With this change, the runtime can correctly capture deferred upvalues.
Change-Id: I1e43b7b1ac2553b1812424adfc8bd08ef77bf1ea
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6339
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
When encountering a deferred local upvalue, the compiler will now mark
the corresponding local as needing a finaliser which makes it possible
to emit the OpFinalise instruction for this stack slot a little bit
down the line.
Change-Id: I3962066f10fc6c6e1472722b8bdb415a811e0740
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6338
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>
This will leave a sentinel value in the upvalue slot in which the
actual value is to be captured after resolution once a scope is fully
set up.
Change-Id: I12b37b0dc8d32603b03e675c3bd039468e70b354
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6336
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
While compiling local bindings, we now know the stack slot of the
value currently being compiled.
This will let us determine whether an upvalue can be captured directly
or whether it needs to wait for a synchronisation point at which the
upvalue can be instantiated.
This machinery lets us avoid unnecessary work at runtime when
instantiating closures that actually do not need complicated recursive
resolution.
This change itself introduces no new functionality, but since the
threading is noisy it is split out as a separate change.
Change-Id: I847c677ee8f6725fda1d2efd689b6a58bdccb779
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6328
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Instead of looking up the local to be initialised by its name again,
we can simply track the index at which it was declared from the point
where the declaration was made.
This reduces some string cloning and removes unnecessary logic. It
also theoretically makes the *current* index available during locals
compilation, which can be used to optimise some recursion cases.
Change-Id: I06f403603d4f86c3d319debfe74b5a52eec00990
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6327
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This actually makes things full-circle, as this tree already had this
implementation once before all the other required components were in
place.
With this commit, the compiler can resolve recursive upvalues within
the same scope (though they will not yet work at runtime).
Change-Id: I6267e477d08f367257c3a6dde054b880d7b47211
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6326
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
In order to resolve recursive references correctly, these two can not
be initialised the same way as a potentially large number of (nested!)
locals can be declared without initialising their depth.
This would lead to issues with detecting things like shadowed
variables, so making both bits explicit is preferable.
Change-Id: I100cdf1724faa4a2b5a0748429841cf8ef206252
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6325
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
The previous closure refactoring introduced a bug in which the same
closure object would get mutated constantly for each instance of a
closure, which is incorrect behaviour.
This commit instead introduces an explicit new Value variant for the
internal "blueprint" that the compiler generates (essentially just the
lambda) and uses this variant to construct the closure at runtime.
If the blueprint ever leaks out to a user somehow that is a critical
bug and tvix-eval will panic.
As a ~treat~ test for this, the fibonacci function is being used as it
is a self-recursive closure (i.e. different instantiations of the same
"blueprint") getting called with different values and it's good to
have it around.
Change-Id: I485de675e9bb0c599ed7d5dc0f001eb34ab4c15f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6323
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
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>
With this change, encountering a dynamic upvalue will thread through
all contexts starting from the lowest context that has a non-empty
`with`-stack.
The additional upvalues are not actually used yet, so the effective
behaviour remains mostly the same. This is done in preparation for an
upcoming change, which will implement proper dynamic resolution for
complex cases of nested dynamic upvalues.
Yes, this whole upvalue + dynamic values thing is a little bit
mind-bending, but we would like to not give up being able to resolve a
large chunk of the scoping behaviour statically.
Change-Id: Ia58cdd47d79212390a6503ef13cef46b6b3e19a2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6321
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This is a common idiom in both Nix and other languages when a local is
declared without actually being used.
Since Tvix warns for unused locals, having this available is useful
and can be included in the final error message as a suggestion if an
unused variable is intentional.
Change-Id: Ia85f704ba183499a3bae657c58166e2e29f9bde5
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6320
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This does not yet correctly resolve them if they are more than one
scope up, however.
Change-Id: I6687073c60aee0282f2b6ffc98b34c1e96a60f20
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6319
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
With this change, it becomes possible for functions to call themselves
as they are being defined in local bindings.
Change-Id: Ib46a39ba17b1452b5673d96fa729d633d237241a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6314
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This is required to efficiently construct the upvalue array at
runtime, as there are situations where during Closure construction
multiple things already have a reference to the closure (e.g. a
self-reference).
Change-Id: I35263b845fdc695dc873de489f5168d39b370f6a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6312
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
`builtins.getFlake` doesn't interest us, of course, but some others may
be worth (or easy) to implement. They are pretty low priority, though,
since nixpkgs has compatiblity wrappers for the ones it uses.
The new debugging-related builtins (break and traceVerbose) are
interesting to note, but may not make sense to implement at all.
Change-Id: Icae547aa3bd9d6ee6b87897ba8210eb9b9b044c7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6332
Tested-by: BuildkiteCI
Autosubmit: sterni <sternenseemann@systemli.org>
Reviewed-by: tazjin <tazjin@tvl.su>
This avoids unnecessary rebuilds of //tvix/eval when working on it
locally using impure cargo.
Change-Id: I028033a6345a9655e2877a534448706b8f85a1a1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6317
Tested-by: BuildkiteCI
Autosubmit: sterni <sternenseemann@systemli.org>
Reviewed-by: tazjin <tazjin@tvl.su>
The `With` struct no longer contained any internals after the cleanup
logic for the stack had been moved into Compiler::compile_with,
leaving the `Vec<With>` to essentially act as a counter for the number
of things on the with stack.
