Name conflicts weren't previously tested, and the diagnostics were just
a TODO, so this is also adding testing for that. But handling too.
Removing AddEntry because I think it's hard to make helpful for this
use-case when we want to do a diagnostic followup (because really,
callers want the full `.insert` result of pointer + success), and unused
otherwise.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
Remove the type canonicalization mechanism and instead rely on constant
canonicalization to deduplicate types.
Rename the `Canonicalize*Type` functions to reflect that they're no
longer performing canonicalization. Switch code that creates types due
to semantic checking, rather than due to source syntax, to directly
create type constants through evaluation rather than creating an
instruction and evaluating it to produce a separate constant
representation.
The mapping from `const (const T)` that was previously performed by type
canonicalization is now implemented in expression evaluation instead.
The value `<error>` is now treated as a constant value, with a special
property that an instruction involving `<error>` that could possibly be
constant evaluates to `<error>`. This helps avoid producing follow-on
errors when an error occurs as a subexpression of an expression, such as
a type, that is intended to be constant.
Rather than producing multiple constants with the same value, fold all
instances of a given constant to the same constant instruction.
A future PR will use this to replace the current type canonicalization
system.
Instructions created by splices during conversion are now evaluated, as
are instructions created in cases where we first create a placeholder
instruction and later replace it by a different instruction.
This also removes the ability to set a parse node and instruction
independently after creating an `InstId`, which could lead to them
accidentally not matching.
The parse nodes are still tracked as part of the same value store
interface in order to ensure parity, but they're split out from Inst
itself in order to reduce the size of Inst -- the expectation is that
they don't need to be passed around quite as much.
This change doesn't actually reduce the passing very much, although
there are hints of it: AddInstAndPush doesn't typically need a separate
parse node from the one on the Inst itself, for example. In a couple
spots I changed code to rely a little more on the InstId until the
ParseNode is needed, but it's very low hanging fruit where done. I think
convert could do more to not eagerly fetch the parse node before its
use, but more cleanup felt it would be easier to handle separately. I'm
currently viewing this as making such cleanup _possible_ rather than
executing on it up-front.
But also, I want to make sure there's a consensus to head in this
direction before pulling the trigger. We speculated that this would
result in the parse node being passed around less, and I do think that's
the case, although it's a bit fuzzy in the change.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
This is accomplished by tracking an extra bit on the ID we store in the
constant values table, and propagating that from subexpressions to the
enclosing expression. This extra bit is not yet computed correctly for
types; that will be addressed in later PRs.
---------
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
Form a side table with constant values for each instruction. Evaluation
is only supported for a few very simple kinds of instruction for now.
This is not observable outside of the SemIR output, because nothing
depends on expressions having a constant value phase yet.
This is primarily being done for performance reasons, removing hash
lookups. The increased memory consumption is accepted.
Refactors LexicalLookup out to its own structure.
Namespaces are copied, which means also adding their name to the
underlying instruction. It happened not to be done previously; the name
was only in name lookup.
Since the only import supported right now is the default import,
functionality is limited; in the future I'll need to deal with namespace
vs package conflicts.
Tests of namespace imports are under "namespace" -- I figured this would
be best for scaling as more instructions get support.
This also improves some debugging-related output that I was trying to
use while trying to build the support.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
This patches bazel_clang_tidy handling of headers. I found an equivalent
change at https://github.com/erenon/bazel_clang_tidy/pull/13, but that
was [already
rejected](https://github.com/erenon/bazel_clang_tidy/pull/13#issuecomment-1047007424).
Per the criticism, this will result in redundant processing of headers.
The project instead uses `HeaderFilterRegex: ".*"`, but that results in
two problems:
1. When running with `-k`, errors are repeated when a header is included
more than once, which is common.
2. clang-tidy including errors from headers that are included from other
modules (e.g., abseil-cpp); filtering correctly is difficult.
Given the trade-offs and options (including forking), I thought patching
was preferable so long as it remains narrow.
