This doesn't split apart the current error type into one that tracks
location and one that doesn't, although that might be easier to do once
we have this.
Instead, this is primarily intended to support custom error types that
lazily materialize the error message in case that can be avoided by
completely handling the error. For example, many file system operations
are *expected* to produce errors even in the hot path and we don't want
to render `ENOENT` (for example) to a pretty string and instead will
directly query the error to understand and handle it in code.
The type parameter ordering isn't the most obvious, but helpfully allows
us to default the error type in a useful way.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
This is to make it easier to debug formatter issues. It means printing
can now result in things like:
```
<unexpected>.inst57.loc4_24: type = bind_symbolic_name ...
```
Where the "unexpected" reflects incorrect construction.
Expand support for `int` and `short` to cover all the other `intN_t` and
`uintN_t` types too. We achieve this by asking Clang what the `intN_t` /
`uintN_t` type that it would use for the given bitwidth is, and checking
if that's the type we're trying to map.
This is trying to make it clearer when vectors are being indexed with
`CheckIRId`.
The only one that I still kind of want to change is the
`SmallVector<std::unique_ptr<CompilationUnit>>`, but because it's a
`unique_ptr` that's a little more complex. I may not bother.
Note, some of the changes around nuanced `SmallVector` interactions were
based on trying to copy the way `SmallVector` itself takes arguments,
like with range passing.
Change impls from `<interface>.impl` to `<self>.as.<interface>.impl`,
and *member* functions to `<parent scope>.<fn>` (non-member functions
exclude their parent scope). Stop special-casing builtin functions,
given the new naming scheme.
The purpose of this is to make it clearer when a member function is
being accessed and, if so, which member function. In particular, we
often access interface `Op` functions. The builtin function
special-casing was intended to help with that, but we still have lots of
`Op` functions. This particular approach should make the interactions
clearer.
This changes up queueing of block IDs a little because, in particular,
we need to process bodies of entities only after constants finish
processing. But, it should also result in less memory usage during
processing because it means we have less on the insts stack at any given
time, since we track a block rather than all instructions contained by
the block.
This switches from the `CollectNamesInBlock` approach for entities, to
instead traversing entities as they're encountered. For example, when
traversing constants, when a type is found, the entity will have its
block queued for processing.
This leads to a change in the traversal order, which affects
disambiguation done by numeric sequencing (since that's just showing the
traversal order).
This will allow for simpler "name based on name" logic. This is
something I plan to use for:
- impls: `<type>.as.<interface>.impl`
- functions: `<entity>.<member function>`
- Note an impl may be used as the entity for a bound function.
By naming the entities as they're encountered, I'll be able to rely on
the generated names rather than recalculating them.
To assist this, I'm also differentiating between the ambiguous and
disambiguated name. Otherwise, we could end up with things like
`<function>.<disambiguator>.<call>.<other disambiguator>`, where the
repeated disambiguator may not be necessary in order to get full
disambiguation. It's also a smaller delta from the current output.
Note, changing `Name` to a class felt appropriate given its shape. I was
also noticing that parts of its API were unused, and the class helps
detect unused private members.
---------
Co-authored-by: Geoff Romer <gromer@google.com>
This will allow reusing existing entity names when there are nested
.Self references in a facet type. They are canonical as the contents of
a .Self reference are all canonical.
- Update bazel to 8.3.1, just to stay reasonably up to date.
- Bazel warned about the platforms version, so I generally updated
packages that have central registry versions.
- Note there's a newer re2 in the central registry, but I got a download
error with it.
- `--experimental_guard_against_concurrent_changes` is deprecated; I
wasn't sure it's worth explicitly setting
`--guard_against_concurrent_changes=full`, but figured it may be
consistent (it's not clear to me -- see
https://github.com/bazelbuild/bazel/pull/25874).
Also update `RequestVtableIfVirtual` to strip `impl` from functions in
classes without a base class - this avoids a duplicate diagnostic where
they're diagnosed as being in a class without a base class, then
diagnosed again because they don't have a matching function in a base
class.
