When impl lookup results in an `ImplWitness`, either directly or later
through `LookupImplWitness` evaluating to an `ImplWitness` inside some
specific, it represents a dependency on the eval blocks of the
`ImplDecl` that the `ImplWitness` comes from. The `ImplWitness` contains
a `SpecificId` of the `ImplDecl`, and the constants in the witness table
must be viewed through that `SpecificId`. The constants may depend on
the declaration or the definition block of the `ImplDecl`'s specific, so
any use of the witness table from `ImplWitness` requires the specific
from the `ImplWitness` to be fully resolved: both the decl and
definition blocks.
The specific declaration block is always resolved immediately when
forming a specific id, so there's nothing to do there. And things are
simple when the `ImplWitness` returned from impl lookup is concrete. We
would simply resolve the specific definition directly in the evaluation
of `LookupImplWitness`.
But when the `ImplWitness` is symbolic, meaning there's a dependency on
generic parameters in its referenced specific, then we have a dependency
on resolving the specific definition in the `ImplWitness` for every
specific that gets applied to the `ImplWitness`, eventually resulting in
a concrete specific.
Previously we attempted to do this by representing a dependency edge
from the eval block containing the `ImplWitness` instruction to the
definition block of the specific inside the `ImplWitness`. We
represented this dependency edge with the `RequireSpecificDefinition`
instruction.
The challenge with the `RequireSpecificDefinition` instruction was
ensuring it always landed in the _same eval block_ as the one holding
the symbolic `ImplWitness`. Ultimately, we've found a way to do this
without the `RequireSpecificDefinition`. The witness table is accessed
through `ImplWitnessAccess`, so we resolve the specific definition
before each access to the witness table. If the definition is already
resolved, this just costs a lookup of the Specific from the value store.
We document the other things we tried, which don't work or were too
complex, below.
In PR #6236, we added the `RequireSpecificDefinition` instruction in the
eval of `LookupImplWitness`, but this is too late. If impl lookup
resolved to a symbolic `LookupImplWitness`, the `LookupImplWitness` is
added some eval block. Later, we form specifics and re-evaluate that
eval block, and the `LookupImplWitness` can evaluate to an
`ImplWitness`. Adding the `RequireSpecificDefinition` instruction at
this point inserts the instruction into whatever generic's eval block
happens to be on the top of the generic stack, not the eval block of the
generic we are building a specific for. And indeed we can't add
instructions to a generic's eval block later like this, as it would be
missing from specifics that were already resolved for the generic. This
incorrect insertion shows up as a crash in eval at the CHECK introduced
but commented out in [eval.cpp in
#7981](https://github.com/carbon-language/carbon-lang/blob/56d9bbc40add1e9263f2f9b2555c3a27292d1725/toolchain/check/eval.cpp#L3507-L3512).
This is covered by the
`require_specific_impl_definition_through_specific_instantiation.carbon`
test.
We tried moving the insertion of `RequireSpecificDefinition` to be
earlier - at the point where the `LookupImplWitness` is constructed, in
the `LookupImplWitness()` function, instead of inside evaluation. This
seemed like a straightforwardly correct answer, as the eval block that
contains the `LookupImplWitness` will then also contain the
`RequireSpecificDefinition`. However, this can cause us to lose the
`RequireSpecificDefinition` entirely. Because it's a separate
instruction, not an operand of `LookupImplWitness`, it does not get
treated as a dependency of the witness. When we do an impl lookup inside
an `alias`, we build a generic for the alias and then discard it. This
discards the eval block containing the `LookupImplWitness`. Then, in an
enclosing generic, we refer to the constructed `AliasDecl` instruction,
which pulls all its dependent operands into the enclosing eval block.
This pulls in the `LookupImplWitness` into the enclosing eval block, but
it does not preserve the `RequireSpecificDefinition`. This is covered by
the
`require_specific_impl_definition_from_lookup_impl_witness_in_class_definition.carbon`
test.
We tried having `ImplWitness` evaluation resolve the specific definition
of the specific `ImplDecl` it refers to. This seems like the simplest
approach by far as we should always be working with the correct eval
block and nothing can get lost. However it does not work because of the
recursive nature of `ImplWitness`. The `ImplWitness` instruction appears
in the declaration of an `ImplDecl`, and the `ImplWitness` holds a
`SpecificId` for that same `ImplDecl`. When building a specific
`ImplDecl`, we evaluate the `ImplWitness` in the `ImplDecl`'s
declaration block which then immediately resolves the specific
definition before the declaration is finished. This leads to missing
instructions in the specific definition. In short, this puts the
dependency edge to the `ImplDecl`'s specific inside the `ImplDecl`
itself, which is problematic.