This change adds a `BindSymbolicName` instruction for generic bindings,
paralleling the existing `BindName`. A mechanism is also added to allow
both kinds of binding to be accessed uniformly, for convenience in the
case where the two different kinds of binding are treated the same.
Generic bindings of type `type` are allowed to be used as types,
although no operations are provided for such types. For now lowering
treats these types as empty structs, which seems like a reasonable
lowering for non-monomorphized unconstrained types.
This is a step towards adding enclosing scopes for imports. It creates
an indirection for all bind names.
We discussed specializing for bindings that are in function scope (i.e.,
not a useful enclosing scope for imports or diagnostics). However, the
thought is to go ahead with this singular approach for now, and only
change structure if it's a performance issues so that we have
incrementally fewer instructions to handle.
Add a mechanism to define instruction categories, to support inspecting
the common representation of similar kinds of instruction. Use that
mechanism to make formatting of branch instructions slightly more
type-safe.
The idea here is to use the existing `Inst` mechanism for converting to
and from structs, extended to operate on a struct representing multiple
different kinds of instruction. In this case, the concrete kind of
instruction is stored in the struct in a `kind` field, rather than being
implied by the type.
Factored out of #3555 where this mechanism is used to provide a common
interface for runtime and symbolic name bindings.
This involves a number of supporting changes:
* The `parse_node;` member of instruction types may now have any type
derived from `Parse::NodeId` and is no longer required to have that
exact type.
* `Parse::Node::Invalid` is now a singleton object of a separate type
that is convertible to `Parse::NodeId` and its descendants. This
replaces the `Invalid` member of its descendants, and avoids having to
write long `NodeIdOneOf<...>` types when initializing variables to
invalid.
* `IndexBase` now allows `==` and `!=` comparisons between its derived
classes and types that are convertible to those types.
* A number of functions in the check stage have been changed to preserve
more type information instead of using `Parse::NodeId`.
* `NodeIdForKind<K>` (also known as `KId`) now has a `Kind` member so it
may be used to declare `NodeIdOneOf<T, U>` types without #including
`parse/typed_nodes.h`.
* `NodeIdForKind<K>` (also known as `KId`) may be implicitly converted
to `NodeIdOneOf<T, U>` if `T::Kind == K` or `U::Kind == K` (executing a
TODO).
Many of the `parse_node` members were not converted since they would
have required more extensive changes. They have been marked with "TODO"
comments.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
When formatting special values, we currently have `NameId` doing `if
(*this == SelfValue)`, `BoolValue` doing `case 0:`, and `TypeId` doing
`if (index == TypeType.index)`. I'm suggesting we consolidate onto the
`*this == SelfValue` approach for consistency, it seems the easiest to
see the mapping of values.
This mixes in #3552 because I'm adding `Exports` there. I'm suggesting
reformatting `Empty` and `Exports` consistent with `Unreachable`.
This adds a block for exported InstIds, rather than scanning the package
scope. This working down a path discussed last month, which we'll need
to add enclosing namespaces to the Inst in order to complete import of
something like `namespace NS; var NS.a;`
Exports could've been a separate `vector<InstId>` on `SemIR::File`, but
using an entry in `inst_blocks` felt more consistent.
Goal is to increase type safety, though more work needs to be done (see
added TODOs).
Note that, after this change, check handlers corresponding to deleted
parse node kinds will no longer compile.
---------
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
These are intended to allow the structure of a parse tree node to be
described more precisely in code, to support these use cases:
- Automated checking that the parse tree conforms to the expected
structure. (Added to `Tree::Verify`.)
- Easier reading and understanding of the structure of the parse tree by
toolchain developers. (See `parse/typed_nodes.h`.)
- Easier navigation of the parse tree, for example for tooling uses and
for use when forming diagnostics.
