi/c/depot/+/7532 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2022-12-06 14:31:09 +01:00
rnix = "0.11.0"
chore(tvix/eval): add 'test-generator' as dev-dependency This crate makes it easy to generate test cases from files on disk, which is used for running the Nix test suite against Tvix. Change-Id: I60ef26484d05e723982679dd42958adf52735916 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6125 Tested-by: BuildkiteCI Reviewed-by: grfn <grfn@gws.fyi>
2022-08-10 17:52:42 +02:00
[dev-dependencies]
chore(tvix): upgrade to clap 4.0 In //tvix/eval: * criterion bumped to 4.0, which at least depends on clap 3.x instead of 2.x, which is less incompatible In //tvix/cli: * no changes required In //tvix/nix_cli: * some minor changes for compatibility with clap 4.0, no functionality changes Change-Id: If793f64b59fcaa2402d3d483ddbab4092f32df03 Reviewed-on: https://cl.tvl.fyi/c/depot/+/7588 Reviewed-by: grfn <grfn@gws.fyi> Tested-by: BuildkiteCI
2022-12-16 12:54:22 +01:00
criterion = "0.4"
refactor(tvix/eval): use pretty_assertions for tests This makes for much more readable output especially when long strings are involved. Change-Id: I43dd73a0480535d7181a760788c42883a9b083f8 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6229 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2022-08-16 14:33:50 +02:00
pretty_assertions = "1.2.1"
test(tvix/eval): Add attr merge benchmarks Add a quick couple of benchmarks for merging attribute sets, large and small. Change-Id: I26940a9cf4e0d30e3d9eb07a7b8c366ca4072ca3 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6286 Tested-by: BuildkiteCI Reviewed-by: tazjin <tazjin@tvl.su> Autosubmit: grfn <grfn@gws.fyi>
2022-09-02 21:02:19 +02:00
itertools = "0.10.3"
test(tvix/eval): Add the start of a nix oracle test suite Add the start of a test suite that compares tvix eval results against nix, using the string repr of the value as the comparison. This shells out to a nix-instantiate binary, which is configurable as an environment variable, to eval - there's some extra machinery there to setup a new nix store as a tempdir to allow running this test inside the nix build for tvix-eval itself. Currently this has a macro that'll allow writing lots and lots of hardcoded tests, but going forward I'm also going to be looking into adding proptest-based generation of expressions to compare. Change-Id: I9f4895fab1e668ed2b7dfd6f92f8c80de1bbb16b Reviewed-on: https://cl.tvl.fyi/c/depot/+/6307 Autosubmit: grfn <grfn@gws.fyi> Reviewed-by: tazjin <tazjin@tvl.su> Tested-by: BuildkiteCI
2022-09-03 17:42:27 +02:00
tempdir = "0.3.7"
chore(tvix/tests): gate Nix test suite behind `nix_tests` feature Once we have full coverage they should be enabled by default. Change-Id: Iace9e1ae9a9f901a0979ad336434004b8028fe8a Reviewed-on: https://cl.tvl.fyi/c/depot/+/6129 Tested-by: BuildkiteCI Reviewed-by: grfn <grfn@gws.fyi>
2022-08-10 18:18:01 +02:00
fix(tvix/eval): use test-generator fork that supports workspaces This should make no difference in Nix builds, but allows running tests locally again with `cargo test` for //tvix/eval. Change-Id: I97d61840143d5c14db61d5862781bf635f9a28e7 Reviewed-on: https://cl.tvl.fyi/c/depot/+/7590 Reviewed-by: grfn <grfn@gws.fyi> Tested-by: BuildkiteCI Reviewed-by: flokli <flokli@flokli.de>
2022-12-18 15:23:35 +01:00
[dev-dependencies.test-generator]
# This fork of test-generator adds support for cargo workspaces, see
# also https://github.com/frehberg/test-generator/pull/14
git = "https://github.com/JamesGuthrie/test-generator.git"
rev = "82e799979980962aec1aa324ec6e0e4cad781f41"
chore(tvix/tests): gate Nix test suite behind `nix_tests` feature Once we have full coverage they should be enabled by default. Change-Id: Iace9e1ae9a9f901a0979ad336434004b8028fe8a Reviewed-on: https://cl.tvl.fyi/c/depot/+/6129 Tested-by: BuildkiteCI Reviewed-by: grfn <grfn@gws.fyi>
2022-08-10 18:18:01 +02:00
[features]
refactor(tvix): split binary (REPL etc.) out from evaluator library The tvix-eval project is independent from any *uses* of the evaluator, such as the tvix-repl. This functionality has been split out into a separate "tvix-cli" crate. Note that this doesn't have to mean that this CLI crate is the "final" CLI crate for tvix, the point of this is not "getting the CLI structure right" but rather "getting the evaluator structure right". This reshuffling is part of restructuring the way that functionality like store communication is injected into language evaluation. Note that at this commit the new CLI crate is not at feature-parity. Change-Id: Id0af03dc8e07ef09a9f882a89612ad555eca8f93 Reviewed-on: https://cl.tvl.fyi/c/depot/+/7541 Autosubmit: tazjin <tazjin@tvl.su> Reviewed-by: grfn <grfn@gws.fyi> Tested-by: BuildkiteCI
2022-12-08 22:19:22 +01:00
default = [ "impure", "arbitrary", "nix_tests", "backtrace_overflow" ]
chore(tvix/eval): gate REPL-only dependencies behind `repl` feature With this change, it becomes possible to compile tvix-eval to webassembly if the `repl` feature is disabled. Change-Id: Icc0a059964cd0bea2054110c682d50fc5c87ec01 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6446 Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org>
2022-09-03 18:40:01 +02:00
chore(tvix/tests): gate Nix test suite behind `nix_tests` feature Once we have full coverage they should be enabled by default. Change-Id: Iace9e1ae9a9f901a0979ad336434004b8028fe8a Reviewed-on: https://cl.tvl.fyi/c/depot/+/6129 Tested-by: BuildkiteCI Reviewed-by: grfn <grfn@gws.fyi>