We tried wrapping `ImplWitness` in another instruction,
`ImplWitnessUse`. Inside the `ImplDecl` declaration there is still
`ImplWitness`, but then in the definition there is `ImplWitnessUse` and
that is what is returned from the evaluation of `LookupImplWitness`. The
`ImplWitnessUse` is an `ImplWitness` plus it ensures the specific
referred to by the witness has its definition resolved. However this
breaks constant comparisons, as we get different witnesses inside an
impl versus outside.
The first iteration of this PR's solution had us insert the
`RequireSpecificDefinition` into the eval block when we sniffed a
`LookupImplWitness` instruction, however that missed cases where impl
lookup immediately resolved to a symbolic `ImplWitness`, which is why we
now sniff for the `ImplWitness` in the eval block. This is covered by
the
`require_specific_impl_definition_from_symbolic_impl_in_class_definition.carbon`
test.
We also originally made a mistake of not inserting
`RequireSpecificDefinition` when inside the impl definition that the
`ImplWitness` refers to. But we can form specifics ImplWitness values
through function signatures, and these must add the dependency edge on
that specific ImplDecl definition. This is covered by the
`require_specific_impl_definition_from_inside_impl_definition.carbon`
test.
Finally, we did have a working solution. It was to sniff instructions
added to the eval block of a generic, and when an `ImplWitness` is added
to the eval block (it is implied that it is symbolic), also add a
`RequireSpecificDefinition` to that eval block immediately. We also
resolved the definition of a specific referred to by `ImplWitness` when
`LookupImplWitness` evaluates to `ImplWitness` to cover the concrete
witness case. This complicated the construction of the eval block
immensely, however. It caused a situation where adding one inst to the
eval block actually adds two, which we needed to avoid when importing.
And it required inst-kind specific behaviour in the very general
construction of an eval block, completely outside of eval itself.
Determining that a type has trivial destruction requires checking
whether all of its subobjects have trivial destruction and whether its
type has a custom witness for `Destroy.Op`. This isn't possible when
`Destroy` is the custodian for `Destroy.Op` and `SubobjectDestroy.Op`,
so we need to make `SubobjectDestroy` its own interface.
`Destroy.SelfDestruct` becomes a standalone function because it depends
on `SubobjectDestroy` and `Destroy` being implemented, and is also
responsible for generating any custom witnesses for both interfaces.
Making `SelfDestruct` an interface was experimented with, but this
required a blanket impl that caused issues for facet types.
`Destroy` depends on `SubobjectDestroy` to be implemented. We'd ideally
use `require impls SubobjectDestroy`, but it isn't complete yet, and it
also causes problems when `Destroy` is implemented inside the class
definition. We work around this by temporarily renaming the `Destroy`
interface to `DestroyImpl`, and introducing a constraint called
`Destroy` so that user code doesn't need to write `Core.SubobjectDestroy
& Core.Destroy`.
Diagnostic rendering was part of the data model: `Loc::FormatLocation`
and
`Loc::FormatSnippet` wrote plain text to a stream, so no capability,
color, or
terminal width could reach them.
A diagnostic is now one message with labels, contexts, and location
information
attached, and `Diagnostics::Renderer` draws it into a
`Terminal::Buffer`. The
design is in `toolchain/docs/diagnostics_rendering.md` (#7668),
authoring in
`toolchain/docs/diagnostics.md`.
Before:
```
example.carbon:6:3: error: 1 argument passed to function expecting 0 arguments
Run0(1);
^~~~~~~
example.carbon:3:1: note: calling function declared here
fn Run0() {}
^~~~~~~~~~~
```
After:
```
error: 1 argument passed to function expecting 0 arguments
╭─┤ example.carbon:6:3
┆
3 │ fn Run0() {}
· ─────┬─────
· ╰──┤ calling function declared here
┆
-> 6 │ Run0(1);
· ───────
╰────────────
```
`--no-diagnostic-snippets` gives a compact form, `example.carbon:6:3-9`.
File
tests use it; the rendering tests under `toolchain/diagnostics/testdata`
ask for
snippets.
C++ interop diagnostics are data now rather than text Clang drew:
`SemIR::ClangSourceLocId` names a `clang::SourceRange`, and a Clang
location is
marked as the point it names.
`StreamConsumer` moves to a `:renderer` target that `:emitter` does not
depend
on, keeping `common/terminal` out of `consumer.h`. Capabilities come
from the
file behind the error stream, so tests get plain ASCII.
`renderer_fuzzer` draws diagnostics built from fuzzer bytes with every
capability combination, since a shape the renderer doesn't expect must
never be
why the compiler dies while reporting a problem.
