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.
Previously, we picked a single Carbon parameter pattern for each C++
parameter pattern. This doesn't work well in cases where the Carbon
semantics and the C++ semantics are not perfectly aligned. In
particular, when a parameter is passed by value in C++, that might mean
either pass-by-move (which in Carbon would best be modeled by a `var`
pattern, as no other form of parameter would perform a move) or
pass-by-copy (which in Carbon would best be modeled by a value
parameter, as a `var` parameter would force an extra copy).
After this change, we compute a passing mode for each parameter based on
the implicit conversion sequence from the argument to the parameter as
determined by C++ overload resolution, and use that to determine the
Carbon pattern corresponding to each C++ parameter. This results in
potentially generating multiple different thunks for the same C++
function if it's called in different ways, but we already did that to
handle default arguments and list-initialization. The passing modes are
included in the thunk mangling.
Add a new value store for clang decl signatures, which capture the
information about parameter passing mode as well as the other existing
information about different ways that a C++ function might be imported
to Carbon.
Most of the rules for computing passing modes are the same as before:
const references use pass by value, non-const lvalue references use
pass-by-ref, non-const rvalue references use pass-by-var. But for C++
non-reference parameters, pick between pass-by-value and pass-by-var
based on whether the implicit conversion sequence was effectively
performing a copy. Prefer pass-by-value if either would work and they'd
do the same thing. We still use pass-by-value for const references, even
when the argument is an lvalue and we could pass a reference; we may
want to change this in future.
For virtual functions, we try to pick a worst-case passing mode, as we
can only pick a single signature for what goes in the vtable. Calls to
virtual functions will still use a thunk to C++, allowing variance in
the calling convention at call sites. We don't allow variance in the
overriders as we don't implement support for thunks for virtual
functions yet. We currently use pass-by-value for const reference
parameters here, but that should probably change at some point.
Assisted-by: Gemini via Antigravity
When importing a C++ function with an rvalue reference parameter, we
previously produced a Carbon value parameter. This would lead to the
toolchain believing it could pass the address of a non-expiring object
to the function, which would lead to a use-after-move.
Instead, we now map non-const rvalue reference parameters to Carbon
`var` parameters. This forces the object passed into C++ to be unique
and owned by the call. While that's not an exact match for C++ rvalue
reference parameters, given that it provides "always move" not
"conditionally move", it's the closest match we have at the moment.
This is a step toward removing the index from `InitForm`, so that equal
form values always have equal representations.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
This TODO had been written before C++ types were generating destroy
implementations, which is resolved now.
Assisted-by: Google Antigravity with Gemini 3 Flash
The primary change in this PR is to split the `Initializing` expression
category into separate `ReprInitializing` and `InPlaceInitializing`
categories, depending on whether initialization uses the types
initializing representation, or is guaranteed to be in place. It also
rationalizes and documents the SemIR-level semantics of those categories
(including where #5545's "ephemeral entire reference" category will
fit), and introduces two new inst kinds to close gaps exposed in the
process.
Some additional secondary changes:
- Consistently format the storage arguments of initializers with `to`,
regardless of whether initialization is in-place, and document the `to`
notation.
- Rename some inst kinds and functions, and restructure some of the
code, for clarity and consistency with the new documentation.
- Resolve a TODO to handle more category conversions in
`CategoryConverter`, in order to make it easier to reason about category
conversions.
See #6588 and the review history of this PR for background.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
The key changes are:
- Function output parameters are now prefixed with `out`, and more
consistently formatted as named parameters.
- Function and inst output arguments are now written as part of the inst
form, rather than as one of the inst arguments.
As a drive-by fix, this also changes `Temporary::storage_id` from
`DestInstId` to `InstId`, because it doesn't represent an output
parameter of the `Temporary` inst itself.
See the review of
[#6532](https://github.com/carbon-language/carbon-lang/pull/6532) and
[this Discord
discussion](https://discord.com/channels/655572317891461132/999638000126394370/1458268977020141589)
for additional background.
Pursuant to recent decisions on #6124, switch `Destroy` to use a
`CustomWitness` for its implementation. Right now this is manufacturing
no-op implementation functions on each lookup, which obviously isn't
ideal but is intended as a first pass. I'm mostly trying to find the
right balance between updating the approach to reflect new decisions,
while still breaking apart work in a way.
The `CoreInterface` logic is intended to build on `CoreIdentifier`
support. We have a number of additional interfaces that require
specialized logic, and that'll extend pretty far with C++ interop, so it
seemed easiest to have a generic function for it. That's what's
replacing the logic inside C++ interop that was doing string comparisons
(which could have already been moved to `CoreIdentifier`, I just missed
it in my first pass).
