Warning: This is probably the biggest refactor in tvix-eval history,
so far.
This replaces all instances of trampolines and recursion during
evaluation of the VM loop with generators. A generator is an
asynchronous function that can be suspended to yield a message (in our
case, vm::generators::GeneratorRequest) and receive a
response (vm::generators::GeneratorResponsee).
The `genawaiter` crate provides an interpreter for generators that can
drive their execution and lets us move control flow between the VM and
suspended generators.
To do this, massive changes have occured basically everywhere in the
code. On a high-level:
1. The VM is now organised around a frame stack. A frame is either a
call frame (execution of Tvix bytecode) or a generator frame (a
running or suspended generator).
The VM has an outer loop that pops a frame off the frame stack, and
then enters an inner loop either driving the execution of the
bytecode or the execution of a generator.
Both types of frames have several branches that can result in the
frame re-enqueuing itself, and enqueuing some other work (in the
form of a different frame) on top of itself. The VM will eventually
resume the frame when everything "above" it has been suspended.
In this way, the VM's new frame stack takes over much of the work
that was previously achieved by recursion.
2. All methods previously taking a VM have been refactored into async
functions that instead emit/receive generator messages for
communication with the VM.
Notably, this includes *all* builtins.
This has had some other effects:
- Some test have been removed or commented out, either because they
tested code that was mostly already dead (nix_eq) or because they
now require generator scaffolding which we do not have in place for
tests (yet).
- Because generator functions are technically async (though no async
IO is involved), we lose the ability to use much of the Rust
standard library e.g. in builtins. This has led to many algorithms
being unrolled into iterative versions instead of iterator
combinations, and things like sorting had to be implemented from scratch.
- Many call sites that previously saw a `Result<..., ErrorKind>`
bubble up now only see the result value, as the error handling is
encapsulated within the generator loop.
This reduces number of places inside of builtin implementations
where error context can be attached to calls that can fail.
Currently what we gain in this tradeoff is significantly more
detailed span information (which we still need to bubble up, this
commit does not change the error display).
We'll need to do some analysis later of how useful the errors turn
out to be and potentially introduce some methods for attaching
context to a generator frame again.
This change is very difficult to do in stages, as it is very much an
"all or nothing" change that affects huge parts of the codebase. I've
tried to isolate changes that can be isolated into the parent CLs of
this one, but this change is still quite difficult to wrap one's mind
and I'm available to discuss it and explain things to any reviewer.
Fixes: b/238, b/237, b/251 and potentially others.
Change-Id: I39244163ff5bbecd169fe7b274df19262b515699
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8104
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Reviewed-by: Adam Joseph <adam@westernsemico.com>
Tested-by: BuildkiteCI
This CL removes calling into_iter on a reference, as it will
not move out it's content into resulting iterator.
Change-Id: Ifcc10b7cf33b98453570cbcec3eb82ffaba2ffcb
Signed-off-by: Aaqa Ishtyaq <aaqaishtyaq@gmail.com>
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8126
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
We need to distinguish explicitly between the paths used for the
scanner, and the paths that populate the derivation inputs. The full
paths must be accessible from the result of the refscanner to populate
drv fields correctly.
This was previously hidden by debug changes that masked actual IO
operations with no-ops.
Change-Id: I037af6e6bbe2b573034d695f8779bee1b56bc125
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8022
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Creates a cache of imported literal files (e.g.
`./default-builder.sh`) which avoids shelling out to Nix for each
instance of the same file.
Note that a better way to tackle this is to create memoizable thunks
for these expressions in the compiler, but we are lacking a little bit
of infrastructure for that at the moment.
Change-Id: Ibc062b20d81e97dd3986e734d225a744e1779fe7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8015
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
Switch out the string-scanning algorithm used in the reference scanner.
The construction of aho-corasick automata made up the vast majority of
runtime when evaluating nixpkgs previously. While the actual scanning
with a constructed automaton is relatively fast, we almost never scan
for the same set of strings twice and the cost is not worth it.
An algorithm that better matches our needs is the Wu-Manber multiple
string match algorithm, which works efficiently on *long* and *random*
strings of the *same length*, which describes store paths (up to their
hash component).
This switches the refscanner crate to a Rust implementation[0][1] of
this algorithm.
This has several implications:
1. This crate does not provide a way to scan streams. I'm not sure if
this is an inherent problem with the algorithm (probably not, but
it would need buffering). Either way, related functions and
tests (which were actually unused) have been removed.
2. All strings need to be of the same length. For this reason, we
truncate the known paths after their hash part (they are still
unique, of course).
3. Passing an empty set of matches, or a match that is shorter than
the length of a store path, causes the crate to panic. We safeguard
against this by completely skipping the refscanning if there are no
known paths (i.e. when evaluating the first derivation of an eval),
and by bailing out of scanning a string that is shorter than a
store path.