That's inefficient of course, so with this commit it actually becomes
an integer (with an encapsulated API within scope::Scope).
Change-Id: I67a00987fc8b46b30d369a96d41e83c8af5b1998
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6311
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
The compiler module is getting quite long and this will help keep some
order.
Right now the scope internals are not very well encapsulated; this
paves a way to reducing the API surface of the `scope` type to the
things that are actually used by the compiler instead of giving access
to its internals.
Change-Id: I8c16c26d263f018baa263f395c9cd80715199241
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6310
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This has no effect yet, other than changing the way in which some
upvalue captures break (that are already not working correctly).
However, after this change the compiler correctly detects
self-recursion and can start emitting the instructions to deal with
this at runtime.
Change-Id: Id3b0ac206c0204739597a4325bcc66f9c806c242
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6309
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Calculating the with_idx (i.e. the stack offset of the "phantom"
variable from which a `with` dynamically reads at runtime) needs to
account for unitialised variables the same way as the resolution of
normal locals does.
Change-Id: I9ffe404535bf1c3cb5dfe8d9e005798c857fff94
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6308
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This is required to correctly clean up the `with` values.
At the moment, the attrset from which identifiers are being taken is
always pushed on the stack. This means that it must also be removed
again, otherwise in an expression like
with { a = 15; }; a
The final stack is `[ { a = 15; } 15 ]` *after the last operation*,
which means that the attrset is still on there as garbage.
This has little practical impact right now because it is always
shadowed by the fact that the actual expression value is at the right
location, but becomes relevant when accounting for upvalue captures.
Change-Id: I69e9745bfaa4d6bbcb60ee71f4dc3f8d8695d16a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6303
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This extends the logic of `Scope::resolve_local` to detect cases where
self-recursion is occuring (i.e. an identifier is being accessed in
its own identifier).
These cases are not yet handled specially, and the logic of when
things are marked initialised (which was previously always at the same
spot as their declaration) has not changed, making this commit a
runtime no-op for now.
Change-Id: I3179642a7c55869ad4465fdd2678b0cd51a20f15
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6302
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Nix does not allow things like `let a = 1; a = 2; in a`, but doing it
across depths is allowed.
Change-Id: I6a259f8b01a254b433b58c736e245c9c764641b6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6301
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
This does not yet change anything semantically, but will be useful for
resolving simple cases of self-recursion etc.
Change-Id: I139ecb7e4a8a81193774392a96e73e0ea6b9f85d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6300
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>
The previous implementation of OpResolveWith manually controlled the
loop iteration, which skipped over the disassembler's tracing
instruction.
Instead, the resolution of dynamic variables has been delegated to a
new helper function. This has the additional benefit that the loop
labels are no longer required, making things a bit cleaner.
Change-Id: If22b74c3d49c74bf3a1ec4497cb761a9ee6cf2a4
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6298
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Instead of tying the popping of the with stack to scope depth, clean
up the stack immediately after processing a with body.
The previous behaviour was actually incorrect, as it would leave
things on the with-stack longer than they were supposed to be there.
This could lead to false positive resolutions in some situations
involving closures.
Change-Id: I7b0638557503f1f71eb602e3d5ff193cdfcb67cc
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6297
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Implements the final bit of logic remaining for wiring up closures,
which is the runtime construction of closure objects.
When encountering an OpClosure, the VM walks through the bytecode
collecting all the upvalue location operands (see commit introducing
the OpCode::Data* variants for details) and stores the runtime values
in the new closures upvalue vector.
After that, the handling of the closure itself becomes functionally
identical to that of lambdas.
With this initial implementation of closures there are several large
optimisation potentials available, the two most notable ones are:
- Distinguish the runtime representation of lambdas and closures
explicitly.
- Detect and handle multiple-arity functions directly in the compiler.
However, for both of these we should wait until we have appropriate
benchmarking infrastructure in place. This is because our test
implementations have shown that the complexity of either of these
changes is quite significant, and we do not yet know if they really
pay off.
Change-Id: I077e977810fd5cb2b1ecd7f1a119e728025dd786
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6295
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
This resolves an upvalue at runtime by pushing it on the stack from
the closure's upvalue vector.
Change-Id: Ic3e7a7ecd9f7032f679114a1995e5bbf83062fcf
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6294
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
In preparation for implementing calling of closures, store a closure
directly in the VMs call frame.
Change-Id: Iad24cd8c49fee4ebd4d0c84ffaa4c2505ee3dfd6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6293
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
... same as the others
Change-Id: I9c8868388c10b0b6484c5bdd3799d801296c6979
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6292
Tested-by: BuildkiteCI
Reviewed-by: grfn <grfn@gws.fyi>
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>
grfn pointed out in cl/6174 that `Result` might cause developers to
believe that this behaves like std::Result, which it does not.
Change-Id: Ia30ab0dcb7e8da7bf842777ee3fe17bcf35cb0c1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6281
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
Refactors the update function to take the attribute sets by value
instead.
To facilitate this, we use an equivalent of the currently unstable
`Rc::clone_or_unwrap` in the VM when encountering attribute sets, so
that in cases where the only references to the attrs being updated are
the ones on the stack those clones are avoided completely.
This does make update() a little bit more tricky internally, as some
optimised branches can directly return the moved value, and others
need to destructure with ownership. For this reason there are now two
different match statements handling the different ownership cases.
Change-Id: Ia77d3ba5c86afb75b9f1f51758bda61729ba5aab
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6279
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
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>