This adds `CustomWitness` support because the `Destroy` witnesses can be
imported cross-file. `CustomWitness` was previously only used for C++
types, which don't yet support import, which is why that wasn't
previously an issue. The addition of `query_specific_interface_id` is
similarly needed in order to get correct sorting of witness blocks when
imported.
This PR also removes builtin constraint logic (note this is in a
separate commit to help review; it's not a separate PR because it's
difficult to split apart without tests breaking). This had been made
generic with the expectation that destroy, copy, move, and conversions
would all need related support. Under the new decision, we are not going
to do blanket impls and will instead just manufacture a `CustomWitness`
for everything.
A lot of SemIR fingerprints change, but that's probably because the
addition of `Destroy` on core classes is yielding structural changes.
This is a prerequisite for support for interop with C++ template names.
No behavior change here, except that it sadly changes the fingerprinting
for a lot of tests.
For now, map C++ reference types to const-qualified Carbon pointer types
rather than picking between a (non-const) pointer or a value type. This
fixes misbehavior in lowering for reference members in classes and
reference return types.
Update the special-case handling for references as function parameters
so that it continues to map const reference parameters to Carbon
pass-by-value, and unify the code paths for `self` parameters and other
parameters, which were mostly doing the same thing but had some subtle
differences.
Add references to the list of types that we can pass to and from C++
directly, without needing an additional layer of thunks.
Remove duplication between determining whether a parameter needs custom
thunk mapping and whether a function needs a thunk. Now a function needs
a thunk if any parameter or the return type does.
This fixes some inconsistencies; previously:
- We would not require a thunk when passing an `unsigned int`, but if we
had a thunk we'd pass `unsigned int` indirectly.
- We would always require a thunk for an enum parameter, even though
we'd actually pass it directly if its underlying type is a 32- or
64-bit integer.
- We would require a thunk for a nullable pointer, even though
we arrange for all pointer types to have the same ABI in Carbon and
C++, including nullable pointers / Optional(T*).
This also causes us to use a thunk for rvalue reference return types,
which we used to miscompile.
Depends on #6276.
When deducing arguments for generic parameters of an `impl`, the
deduction calls `Convert` on the input arguments. Often, the input
argument is a facet, and needs to be converted to a type via
FacetAccessType in order to produce a different facet. These
instructions end up being added to the semir, but only their constant
values are needed for the resulting specific returned from Deduce.
In the best case, these extra instructions are just noise in the semir,
or they just cause instruction names to get differentiated with larger
suffixes.
In the worst case, these extra instructions contain references to
instructions from a generic context, and leak them out of that generic
context and into another. In particular, when importing a
LookupImplWitness instruction, the re-evaluation of it can do deduce
(when the lookup is against a generic `impl`). The instructions created
in Deduce are not part of the import, and end up referring to imported
instructions from the local context, which leads to confusion in the
toolchain, and can crash.
The `import_self_specific.carbon` test demonstrates this. It causes the
`I.F` function to be imported from the `I` interface when building the
witness table for the `impl`. Doing so imports the specific of `C` which
includes a LookupImplWitness for `Self.Accoc` in `I`. The `Self` is a
BindSymbolicName with generic binding index 0, in `I`. When Convert
creates instructions in the generic `impl forall D`, however, they end
up referencing and including this BindSymbolicName into its eval block.
But the generic binding 0 in the `impl` is a very different thing (a
value of type `E`). This confusion leads to crashes.
The SymbolicBindingType refers to the type value that will be
substituted in for the BindSymbolicName, but holds onto the EntityNameId
from the BindSymbolicName instead of (or in addition to, for now) the
instruction.
The EntityNameId will be used to look in the ScopeStack to find the
witnesses either from the BindSymbolicName instruction, or other
instructions that specify `impls` constraints against the EntityName.
This will allow us to have the `T` in `I(T)` resolve to a `.Self`
reference in the type so that we get type equality with the binding's
type: `T:! I(.Self)`.
After this change, we correctly increment the offset by the next switch
case type.
Before this change, we accidentally incremented the offset by
`functions()` size instead of `cpp_overload_sets()` size and vice versa.
Also sorted the switch cases according to the order of the enum, for
consistency. This might help prevent a future similar incident.
This fix prevents crashing in the newly introduced test
`multiple_too_few_args_calls`.
This also has the side effect of showing `null name` for
`cpp_overload_set_type` and `cpp_overload_set_value`, instead of having
an arbitrary name.
