mirror of
https://github.com/carbon-language/carbon-lang.git
synced 2026-11-14 14:08:50 +00:00
Lexer diagnostics used a single `const char*` as their source location, which only allowed pointing at a single character rather than a run of characters. This limited the underlining of unrecognized tokens, and also made escape sequence diagnostics particularly strange by stopping at whatever point the lexer discovered the error and pointing only at that location. For example: `"\w"` was reported at the `w`, and `"\x"` at the closing `"`. This change tries to always keep the primary location of a diagnostic the point at which typing (inserting or deleting) might fix the error. So for `"\x"`, `\x` is underlined, but the position is at the `"` as that's where you would type a hex code. The lexer's emitter now takes a special lexing source location that represents either a single position _between_ two characters, a single character, or a range of the source buffer. The lexer diagnostics are updated to construct these for the character range as appropriate. Parser diagnostics for a missing token were located at the _next_ token, which is particularly confusing for terminators like missing `;`s. Now they are located at a position after the previous token, with or without a space depending on the token, but typically on the same line. Diagnostics still sort by how far the parser has read. The parser diagnostics for unclosed groups also now attaches a label to the opening token when it is on a separate line. And the diagnostics for mismatched indent in a multi-line string attaches a label to the closing `'''` that establishes that indent. There are likely more location improvements we should do here, but this seemed like a reasonable incremental improvement and is what Claude found when doing an initial audit. One example where I think we should make further improvements is around suggesting raw identifiers when we find an unexpected `Cpp` or `Core` or `base`. We do this consistently now, but raw identifiers should be rare. A TODO is added for these cases. Assisted-by: Claude Code --------- Co-authored-by: Richard Smith <richard@metafoo.co.uk>
147 lines
5.5 KiB
C++
147 lines
5.5 KiB
C++
// Part of the Carbon Language project, under the Apache License v2.0 with LLVM
|
|
// Exceptions. See /LICENSE for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
|
|
#include "toolchain/diagnostics/diagnostic.h"
|
|
#include "toolchain/diagnostics/format_providers.h"
|
|
#include "toolchain/lex/token_kind.h"
|
|
#include "toolchain/parse/context.h"
|
|
#include "toolchain/parse/handle.h"
|
|
#include "toolchain/parse/node_kind.h"
|
|
|
|
namespace Carbon::Parse {
|
|
|
|
auto HandleLet(Context& context) -> void {
|
|
auto state = context.PopState();
|
|
|
|
// These will start at the `let`.
|
|
context.PushState(state, StateKind::LetFinishAsRegular);
|
|
context.PushState(state, StateKind::LetAfterPatternAsRegular);
|
|
|
|
context.PushStateForPattern(StateKind::Pattern, /*in_var_pattern=*/false,
|
|
/*in_unused_pattern=*/false,
|
|
/*in_field_shorthand_pattern=*/false,
|
|
BindingContext::ExplicitParam,
|
|
PrecedenceGroup::ForTopLevelPattern());
|
|
}
|
|
|
|
auto HandleAssociatedConstant(Context& context) -> void {
|
|
auto state = context.PopState();
|
|
|
|
// Parse the associated constant pattern: identifier : type. The binding is a
|
|
// checked generic by context, so no phase keyword is used.
|
|
auto identifier = context.ConsumeIf(Lex::TokenKind::Identifier);
|
|
if (!identifier) {
|
|
auto kind = context.PositionKind();
|
|
if (kind == Lex::TokenKind::Generic || kind == Lex::TokenKind::Runtime ||
|
|
kind == Lex::TokenKind::Template) {
|
|
// Diagnose a phase keyword specifically, rather than reporting a missing
|
|
// associated constant name.
|
|
CARBON_DIAGNOSTIC(
|
|
PhaseKeywordInAssociatedConstant, Error,
|
|
"phase keyword is not allowed on an associated constant");
|
|
context.emitter().Emit(*context.position(),
|
|
PhaseKeywordInAssociatedConstant);
|
|
state.has_error = true;
|
|
} else {
|
|
// A word that can't be a name, followed by the `:`, was most likely
|
|
// meant as the name, and is what the message is about. Otherwise the
|
|
// name is missing, and goes after `let`.