2022-08-10 18:18:01 +02:00
# Enables running the Nix language test suite from the original C++
# Nix implementation (at version 2.3) against Tvix.
nix_tests = []
feat(tvix/eval): Add initial benchmarking infrastructure Bootstrap some (initially very simple, mostly proof-of-concept) benchmarking infrastructure using Criterion, using the newly-exposed lib from tvix-eval. Change-Id: I4bb93c142ba8d018d7e67e58ac8907a0429398a5 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6156 Reviewed-by: tazjin <tazjin@tvl.su> Tested-by: BuildkiteCI Autosubmit: grfn <grfn@gws.fyi>
2022-08-25 17:04:07 +02:00
feat(tvix/eval): Sketch out impure builtins Sketch out a new set of "impure" builtins, which supplement the existing set of "pure" builtins but are gated behind a feature flag, which allows them to be omitted by crates depending on tvix-eval that only want pure evaluation, such as tvixbolt. Change-Id: I2736017b5c9b4776bbba8758e108ec84887abd66 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6655 Reviewed-by: wpcarro <wpcarro@gmail.com> Tested-by: BuildkiteCI Reviewed-by: sterni <sternenseemann@systemli.org> Reviewed-by: tazjin <tazjin@tvl.su>
2022-09-18 22:34:41 +02:00
# Enables operations in the VM which depend on the ability to perform I/O
impure = []
chore(tvix/eval): Add deps on proptest+test_strategy It's proptest time! Add deps on the proptest and test_strategy crates, gated behind a new (default-enabled) "arbitrary" feature flag so that they don't affect dependencies of things like tvixbolt that depend on tvix. These are going in dependencies, not dev-dependencies, so that we can impl Arbitrary for stuff outside of test modules (which will be important for integration suites which want to run proptests) Change-Id: I1613bd3ea9a835e22986ad4e59700e8736007963 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6624 Autosubmit: grfn <grfn@gws.fyi> Tested-by: BuildkiteCI Reviewed-by: tazjin <tazjin@tvl.su>
2022-09-17 19:52:02 +02:00
# Enables Arbitrary impls for internal types (required to run tests)
arbitrary = [ "proptest", "test-strategy" ]
feat(tvix/eval): use backtrace-on-stack-overflow crate The backtrace-on-stack-overflow create provides best-effort stack traces when a stack overflow happens. Since it's running on the (usually tiny) signal alternate stack this isn't easy. This is guarded by a new `backtrace_overflow` feature flag and never enabled (even if that feature is selected) for release builds. This is strictly for debugging; there's crazy unsafe voodoo in there. https://lib.rs/crates/backtrace-on-stack-overflow Example output: ``` Stack Overflow: 0: backtrace_on_stack_overflow::handle_sigsegv at /home/amjoseph/.cargo/registry/src/github.com-1ecc6299db9ec823/backtrace-on-stack-overflow-0.2.0/src/lib.rs:93:40 1: <unknown> 2: __rust_probestack 3: tvix_eval::vm::VM::run_op at src/vm.rs:399 4: tvix_eval::vm::VM::run at src/vm.rs:388:23 5: tvix_eval::vm::VM::enter_frame at src/vm.rs:360:22 6: tvix_eval::value::thunk::Thunk::force at src/value/thunk.rs:116:25 7: tvix_eval::vm::VM::run_op at src/vm.rs:801:37 8: tvix_eval::vm::VM::run at src/vm.rs:388:23 9: tvix_eval::vm::VM::enter_frame at src/vm.rs:360:22 10: tvix_eval::value::thunk::Thunk::force at src/value/thunk.rs:116:25 ... ``` Signed-off-by: Adam Joseph <adam@westernsemico.com> Change-Id: I1d8a2017f836be7bf91a2223e7adacb86fa1dbb2 Reviewed-on: https://cl.tvl.fyi/c/depot/+/7354 Tested-by: BuildkiteCI Reviewed-by: tazjin <tazjin@tvl.su>
2022-11-23 04:42:06 +01:00
# For debugging use only; not appropriate for production use.
backtrace_overflow = [ "backtrace-on-stack-overflow" ]
feat(tvix/eval): Add initial benchmarking infrastructure Bootstrap some (initially very simple, mostly proof-of-concept) benchmarking infrastructure using Criterion, using the newly-exposed lib from tvix-eval. Change-Id: I4bb93c142ba8d018d7e67e58ac8907a0429398a5 Reviewed-on: https://cl.tvl.fyi/c/depot/+/6156 Reviewed-by: tazjin <tazjin@tvl.su> Tested-by: BuildkiteCI Autosubmit: grfn <grfn@gws.fyi>
2022-08-25 17:04:07 +02:00
[[bench]]
name = "eval"
harness = false
Reference in a new issue Copy permalink