The logic in this method is *very* similar to `compile_lambda`. It is
intended to be called around any expression that should be
thunked (such as function applications, attribute set values, etc.).
Change-Id: Idfbb2daa9f4b735095378fb9c39a2fd07c8cff91
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6344
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This instruction finalises the initialisation of deferred upvalues in
closures (and soon, thunks).
The compiler does not yet emit this instruction, some more accounting
is needed for that.
Change-Id: Ic4181b26e19779e206f51e17388559400da5f93a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6337
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Uses the threaded through slot offset to determine whether
initialisation of a captured local upvalue must be defered to a later
point where all values of a scope are available.
This adds a new data representation to the opcode for this situation,
but the equivalent runtime handling is not yet implemented. This is in
part because there is more compiler machinery needed to find the
resolution point.
Change-Id: Ifd0c393f76abfe6e2d91483faf0f58947ab1dedc
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6329
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
This puts together the puzzle pieces for threading dynamic
upvalues (that is, upvalues resolved from the `with`-stack) all the
way through.
Reading the test case enclosed in this commit and walking through it
is recommended to understand what problem is being tackled here.
In short, because the compiler can not statically know *which*
with-scope a dynamic argument is resolved from it needs to lay the
groundwork for resolving from *all* possible scopes.
There are multiple different approaches to doing this. The approach
chosen in this commit is that if a dynamic upvalue is detected, the
compiler will emit instructions to close over this dynamic value
in *all* enclosing lambda contexts.
It uses a new instruction for this that will leave around a sentinel
value in case an identifier could not be resolved, and wire the
location of this found value (or sentinel) up through the upvalues to
the next level of nesting.
In this tradeoff, tvix potentially closes over more upvalues than are
needed (but in practice, how often do people create *really* deep
`with`-stacks? and in *this* kind of code situation? maybe we should
even warn for this!) but avoids keeping the entire attribute sets
themselves around.
Looking at the test case, each surrounding closure will close
over *all* dynamic identifiers that are referenced later on visible to
it, but only the last one for each identifier will actually end up
being used.
This also covers our bases for an additional edge-case this creates,
in which an identifier potentially resolves to a dynamic upvalue *and*
to a dynamic value within the function's own scope (again, would
anyone really do this?) by introducing a resolution instruction for
that particular case.
There is likely some potential for cleaning up this code which is
quite ugly in some parts, but as this implementation is now carefully
calibrated to work I decided it is time to commit it and clean it up
in subsequent commits.
Change-Id: Ib701e3e6da39bd2c95938d1384036ff4f9fb3749
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6322
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
These need to be handled specially by the runtime if the compiler
determines that a given local must be resolved via `with`.
Note that this implementation has a bug: It currently allows `with`
inside of nested lambdas to shadow statically known identifiers. This
will be cleaned up in the next commit.
Change-Id: If196b99cbd1a0f2dbb4a40a0e88cdb09a009c6b9
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6299
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Fully implements the instructions for compiling closure
objects (without runtime handling yet).
Closure (and thunk) objects are created at runtime by capturing all
known upvalues. To represent this, the instructions for creating them
need to have a variable number of arguments. Due to that, this commit
introduces new variants in OpCode that are not actually operations,
but data.
If the VM is implemented correctly, the instruction pointer should
never point at these. Due to this, the VM will panic if it sees a data
operand during an execution run.
Change-Id: Ic56b49b3a42736dc437751e76df0e89c8d0619c6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6291
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
This adds a new upvalue tracking structure in the compiler to resolve
upvalues and track their positions within a function when compiling a
closure.
The compiler will emit runtime upvalue access instructions after this
commit, but the creation of the runtime closure object etc. is not yet
wired up.
Change-Id: Ib0c2c25f686bfd45f797c528753068858e3a770d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6289
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
This adds a transparent wrapper around `usize` used for jump offsets
in the opcodes. This is a step towards getting rid of ambiguous plain
`usize` usage in the opcode.
Change-Id: I21e35e67d94b32d68251908b96c7f62b6f56a8bb
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6282
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Nix functions always have a single argument and we do not yet make
efforts to optimise this in Tvix for known multi-argument functions
being directly applied.
For this reason, the call instruction is fairly simple and just calls
out to construct a new call frame.
Note that the logic for terminating the run loop has moved to the top
of the dispatch; this is because the loop run needs to be skipped if
the call frame for the current lambda has just been dropped.
Change-Id: I259bc07e19c1e55cd0a65207fa8105b23052b967
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6249
Reviewed-by: sterni <sternenseemann@systemli.org>
Reviewed-by: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
This implements `assert`, which evaluates an expression and aborts
evaluation if the value is not `true`.
At this point we should introduce eval-failed-* tests; probably
asserting against some representation of the error enum?
Change-Id: If54c8f616d89b829c1860a4835dde60a2cd70d7a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6230
Reviewed-by: grfn <grfn@gws.fyi>
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Adds an additional structure to the compiler's scope to track the
runtime "with stack", i.e. the stack of values through which
identifiers should be dynamically resolved within a with-scope.
When encountering a `with` expression, the value from which the
bindings should be resolved is pushed onto the stack and tracked by
the compiler in the "with stack", as well as with a "phantom value"
which indicates that the stack contains an additional slot which is
not available to users via identifiers.