Examples that demonstrate the old name is arbitrary can easily be seen
in tests like `cpp_namespace.carbon` and `decayed_param.carbon`, but
careful review would show that all old names are arbitrary, though often
luckily almost make sense.
We might want to have a proper name for these, but it's beyond the scope
of this crash fixing change.
See #6156.
Part of #5915.
This also does a little restructuring in the same direction, following
#6124.
Leads want `Destroy` to work similarly now for all types. As a
consequence, there doesn't seem to be as much benefit to splitting off
aggregate destruction. In this PR, the `type.destroy` function can now
be expected to destroy anything that's destructible; that means it'll be
usable for the `final fn` once that support is available.
Similarly, this gets rid of the impls other than the single blanket
impl, now using `type.can_destroy`. Since they all need to use the same
function, there's no benefit to splitting approaches. Also, now it can
just be a `final impl` since there should be no need for people to
create specializations -- if this blanket impl applies, it means the
`final fn` is the same.
This also slips in `partial` support since there's no reason to have it
diverge anymore. Also `abstract`, which I'm not sure is broadly testable
since most cases it'd come up, the `abstract` keyword is explicitly
detected/rejected.
Note though that this doesn't make any really big changes. It's just
realigning on the leads decision. I'm going this way to try to reduce
name-related churn for other changes.
The general strategy here is to force use of a thunk when we want to use
default arguments, and have Clang generate uses of the default arguments
on its side of the thunk.
To support this, change the key type used in `clang_decls` from being
just a `Decl*` to being a pair of `Decl*` and number of parameters in
the case of function decls. Import distinct `SemIR::Function`s for each
number of parameters that's used, and corresponding distinct thunks.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
The main direction of this change is the edits to `destroy.carbon`
(matching in both prelude and min_prelude).
Previously there was a no-op blanket impl for `Destroy`, which hid all
missing implementations of `Destroy`. This does a few things:
- Sets up builtin aggregate destruction for struct and tuple types as
before, but also adds C++ class types and array types to the same
handling. (all as a TODO for actual implementation)
- Also maybe-unformed destruction, for now at least. (there's a chance I
may try a different approach on this, but the impl lookup wasn't working
as I'd hope in order to write it in code)
- Adds handlers for simple things that are easy to do in code: `type`,
`bool`, pointers. (because these are no-op destruction)
- Redirect `const T` destruction to `T` destruction.
This leaves as future issues:
- `partial T` destruction. (this can't be done similar to `const`
because it only works for non-`final` class types; I think `class`
definitions should just generate what's needed)
- Destruction of other prelude-provided types. (will probably come up as
we implement class destruction, that the adapted builtin type doesn't
implement `Destroy` -- but may end up special-casing that in a way that
moots it)
This moves the `&` operator from `facet_types.carbon` to
`convert.carbon` because more things need to handle type and now that
we're getting separate copy and destroy interfaces. It should be
low-cost (an interface and builtin) so hopefully this is the right
balance for complexity and re-use.
A few tests are also edited in order to focus them more on what they
intend to test, and avoid a `Destroy` dependency.
Mark C++ functions as used when overload resolution selects them, and
trigger Clang's end-of-TU processing at the end of the Carbon
compilation to perform instantiation and other pending cleanup steps.
This is in support of a goal of changing the blanket `destroy` impl to
use (roughly):
```
private fn CanAggregateDestroy() -> type = "type.can_aggregate_destroy";
// Handles aggregate type destruction.
impl forall [AggregateDestroyT:! CanAggregateDestroy()] AggregateDestroyT as Destroy {
fn Op[addr self: Self*]() = "type.aggregate_destroy";
}
```
That isn't done here because there's still other issues that migrating
raises. What this *does* do is add the builtin functions, and in
particular, support to `FacetTypeInfo` to make `CanAggregateDestroy`
work.
The "special requirement" approach in `FacetTypeInfo` allows us to
support restricting a blanket impl under the current approach of impls.
Maybe we'll find a cleaner approach that can work in the future, but
this fits into the current model by propagating similar to other
requirements. I'm using an enum mask because we have a number of similar
things to add (e.g. copy, move) but I'm not sure we need a full vector.
A few alternatives considered were:
- Supporting syntax more like `where .Self impls
TypeCanAggregateDestroy(.Self, SupportedInterface,
UnsupportedInterface)`. I think it'd be a little cleaner, but requires
better compile-time evaluation in order to assess the type of the call.
Right now it's expected to be a `FacetType` too early to make this work,
and I was concerned about pouring too much more time down this route.