TODOs mark the follow-ups: line weights and crossing lines, wrapping
that keeps
a quoted type whole, and ordering diagnostics that share a byte offset.
Also: `IncorrectExtension`'s hint moves into its message, since its note
had no
location of its own, and `strip_noise` in `check_diagnostics.py` blanks
multi-line `#define`s instead of deleting them.
Assisted-by: Claude Code
When identifying a facet type, defer evaluation of LookupImplWitnesses
by redirecting them to evaluate into IdentifiedWitness. Then the caller
of identify will need to substitute the IdentifiedWitnesses, either with
witnesses of their own, or with LookupImplWitness to cause a lookup
operation.
This allows:
- An `impl as` on a named constraint can subst the witnesses in
designators from the named constraint with a witness to the impl itself.
This prevents them from looking for and finding some other overlapping
impl. (This PR)
- Impl lookup to collect rewrite/equality requirements from all named
constraints and then use the impls it has already found to evaluate any
designators in those requirements instead of performing impl lookup
again (creating a cycle). (Done in
https://github.com/carbon-language/carbon-lang/pull/7944)
- Similarly, an `impl as` can collect impls and equality requirements
from all named constraints and verify them correctly by replacing
witnesses that should refer to the impl with the correct witness instead
of performing a lookup and finding an overlapping impl. (Done in
https://github.com/carbon-language/carbon-lang/pull/7937)
- ImplWitnessAccess to be able to find rewrites from a named constraint,
not just from the initial DeclaredFacetType. (TBD)
To do this we need to introduce the concept of "identified" specifics.
In order to identify a facet type, we must construct a specific for each
require decl in a named constraint. These specifics will have
IdentifiedWitnesses inside them, but their args would otherwise match a
specific for the require decl constructed in another context. To avoid
polluting the canonical specific namespace, we introduce a second
namespace of specifics that are "identified" since they are created by
identifying.
As we have introduced a new global state that modifies eval behaviour,
we must disable that state when resolving a specific (as per
https://github.com/carbon-language/carbon-lang/pull/7923). But inside an
"identified" specific, we preserve the behaviour so that the witnesses
modified by identify (through SubstPeriodSelf or through evaluating a
require specific) have their evaluation deferred.
A specific's definition block may contain `InstId::None`, indicating it
has no constant value. Display this as `<not constant>` instead of
`invalid`, to avoid it sounding like an error.
Add support for deferring initialization as a template action, and
performing the deferred initialization during template instantiation.
This is substantially more complex than other conversion actions, for
two primary reasons:
* The initializer in the generic may have storage arguments as inputs.
We model an initializing expression as having a "slot" where
initialization writes the location that should be initialized by that
initializing expression, and that needs to be an output of the
initialization action.
* Initialization from a tuple or struct literal needs to recurse into
that literal, and the literal will have been spelled in the generic,
meaning we don't have an `InstId` that can be used to name the specific
version of the initializer as input for nested conversions.
These issues are addressed by introducing two new features to the action
machinery:
In addition to `InstAction`, we now have `MultiInstAction`, which is an
action that produces a tuple of instruction values instead of a single
instruction value. Initialization actions produce one instruction for
the final result, which is spliced at the point of initialization, plus
one instruction for each storage argument, which are spliced into the
storage argument slots in the original generic. During initialization,
if we find one of those splices in the storage argument of an
initializing expression, we return the new storage argument back to the
initialization action to be included in the specific, instead of
overwriting the storage argument in the generic.
Actions whose `PerforrmAction` takes a `SpecificId` as input no longer
perform automatic refinement of their operands to specific instructions.
Instead, the action is given control over when and where it performs
that refinement. In `InitializeAction`, we use this freedom to form a
`SpecificInst` for the initializer in the primary output block, and form
a `SpecificInst` for the target in the target block. When detecting
whether we are initializing from a tuple or struct literal, we step over
the `SpecificInst` and track its `SpecificId`, and if necessary create a
new `SpecificInst` wrapping the sub-initializer when we recurse into the
nested element conversion.
Assisted-by: Claude and Gemini via Antigravity
This change partially implements [PR #7362], which revises how objects
are destroyed. It is a partial implementation for two reasons:
1. This change moves `Destroy.Op`'s current behaviour into
`Destroy.SubobjectDestroy`, but it doesn't add support for objects with
non-trivial destruction.
2. `Destroy.SubobjectDestroy` is a workaround for `require impls
SubobjectDestroy`. We aren't able to use the latter until the dependents
add their requirements' implementations to their own witness tables.