On the upside, this reduces overall runtime to less 1/5 of what it was
before when evaluating `pkgs.stdenv.drvPath`.
[0]: Frankly, it's a random, research-grade MIT-licensed
crate that I found on Github:
https://github.com/jneem/wu-manber
[1]: We probably want to rewrite or at least fork the above crate, and
add things like a three-byte wide scanner. Evaluating large
portions of nixpkgs can easily lead to more than 65k derivations
being scanned for.
Change-Id: I08926778e1e5d5a87fc9ac26e0437aed8bbd9eb0
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8017
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This doesn't require any other corresponding handling *yet*, as the
actual replacements happen in the builder logic (which we delegate to
cppnix at the moment).
Change-Id: I034147c933f05ae427c7a8794647132d108d0ede
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7972
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
Put this in its src/derivation.
Change-Id: Ic047ab1c2da555a833ee454e10ef60c77537b617
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7967
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Autosubmit: flokli <flokli@flokli.de>
We only do logic here if algo and hash_mode are Some(_)
(and there's an `out` output).
The fact we don't do anything in all in other cases is a bit hidden at
the bottom. Use if let for the destructuring, and drop the other case.
Change-Id: Icc0e38e62947d52d48ef610f754749737977fca9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7966
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
This helper function only was created because
populate_output_configuration was hard to test before cl/7939.
With that out of the way, we can pull it in.
Change-Id: I64b36c0eb34343290a8d84a03b0d29392a821fc7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7961
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
This is repetitive and error prone (e.g. switching around
to_string/as_str has drastic consequences) due to the ToString
overloads.
Change-Id: I9b16a2e0e05e4c21e83f43e9f603746eb42e53f7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7947
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Autosubmit: tazjin <tazjin@tvl.su>
This is much less code, and makes it much easier to read.
Change-Id: I9028f226105f905c2cc2cabd33907ff493e26225
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7938
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Autosubmit: flokli <flokli@flokli.de>
Instead of being called with `md5`, `sha1`, `sha256` or `sha512`,
`fetchurl.nix` (from corepkgs / `<nix`) can also be called with a `hash`
attribute, being an SRI hash.
In that case, `builtin.derivation` is called with `outputHashAlgo` being
an empty string, and `outputHash` being an SRI hash string.
In other cases, an SRI hash is passed as outputHash, but outputHashAlgo
is set too.
Nix does modify these values in (single, fixed) output specification it
serializes to ATerm, but keeps it unharmed in `env`.
Move this into a construct_output_hash helper function, that can be
tested better in isolation.
Change-Id: Id9d716a119664c44ea7747540399966752e20187
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7933
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
This adds an implementation of this builtin which correctly calculates
paths, but does not actually write anything to the store or verify
references.
Change-Id: Ie9764cbc1d13a73d8dc9350910304e2b7cad3fe8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7910
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This uses the actual upstream Nix code for
`builtins.derivation` (which is not a primop in C++ Nix) to implement
`builtins.derivation` as a wrapper around `builtins.derivationStrict`.
We're doing it this way to ensure that our thunking logic is correct.
An initial Rust-native rewrite (see e.g. cl/7363) is pretty difficult
to debug while there are still other issues to root out, but
eventually we might want to turn this into native code.
Change-Id: I5845e18073e103b8670e40648bd7fd9b511058e0
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7902
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
Implements the logic for converting an evaluator value supplied as
arguments to builtins.derivationStrict into an actual,
fully-functional derivation struct.
This skips the implementation of structuredAttrs, which are left for a
subsequent commit.
Note: We will need to port some eval tests over to CLI to test this
correct, which will be done in a separate commit later on.
Change-Id: I0db69dcf12716180de0eb0b126e3da4683712966
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7756
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Adds a helper function which handles special parameters to
`builtins.derivation` that are not just blindly passed through to the
builder environment, but populate other specific fields of the
derivation (outside of the ones handled by other, more complex helpers
from previous commits).
Change-Id: I82d1edf9af714fc4591e9071c0b83ece83be7eee
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7901
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
This threads through the fields that control whether a derivation is a
fixed-output derivation or not.
Change-Id: I49739de178fed9f258291174ca1a2c15a7cf5c2a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7900
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This adds a helper function which takes the output of the reference
scanner used on derivation inputs and populates the `input_sources`
and `input_derivations` field of the derivation accordingly.
Note that we have a divergence from C++ Nix here, as we do not
populate the entire FS closure of a literally referred derivation (and
our standing theory is that this is unnecessary for nixpkgs).
Change-Id: Id0f605dd8c0a82973c56605c2b8f478fc17777d6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7899
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
Adds a small helper function which uses a Nix value supplied to
`builtins.derivation{Strict}` to populate the `outputs` field of the
`Derivation` struct.
Change-Id: Iccc7a4f293b3d913140aed576a573a8992241e46
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7898
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
We only stripped one of the two uses of this string, leading to
extraneous newlines in the refscanner.