On this last point, an object representing the file may be inspecting
using `Tree::ExtractFile`, as in:
```
auto file = tree->ExtractFile();
for (AnyDeclId decl_id : file.decls) {
// `decl_id` is convertible to a `NodeId`.
if (std::optional<FunctionDecl> fn_decl =
tree->ExtractAs<FunctionDecl>(decl_id)) {
// fn_decl->params is a `TuplePatternId` (which extends `NodeId`)
// that is guaranteed to reference a `TuplePattern`.
std::optional<TuplePattern> params = tree->Extract(fn_decl->params);
// `params` has a value unless there was an error in that node.
} else if (auto class_def = tree->ExtractAs<ClassDefinition>(decl_id)) {
// ...
}
}
```
The `Extract...` functions collect the child nodes into the typed parse
node's fields (internally using a `Tree::SiblingIterator`) for easy
access. However, this is not as fast as directly observing the tree
structure using the postorder strategy being used by the check stage.
These functions rely on using struct reflection on the typed parse node
definitions from `parse/typed_nodes.h` to get the expected structure of
child nodes and then populate them.
Note that validating these in `Tree::Verify` adds significant cost to
it, and is currently included in the parsing stage. Without this change,
a 10 mloc test case of lex & parse takes 4.129 s ± 0.041 s. With this
change, it takes 5.768 s ± 0.036 s.
This builds upon and completes #3393.
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
The goal here is to make the representation more uniform so that we can
start adding different kinds of binding -- checked generic bindings and
template bindings -- across both function parameters and local `let`
declarations.
With this change, the entry in the parameter list for the function is
the name binding, not the Param itself, which has some ripple effects on
consumers of that list that want to access the parameter rather than the
binding. This is expected to change again when we start adding more of
the pattern matching SemIR, but this seems good enough for now.
By adding an InstId to the NameScope, we can determine whether the
declaration is being added to a scoped entity (versus a namespace).
The choice of InstId on NameScope is chosen versus other solutions
because, for imports, we want to just have a list of InstIds to import
and, from those, get the containing namespaces for addition. Similar may
also be desirable for printing fully qualified names given a singular
InstId. That means an InstId must have a path to find enclosing name
scopes.
What we're looking at here is:
- NameScopeId knows its InstId. (done here)
- Inst knows the enclosing NameScopeId. (future work)
- To walk up enclosing scopes for an Inst:
1. Fetch the Inst.
2. Find its enclosing NameScopeId (which will be per-declaration due to
Function etc complexity).
3. Fetch the NameScope if not Package scope. (if Package scope, done)
4. Use the InstId on the NameScope to go back to step 1.
We have `StringLiteral`s in multiple other `Carbon` sub-namespaces.
Rename to a more specific name to avoid collisions.
We should likely also rename `Carbon::IntId` -> `Carbon::IntValueId` and
`Carbon::RealId` -> `Carbon::RealValueId`, but this collision is
prioritized because it was blocking work on typed parse nodes which
introduces a `Carbon::Parse::StringLiteralId`.
This builds out a little infrastructure for one name scope to `extend`
another. We'll need more refinement here to cover other cases, but this
should provide some foundation for that future work.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
Adds a `LazyImportRef` instruction. Versus `CrossRef`, this is intended
to represent an instruction which cannot be used directly, and must be
replaced when it comes up due to name lookup. The intent is to use this
to avoid recursive loading of imported IR instructions.
Note, under this model, when `ResolveIfLazyImportRef` is called, it
essentially needs to load both inst and type information to a sufficient
point where any further attempts would hit name lookup again. That will
probably be complex, and the current implementation is just touching the
surface of the issue. I was heading down this route because it would
mean we have a limited number of points that need to consider whether
they're going to talk about a `LazyImportRef`.
I'm considering whether `CrossRef` should be dropped in favor of more
specific `Builtin` special-casing, due to the divergence of desired
behaviors. This could mean dropping the `builtins` IR since it's not
looking useful right now.
Modify `NameScope` to track whether the scope is associated with a load
error. This is to handle cases where one or more imports failed, so we
do not want to issue warnings for related scopes.