[PR #7362]: https://github.com/carbon-language/carbon-lang/pulls/7362
The type `type` is now a `FacetType` inst with no constraints. This
brings the model implemented in the toolchain into better alignment with
the language design. The `SemIR::TypeType` struct remains as a scope for
holding the `TypeInstId`, `ConstantId`, and `TypeId` constants, but is
not an `InstKind` anymore.
The `TypeType` inst looks a lot like singletons, but there are many
`FacetType` insts so it doesn't quite fit that model. So we put it
alongside singletons with a fixed inst id but refer to it as a more
general "builtin" inst that is not a singleton.
`Namespace::PackageInstId` is similar, and we group it with `TypeType`
conceptually as another builtin instruction with a fixed id.
No conversion is needed anymore to use a `type` as a facet, since types
also have a `FacetType` type. This simplifies and removes a number of
helpers and branches throughout the code.
The `TypeType` inst is now part of the constant store, so we end up
printing it in the constants block in every test. But it's also named
`type` rather than `%type` to preserve the majority of existing
formatting behaviour, though this does look different from other
constants.
Assisted-by: Opus 5 was used to generate a first draft and validate the
refactoring. Though nearly everything non-trivial the tool wrote has
been modified or rewritten.
This action was created to wrap any `MetaInstId` operand of an action
instruction. This served two purposes:
1) It had a special hook in `OperandIsDependent` to allow it to be
performed while it had a dependent operand (the reference to the
instruction in the generic).
2) It created a `specific_inst` so that the downstream action saw an
instruction in the specific instead of one in the generic.
These are both replaced: the special case in `OperandIsDependent` for
`RefineInstAction` is replaced by a special case for `MetaInstId`s in
general, and the `SpecificInst` is now created as part of performing the
downstream action, rather than as a separate step carried out
beforehand.
This simplifies the produced SemIR and reduces the number of splices
significantly. It also prepares us to handle actions like
initialization, where we don't actually want to create `SpecificInst`s
immediately in the location where the action is performed, because they
actually belong somewhere else in the IR.
Assisted-by: Claude Opus 5 and Gemini via Antigravity
Add basic support for lowering templates: we can now lower `SpliceInst`
in the case where the generic and specific are from the same file (and
we don't support importing templates from other files yet in general).
In order for this to work, lowering needs to be able to query the
expression category, and to handle instructions that appear to be
(template) constants in the generic but turn out to be non-constant in
the specific, so support for that is added.
Switch `type_of_inst` from being added as an action inst to being added
as a normal inst, since it's not an action and the old approach led to a
crash in lowering.
Replace `refine_type_action` with `refine_inst_action`, and generate a
`specific_inst` instead of an `as_compatible` to represent the specific
version of an instruction that's used as an input to a template action.
This gives us a place to handle other properties of the instruction that
might vary from generic to specific beyond its type, such as its
constant value and its expression category.
For now, we provide a non-template-dependent constant value to the
`specific_inst` in addition to the non-template-dependent type we have
traditionally provided. This doesn't seem to matter for any current
actions, but sets us up to better handle future actions. The
`specific_inst` representation also allows downstream consumers of the
instruction to track which specific they should be requesting
information from. Providing a correct expression category for
`specific_inst` will be handled in a future PR.
`TemplateInst` wraps another inst. If that inst is symbolic, it is
treated as a template by `OperandDependence`.
Use this to replace `CallCppTemplateAction` with the more general
`CallAction`.
`GetCallee` is sometimes called on a spliced instruction, and is
checking its exact inst operand to see if it's a `BoundMethod`. This
fails if the `BoundMethod` is wrapped in another instruction, such as a
splice. Normally our approach for such a situation would be to
constant-evaluate the operand, but that doesn't work here because the
`BoundMethod` will be non-constant if its bound `self` is. So instead we
now step through splice instructions manually when looking for the
`BoundMethod`.
Calls with template callee or args can now be deferred via an
InstAction. This allows code like this to check:
```carbon
import Cpp inline '''
template<typename T>
struct C {};
''';
fn F(generic T: type) {
let unused c: Cpp.C(T) = Cpp.C(T).C();
}
```
Example:
```carbon
import Cpp library "<vector>";
class C(T: type) {
var v: Cpp.std.vector(T);
}
inline Cpp '''
void F() {
Carbon::C<int> c;
c.v.push_back(123);
std::cout << c.v.back() << std::endl;
}
''';
```
A new `CallCppTemplateAction` is used to delay performing the C++
template call until non-symbolic arguments are known.
Implementing interface modifiers causes an infinite loop when generating
fingerprints because the witness value generates a fingerprint that's
dependent on something dependent on the witness value. We've debugged
this to the witness table's `elements_id` field.