|
|
bool recover_as_raw =
|
|
kind.is_word() &&
|
|
context.PositionKind(Lookahead::NextToken) == Lex::TokenKind::Colon;
|
|
CARBON_DIAGNOSTIC(
|
|
ExpectedAssociatedConstantIdentifier, Error,
|
|
"expected identifier in associated constant declaration"
|
|
"{0:; prefix reserved word with `r#` to form a valid identifier|}",
|
|
Diagnostics::BoolAsSelect);
|
|
context.emitter().Emit(recover_as_raw ? DiagnosticLoc(*context.position())
|
|
: context.PointAfterPreviousToken(),
|
|
ExpectedAssociatedConstantIdentifier,
|
|
recover_as_raw);
|
|
if (recover_as_raw) {
|
|
identifier = context.tokens().AddPostLexingRecoveryTokenAsIdentifier(
|
|
context.Consume());
|
|
} else {
|
|
state.has_error = true;
|
|
}
|
|
}
|
|
}
|
|
|
|
auto colon = context.ConsumeIf(Lex::TokenKind::Colon);
|
|
if (identifier && !colon) {
|
|
CARBON_DIAGNOSTIC(ExpectedAssociatedConstantColon, Error,
|
|
"expected `:` in associated constant declaration");
|
|
context.emitter().Emit(context.PointAfterPreviousToken(),
|
|
ExpectedAssociatedConstantColon);
|
|
state.has_error = true;
|
|
}
|
|
|
|
if (!identifier || !colon) {
|
|
auto end_token = context.SkipPastLikelyEnd(*(context.position() - 1));
|
|
context.AddNode(NodeKind::AssociatedConstantDecl, end_token,
|
|
/*has_error=*/true);
|
|
state.has_error = true;
|
|
return;
|
|
}
|
|
|
|
context.AddLeafNode(NodeKind::IdentifierNameNotBeforeSignature, *identifier);
|
|
state.token = *colon;
|
|
context.PushState(state, StateKind::LetFinishAsAssociatedConstant);
|
|
context.PushState(state, StateKind::LetAfterPatternAsAssociatedConstant);
|
|
context.PushState(StateKind::Expr);
|
|
}
|
|
|
|
static auto HandleLetAfterPattern(Context& context, NodeKind init_kind)
|
|
-> void {
|
|
auto state = context.PopState();
|
|
|
|
if (state.has_error) {
|
|
if (auto after_pattern =
|
|
context.FindNextOf({Lex::TokenKind::Equal, Lex::TokenKind::Semi})) {
|
|
context.SkipTo(*after_pattern);
|
|
}
|
|
}
|
|
|
|
if (auto equals = context.ConsumeIf(Lex::TokenKind::Equal)) {
|
|
context.AddLeafNode(init_kind, *equals);
|
|
context.PushState(StateKind::Expr);
|
|
}
|
|
}
|
|
|
|
auto HandleLetAfterPatternAsRegular(Context& context) -> void {
|
|
HandleLetAfterPattern(context, NodeKind::LetInitializer);
|
|
}
|
|
|
|
auto HandleLetAfterPatternAsAssociatedConstant(Context& context) -> void {
|
|
auto state = context.PopState();
|
|
context.AddNode(NodeKind::AssociatedConstantNameAndType, state.token,
|
|
state.has_error);
|
|
context.PushState(state);
|
|
HandleLetAfterPattern(context, NodeKind::AssociatedConstantInitializer);
|
|
}
|
|
|
|
static auto HandleLetFinish(Context& context, NodeKind node_kind) -> void {
|
|
auto state = context.PopState();
|
|
|
|
auto end_token = state.token;
|
|
if (context.PositionIs(Lex::TokenKind::Semi)) {
|
|
end_token = context.Consume();
|
|
} else {
|
|
context.DiagnoseExpectedDeclSemi(Lex::TokenKind::Let);
|
|
state.has_error = true;
|
|
end_token = context.SkipPastLikelyEnd(state.token);
|
|
}
|
|
context.AddNode(node_kind, end_token, state.has_error);
|
|
}
|
|
|
|
auto HandleLetFinishAsRegular(Context& context) -> void {
|
|
HandleLetFinish(context, NodeKind::LetDecl);
|
|
}
|
|
|
|
auto HandleLetFinishAsAssociatedConstant(Context& context) -> void {
|
|
HandleLetFinish(context, NodeKind::AssociatedConstantDecl);
|
|
}
|
|
|
|
} // namespace Carbon::Parse
|