The 0-size on `ValueStore` comes up due to the changes to `NameScope`,
which make it too large for the default handling. After discussion with
zygoloid, the thought was we might want to try reserving a roughly
correct value based on parse node counts, but the stack default wasn't
buying much.
Fixes a bug where the implicit import used the package name instead of
the invalid identifier.
This is also doing the parse node split, allowing lower reliance in
formatter on the tokenized buffer (something that I may be touching more
due to import handling).
BinaryOperatorAdd had been added early on as a proof-of-concept for
toolchain design for parsing -> checking flow. However, it doesn't
reflect the interface direction for operators, and now other portions of
the toolchain demonstrate the relevant logic. Instead, it's just a bit
of an outlier versus other instructions which have evolved over time.
Most of the calls to `StringifyType` already passed `true` for
`in_type_context`. Checking the rest, I found that every one of them was
already sufficiently clear that they were printing a type, or could be
made so with a very small change to the diagnostic text.
This creates a namespace for `package` scope.
It looks like names of class_decls in namespaces lead to an unexpected
instref. This is already true, as best as I can tell. I'm not sure if
there's a preferred approach to address that, so I've left a TODO for
now.
There's a trade-off here of explicitness in the use versus repetition,
but I'm hoping the Id offers sufficient info (also, some Ids already
relied on this, so this builds consistency). The forward declarations
I'm mixed on, but they are difficult to avoid if heading down this route
due to interdependencies between ids and types which contain ids.
This removes the filename from the file-scoped block, and places it
above to make it clear where the full SemIR begins (with multifile,
providing a barrier between).
This reflects how we're naming classes that derive from these classes,
and matches usage for each existing `Id` and `Index` type, except:
- `Parse::NodeId` previously inherited from `ComparableIndexBase`, and
is no longer comparable.
- `SemIR::MemberIndex` previously inherited from `IndexBase`, and is now
comparable.
Making `Parse::NodeId` non-comparable reflects that it's intended to be
an opaque identifier for a node and that the ordering is an
implementation detail rather than part of the intended public interface.
`PostorderIterator` and `SiblingIterator` still rely on the numerical
meaning of `NodeId`s, but that's OK since they're part of the node
implementation.
I was suggesting this because `FloatingPoint` is pretty long. `int` and
`float` should be familiar abbreviations. `unsigned` should be familiar
to developers too, but `UnsignedInt` still feels usefully clearer for
the additional chars.
These are manual fixes; mostly from clang-tidy, some from clangd (which
notes unused includes).
In typed_insts, adding inlline due to misc-definitions-in-headers. Per
discussion, clang-tidy is wrong, but inline silences it.
For parameter name skew in definition versus declaration, I'm just using
the name from the definition.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
This adds instructions so that we get printing. I may adjust the
instruction format a little further to add a type, but I think the basic
setup will remain.
Note this builds on #3414
Per [#toolchain
discussion](https://discord.com/channels/655572317891461132/655578254970716160/1176632520834560211)
We'd at one point been trying to put `[[nodiscard]]` everywhere, but
then we stopped because it had felt verbose without finding many issues
(plus, people plain forgot to add it). Some history in #888.
Since newer code gets added without it, we now have code like:
```
auto GetLineInfo(Line line) -> LineInfo&;
[[nodiscard]] auto GetLineInfo(Line line) const -> const LineInfo&;
auto AddLine(LineInfo info) -> Line;
auto GetTokenInfo(Token token) -> TokenInfo&;
[[nodiscard]] auto GetTokenInfo(Token token) const -> const TokenInfo&;
auto AddToken(TokenInfo info) -> Token;
[[nodiscard]] auto GetTokenPrintWidths(Token token) const -> PrintWidths;
```
Here, the lack of `[[nodiscard]]` doesn't mean anything: for example,
`GetLineInfo` should not have its result discarded if it's called. But
the mix could be confusing for readers.
As a resolution, remove the attribute. `[[nodiscard]]` should be treated
like other attributes going forward, which essentially means "avoid in
general, add a comment to explain why the attribute is needed" rather
than use-as-default.