Change-Id: I25d9119be082c487352f0cf66b97ecdcc3e1de06
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7932
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
These will be threaded through to eval through the new `TvixError`
variant.
Change-Id: Ia0d3f8710dcf26bb95015cd2a6a2b2911f06343f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7842
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
No situation should be allowed in which a path is inserted into
known_paths with different types twice, which we previously enforced
only for some path types.
Change-Id: I8cb47d4b29c0aab3c58694f8b590e131deba7043
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7843
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Replacement strings are some weird internal feature of Nix that is
required for calculating derivation hashes. We need to track these
like other paths, as they need to be re-used on builds with
dependencies on values from previous builds.
Change-Id: Ie955b3fb5ae3685cfadfbe4d06ea6b5e219590c7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7828
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
When adding things to a C++ Nix store, ensure that the path is tracked
in the tracker.
Since the mechanism for propagating the tracker instance isn't
finalised yet, I've opted to take an Rc<RefCell> parameter for it. How
exactly that ends up there is going to become clear in the next
commits, but for now it's just instantiated in main with
Default::default.
Change-Id: I90f0b44f2d4f292dedc98ff1aa39041d279b61fd
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7833
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This gets very complex very quickly otherwise, as all the construction
paths for a reference scanner and all the access patterns for the
KnownPaths structure are not yet fully understood.
Change-Id: Ibadf1f18b476695f3c286fc6896ae557760edf63
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7827
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
This module implements types used to track the set of known paths in
the context of an evaluation.
These are used to determine the build references of a derivation.
Change-Id: I81e15ae33632784e699128916485751613b231a3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7816
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
Using yet more machinery from the pretty comprehensive aho_corasick
crate, this makes it possible to pass anything implementing `io::Read`
to the `ReferenceScanner` to accumulate matches.
Change-Id: I5b0e28eb44ea4df24010f40831e29f2cbb8c1f80
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7810
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
This module implements a ReferenceScanner struct which uses the
aho_corasick crate to scan string inputs for known, non-overlapping
candidates (store paths, in our case).
I experimented with several different APIs, and landed on this version
with an initial accumulator in the scanner. The scanner is
instantiated from the candidates and "fed" all the strings, then
consumed by the caller to retrieve the result.
Right now only things that look vaguely like bytestrings can be fed to
the scanner, there is no streaming support in the API yet.
Change-Id: I7782f0f0df5fc64bccd813aa14712f5525b0168c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7808
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
With this, tvix/cli can be run on files and only produce compiler
errors and warnings.
Change-Id: I5dd9229fc065647787daafd17d7c1540579a1d98
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7764
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Previously the construction of globals (a compiler-only concept) and
builtins (a (now) user-facing API) was intermingled between multiple
different modules, and kind of difficult to understand.
The complexity of this had grown in large part due to the
implementation of `builtins.import`, which required the notorious
"knot-tying" trick using Rc::new_cyclic (see cl/7097) for constructing
the set of globals.
As part of the new `Evaluation` API users should have the ability to
bring their own builtins, and control explicitly whether or not impure
builtins are available (regardless of whether they're compiled in or
not).
To streamline the construction and allow the new API features to work,
this commit restructures things by making these changes:
1. The `tvix_eval::builtins` module is now only responsible for
exporting sets of builtins. It no longer has any knowledge of
whether or not certain sets (e.g. only pure, or pure+impure) are
enabled, and it has no control over which builtins are globally
available (this is now handled in the compiler).
2. The compiler module is now responsible for both constructing the
final attribute set of builtins from the set of builtins supplied
by a user, as well as for populating its globals (that is
identifiers which are available at the top-level scope).
3. The `Evaluation` API now carries a `builtins` field which is
populated with the pure builtins by default, and can be extended by
users.
4. The `import` feature has been moved into the compiler, as a
special case. In general, builtins no longer have the ability to
reference the "fix point" of the globals set.
This should not change any functionality, and in fact preserves minor
differences between Tvix/Nix that we already had (such as
`builtins.builtins` not existing).
Change-Id: Icdf5dd50eb81eb9260d89269d6e08b1e67811a2c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7738
Reviewed-by: sterni <sternenseemann@systemli.org>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This type allows for temporarily compatibility with the C++ Nix store,
specifically (for now) it gives us the store directory used by Nix and
imports files the same way.
Change-Id: I4767794ef2863eba49661315c63c4e17de946d60
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7587
Reviewed-by: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
... until we have a store-I/O layer, or something that intercepts the
store-related stuff appropriately.
Change-Id: I22f63435b3f9e118e3faeb2924fda8373a23ea7f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7568
Reviewed-by: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
Users can again pass flags for dumping the AST, bytecode, and runtime
trace.
With this commit the CLI is at feature-parity with what we had before,
but entirely through the new API.
Change-Id: I30fe26f243224b25d1e4f828fec607325ef88306
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7550
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
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