This hack is a workaround for creating a new block type whose value is
not codependent with its identity.
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Add `Core.CppCompat.[U]Long64` to represent a 64-bit long that is not
`i64`. Treat it as being "just slightly smaller than" `i64`, like we
treat `Core.CppCompat.LongLong64` as being "just slightly larger than"
`i64`, so that we get implicit conversions `Cpp.long` -> `i64` ->
`Cpp.long_long` on all targets.
This follows the direction of proposal #5448, and seems like the obvious
extension of the `[U]Long32` and `[U]LongLong64` types added in #6275
for targets of this "shape".
Assisted-by: Gemini via Antigravity
This adds SemIR structs and implements building `observe` lists, as well
as naming, formatting, and importing `observe` declarations.
---------
Co-authored-by: Dana Jansens <danakj@orodu.net>
In some cases the pattern block can depend on the initializer, so it
must be sequenced after it. See #7469 for a more detailed explanation of
why this is necessary.
The bulk of this change is changing most pattern insts to be `Always`
rather than `AlwaysUnique` constants, so that they can be wrapped in
`SpecificConstant`s to perform substitution. That then lets thunking
rely much more on `SpecificConstant` wrappers instead of deep-copying
the inst tree with modified types.
This approach to thunking should scale better, particularly as things
like form generics make function signatures more complex, because we can
leverage the existing support for constant evaluation and substitution.
Unfortunately, applying this approach to binding patterns will require
more work; see the TODO near the top of `thunk.cpp` for details.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Fix import logic to make all imported packages be children of the
`NameScopeId::Package` scope. Previously, indirectly-imported packages
would end up as children of their importing package's scope, which
resulted in them not being treated as packages at all, and in particular
not being fingerprinted as packages.
Fixing that caused a failure in the fingerprinting logic as we started
to encounter packages with no correspoding import scopes. Instead of
looking for import scopes, use a simpler mechanism to map packages to
their package names, and clean up.
Unfortunately the latter change churns all the fingerprints again :(
Hopefully this is the last time for a while.
When we import from another library in the same package, its entities
end up with our library as their parent scope, resulting in cross-file
fingerprint mismatches. Instead, only include the library ID when
fingerprinting either a package-private entity or an `ImportIRId` that
refers to a particular `SemIR::File`.
Carbon-side thunks (for example the `Copy`/`Destroy` witness thunks
generated for imported C++ types) are mangled by Carbon, and their names
incorporate a fingerprint of the involved types. The instruction
fingerprinter identifies a class only by its name and parent scope,
which is sufficient for Carbon classes but not for imported C++ classes:
different specializations of one class template (and other cases such as
types in anonymous namespaces) share a Carbon name and parent scope. As
a result, the thunks for two distinct specializations could mangle to
the same name, producing a single LLVM function with two definitions and
failing `verifyModule` during lowering.
When fingerprinting a class imported from C++, also include the Clang
mangled name of its type.
Test: toolchain/lower/testdata/interop/cpp/thunks.carbon gains a split
with two specializations of one class template, each requiring a thunk;
their thunks now get distinct mangled names instead of colliding.
Assisted-by: Claude Code
---------
Co-authored-by: Christopher Di Bella <cjdb.ns@gmail.com>
Fixes link failures when referencing a symbol involving a fingerprint
from a different package.
Previously we included the `Namespace`'s `import_id` as part of its
fingerprint, which caused local and imported namespaces to get different
fingerprints. We now store the `import_id` on the `NameScope` instead of
on the `Namespace` inst to avoid this problem.
Also, when we reach a package-level `NameScopeId`, consistently
fingerprint it as a (package name, library name) pair. Previously the
fingerprinting depended on whether it was imported or not, as an
imported `NameScopeId` had a parent scope (the current package). We need
to include the library name here so that private entities with the same
name in different libraries have different fingerprints.
Provide a source location to the member expression used when a thunk
calls a member function. This ends up being used as the point of
instantiation when the return type triggers a template instantiation;
the absence of this location previously caused assertion failures within
clang.
Assisted-by: Gemini via Antigravity
See
[here](https://docs.google.com/document/d/1rWcueFwIfZox6GKVGxiUG4cBzjrZ6djXiIDGyJDtrE4/edit?tab=t.0)
for the design doc.
This also removes the default value of the `result_type_inst_id`
parameter of `HandleAction`, moves it before the action in the parameter
list, and documents it. This solves two problems:
- The default made it easy to forget, leading to unnecessary
`TypeOfInst` instructions.
- When it was present, putting it after the fairly "bulky" action
argument tended to make the callsite harder to read.