- Providing an actual interface, in particular doing name lookup back
into `Core.` for an interface. This would've added name lookup overhead,
and the question of whether an `impl` exists.
- Generating an interface. This avoids the name lookup, but would still
raise the question of whether an `impl` should also be generated. Work
I've previously done generating interfaces for class destruction also
feels complex to both write and understand (an unfortunate issue).
- Still modeling as an `ImplsConstraint`, for example by defining a
special `InterfaceId::CanAggregateDestroy = -2` similar to what we do on
other ids. I was hesitant because of how this expands the number of
modes of `InterfaceId`, and things for consuming code to watch out for,
for what feels like a relatively niche set of use-cases that are only
interface-like.
---------
Co-authored-by: Dana Jansens <danakj@orodu.net>
As proposed in [Carbon: C++ interop for overloaded functions and
function
templates](https://docs.google.com/document/d/1KUxumZtNe3mY3TsjW2s_ZADOlAaFlrtsLKHVILtqIaM/edit?tab=t.0),
Clang is used to perform the overload resolution using C++ rules, when
an overloaded C++ set is called from Carbon. Once a function is
selected, it's converted into a Carbon function and called using the
Carbon rules including argument conversions.
A single non-templated function is treated the same way as an overload
set and the same rules apply for its call.
Template functions are not supported yet.
Demo:
a) Non-templated function calls:
```c++
// --- overloads.h
auto foo(int a, short b) -> void;
auto foo(double a) -> void;
auto foo(int a) -> void;
```
```c++
// overloads.cpp
#include "overloads.h"
#include <cstdio>
auto foo(int a, short b) -> void {
printf("hello from foo_int_short(%d, %d) \n", a, b);
}
auto foo(double a) -> void { printf("hello from foo_double(%f) \n", a); }
auto foo(int a) -> void { printf("hello from foo_int(%d) \n", a); }
```
```c++
library "Main";
import Cpp library "overloads.h";
fn Run() -> i32 {
Cpp.foo(1.1 as f64);
return 0;
}
```
```
$ clang -c overloads.cpp
$ bazel-bin/toolchain/carbon compile main.carbon
$ bazel-bin/toolchain/carbon link overloads.o main.o --output=demo
$ ./demo
hello from foo_double(1.100000)
```
b) Constructors:
```c++
// --- constructor_overloads.h
class C {
public:
C();
C(int a, int b);
};
```
```c++
// constructor_overloads.cpp
#include "constructor_overloads.h"
#include <cstdio>
C::C() { printf("hello from C() \n"); }
C::C(int a, int b) { printf("hello from C(%d, %d) \n", a, b); }
```
```c++
library "Main";
import Cpp library "constructor_overloads.h";
fn Run() -> i32 {
let c1: Cpp.C = Cpp.C.C();
let c2: Cpp.C = Cpp.C.C(1, 2);
return 0;
}
```
```
$ clang -c constructor_overloads.cpp
$ bazel-bin/toolchain/carbon compile main.carbon
$ bazel-bin/toolchain/carbon link constructor_overloads.o main.o \--output=demo
$ ./demo
hello from C()
hello from C(1, 2)
```
Follow-ups:
- `Cpp.foo({})` - proper handling of struct literals as call args.
- Fix access for overloaded sets.
- Fix tests:
- Method calls: `error: missing object argument in method call
[MissingObjectInMethodCall]` in tests.
- Fix `toolchain/check/testdata/interop/cpp/import.carbon` test.
- Fix `enums` support.
- Fix `str` -> `std::string_view` mapping.
Part of #5915
This uses each vector's size as a barrier between lists, to eliminate
the possibility of incidental collisions between entries of different
lists. This is the same as is done inside `AddBlock`.
Attach the cleanup to the `Temporary` instruction instead of to the
`TemporaryStorage` instruction. We create `TemporaryStorage`
instructions speculatively when creating an initializing expression, and
may overwrite those instructions with other instructions if it turns out
that a temporary is not required. Instead, wait until we finalize the
temporary and create a `Temporary` instruction to register the cleanup.
We import C++ enum types as Carbon class types as adapters for the
corresponding builtin integer type, and we import enumerator constants
as integer constants of that class type.
No operators are supported on such values for now; eventually once we
start asking Clang to implement operators on C++-owned types, these
types should be handled in the same way. However, they can be converted
to the corresponding integer type with `as` via adapter conversion, and
integer builtin functions can operate on them.
Chandler CarruthandChristopher Di Bella
Richard Smith
Geoff Romer
Dana Jansens
Boaz Brickner
Ivana Ivanovska