Runtime handling of this is not yet implemented.
Change-Id: I5e96fb55b6378e8e2a59c20c8518caa6df83da1c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6217
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This makes basic `let ... in ...` statements work correctly. It does
not yet account for the call frames pushed into the VM during function
application.
Change-Id: I67155171daf1a43011b96716dd9d1ab04b27db33
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6190
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
These expressions now leave the binding values on the stack, and clean
up the scope after the body of the expression.
While variable access is not yet implemented (as the identifier node
remains unhandled), this already gives us the correct stack behaviour.
Change-Id: I138c20ace9c64502c94b2c0f99a6077cd912c00d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6188
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
This makes it possible to quickly detect code errors that might blow
up the size of the OpCode type.
Change-Id: I7662dd0aa30c4762c0f9e4fa346418c9ca8b9994
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6177
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
This sentinel value is going to be used for certain nested accesses
into attribute sets.
There is a new instruction similar to `OpAttrsSelect` which leaves the
sentinel on the stack if a key is not found, instead of raising an
error.
Additionally, a new jump instruction makes its jump operation
conditional on finding such a sentinel value.
Change-Id: I2642f0a0bcc85bbe0ead68ea09a7dd794dbedeac
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6166
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This makes it possible to check things like `{} ? a` with a single
level of nesting.
Change-Id: I567c36fcfd2f9e2f60071acd3ebfe56dea59b26f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6161
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
Reviewed-by: sterni <sternenseemann@systemli.org>
Fairly straightforward, handling the optimised representations
manually and otherwise delegating to BTreeMap.
Note that parsing of raw identifiers is not yet implemented.
Encountering an identifier node usually means that there is locals
access going on, so we need a special case for compiling a node in
such a way that an identifier's literal value ends up on the stack.
Change-Id: I13fbab7ac657b17ef3f4c5859fe737c321890c8a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6158
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Reviewed-by: grfn <grfn@gws.fyi>
This operation is required because both sides of the logical operators
are strictly evaluated by Nix, even if the resulting value is not used
further.
For example, in our implementation of `&&`, if the left-hand side is
`true`, then the result of the expression is simply the right-hand
side value. This value must be asserted to be a boolean for the
semantics of the language to work correctly.
Change-Id: I34f5364f2a444753fa1d8b0a1a2b2d9cdf7c6700
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6157
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Reviewed-by: grfn <grfn@gws.fyi>
These expressions use simple jumps to skip the correct expression
conditionally in the bytecode by advancing the instruction pointer.
Note that these expressions are already covered by a test behind the
`nix_tests` feature flag, but adding more is probably sensible.
Change-Id: Ibe0eba95d216321c883d3b6b5816e2ab6fe7eef1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6148
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
Reviewed-by: sterni <sternenseemann@systemli.org>
This is accomplished by simply delegating to the Rust implementations
of (Partial)Ord and (Partial)Eq, which are implemented for Value and
underlying wrapper types to behave like they do in Nix.
To ease the implementation overhead, a new comparison operator macro
has been added to the VM module.
Incomparable types will raise a new error variant when a comparison is
attempted, containing both supplied types. This mimics the information
carried in the error thrown by C++ Nix.
Change-Id: Ia19634d69119d40722f3ca672387bc3a80096998
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6143
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
The underlying implementation does a few tricks based on which pair of
attrset representations is encountered.
Particularly the effect of short-circuiting the empty cases might be
relevant in nixpkgs/NixOS, due to the use of lib.optionalAttrs.
Change-Id: I22b978b1c69af12926489a71087c6a6219c012f3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6140
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
This is required for constructing nested attribute sets at runtime.
There'll be quite a lot of optimisation potential with this solution
eventually, if it should turn out to be a bottleneck.
This introduces a conceptual change, in that the `Value` enum is now
an enum representing "all runtime values" instead of "all Nix language
types". This makes sense in general, as this type will also contain
Chunk representations etc. which are not exposed to users.
Change-Id: Ic5f72b2a0965b146c6a451efad34c6a81ca1aad8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6103
Reviewed-by: grfn <grfn@gws.fyi>
Tested-by: BuildkiteCI
This adds a new instruction which assembles an interpolated string
from a specified number of fragments, which are already going to be
located on the stack in the right position.
This will raise a type error if any of the fragments do not evaluate
to a string.
Change-Id: I5756248fa3e9fcc3d063c14db40b332f7e20a588
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6098
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
Implements attribute set literals without nesting. Technically this
already supports dynamic key fragments (evaluating to strings), though
the only way to create these (interpolation) is not yet implemented.
However, creating simple attribute sets like `{ }`, or `{ a = 15; }`
or `{ a = 10 * 2; }` works.
Recursive attribute sets are not yet implemented as we do not have any
kind of scope access yet anyways.
This is implemented using a new instruction that creates an attribute
set with a given number of elements by popping key/value pairs off the
stack.
Change-Id: I0f9aac7a131a112d3f66b131297686b38aaeddf2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6091
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
Implements simple arithmetic operations (+, -, *, /).
There is some scaffolding included to pop and coerce pairs of numbers,
as the Nix language will let arithmetic operators apply to arbitrary
pairs of number types (always resulting in floats if the types are
mixed).
Change-Id: I5f62c363bdea8baa6ef812cc64c5406759d257cf
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6074
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