Versus #5823, this is manually updated (still verified with `pre-commit
run -a`).
I also looked at updating prettier; adding a note why I'm not doing
that.
Use `full.carbon` min-prelude for any tests using the full prelude. A
few tests were using it unnecessarily and are changed to a more minimal
one in the process.
Any test which does not include some min-prelude will now fail with an
error.
Before this change we crash in this case when trying to use the outer
type.
After this change we diagnose a TODO.
Will add the missing support for this in a separate PR.
Part of #5533.
This changes ConstantValueStore to use ValueStore so that we get the
allocation flow that we're leaning towards there. This adds
`ConstantId::SymbolicId` to represent what was previously an int32
"symbolic_index" (although called an index, it's consistent with ID
usages).
I'm also changing Chunk::at/push to be consistent with the wrapping
Get/Add naming; the naming difference sticks out more with
UninitializedFill being added.
For example, this would allow using `std::string::size`, which is
templated because it's a member of the class template
`std::basic_string`, but isn't itself a function template.
This is instead of having one test file that has different return types
and two test files for `i16` and `i32` param tests, which doesn't seem
consistent.
First commit does file renames for easier review.
Part of #5063.
Added test coverage to demonstrate name lookup following import of a
type of a function parameter.
SemIR changes show that the same decl isn't imported multiple times.
Part of #5533.
Create a `self` parameter when importing a C++ non-static member
function. That seems to be all we need to get basic method calls
working! For now, `const` methods have by-value self parameters, and
non-`const` methods get an `addr self: Self*` parameter.
Incidentally also supports import of pointers-to-pointers.
Imported const-qualified types aren't especially useful just yet,
because on the Carbon side we don't yet permit conversions from
non-const to const types, so most of the tests still fail, but for
different reasons now.
Some specific features:
* Use `SpecificFunction` for vtable entries for generic classes.
* Create specific constants for vtable entries in classes derived from
generic classes to reference the appropriate specific of the function
in the context of such a derived class.
* Create specific constants for vtable_ptrs for uses of specific generic
classes.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
There are some very useful default arguments, so teach the runner script
to directly provide them. They can be easily overridden if needed. As
part of this, change the default run count to 10 which much more often
produces statistically significant error bars and results.
Also, tweak the processing of the results to provide a stable order
based on the source order, even when randomized interleaving is enabled.
The randomized interleaving improves the statistical strength of the
benchmarks significantly, but displaying the results in the source order
is much more understandable. This should give roughly the best of both
worlds.
Direct use of an error means we can delay looking at the actual type,
and also use the `self_id` of the stored `ImplInfo` directly.
Note, I'm assuming we should be okay with this not always being a
`NameRef`. In error cases though, it seems like we shouldn't mind it
being an error instead of a name reference to an error.
In `BuildUnaryOperator`, `GetOperatorOpFunction` is treated as
desugaring, but `PerformCompoundMemberAccess` and `PerformCall` are not.
This treats all of destruction as desugaring.
This leads to some instructions being elided, because of `GetOrAddInst`
behaviors:
> // If the instruction has a desugared location and a constant value,
returns
> // the constant value's instruction ID. Otherwise, same as AddInst.
This changes instructions that previously had a non-desugared location
to instead have a desugared location, so if they also have a constant
value then the constant value can be used directly.
Unify code paths for importing classes by name and importing them
indirectly when their type is referenced. Switch to using the general
type import machinery to import all type declarations, which allows
typedef declarations naming importable types to be used too.
Fix up handling of error cases to consistently only produce an Error
InstId after actually producing an error message, so that we can produce
exactly one diagnostic in failure cases.
Remove TODO error for unions that previously was only produced when
importing them indirectly, not when importing them by name. Import of
unions is exactly as complete / incomplete as import of other class
types, so treating them differently doesn't seem necessary.
Chandler CarruthandJon Ross-Perkins
Richard Smith
dependabot[bot]
Dana Jansens
Boaz Brickner
David Blaikie