[lldb] [llvm] [LLDB] Add `ScalarLiteralNode` and literal parsing in DIL (PR #152308)
Ilia Kuklin via llvm-commits
llvm-commits at lists.llvm.org
Wed Aug 6 06:51:29 PDT 2025
https://github.com/kuilpd created https://github.com/llvm/llvm-project/pull/152308
This patch introduces `ScalarLiteralNode` without any uses by other nodes yet. It also includes lexing and parsing for integer and floating point numbers, and makes a function `getAutoSenseRadix` in `StringRef` global.
>From 330a34ed85f80563b9d6f204b59e1f3172c9f09a Mon Sep 17 00:00:00 2001
From: Ilia Kuklin <ikuklin at accesssoftek.com>
Date: Fri, 1 Aug 2025 20:14:59 +0500
Subject: [PATCH 1/3] Add ScalarLiteralNode
---
lldb/include/lldb/ValueObject/DILAST.h | 23 +++++++++
lldb/include/lldb/ValueObject/DILEval.h | 2 +
lldb/include/lldb/ValueObject/DILParser.h | 2 +
lldb/source/ValueObject/DILAST.cpp | 5 ++
lldb/source/ValueObject/DILEval.cpp | 28 +++++++++++
lldb/source/ValueObject/DILParser.cpp | 48 +++++++++++++++++++
.../TestFrameVarDILArraySubscript.py | 2 +-
.../Indirection/TestFrameVarDILIndirection.py | 2 +-
8 files changed, 110 insertions(+), 2 deletions(-)
diff --git a/lldb/include/lldb/ValueObject/DILAST.h b/lldb/include/lldb/ValueObject/DILAST.h
index 709f0639135f1..a174e28ea5c06 100644
--- a/lldb/include/lldb/ValueObject/DILAST.h
+++ b/lldb/include/lldb/ValueObject/DILAST.h
@@ -23,6 +23,7 @@ enum class NodeKind {
eErrorNode,
eIdentifierNode,
eMemberOfNode,
+ eScalarLiteralNode,
eUnaryOpNode,
};
@@ -178,6 +179,26 @@ class BitFieldExtractionNode : public ASTNode {
int64_t m_last_index;
};
+class ScalarLiteralNode : public ASTNode {
+public:
+ ScalarLiteralNode(uint32_t location, lldb::BasicType type, Scalar value)
+ : ASTNode(location, NodeKind::eScalarLiteralNode), m_type(type),
+ m_value(value) {}
+
+ llvm::Expected<lldb::ValueObjectSP> Accept(Visitor *v) const override;
+
+ lldb::BasicType GetType() const { return m_type; }
+ Scalar GetValue() const & { return m_value; }
+
+ static bool classof(const ASTNode *node) {
+ return node->GetKind() == NodeKind::eScalarLiteralNode;
+ }
+
+private:
+ lldb::BasicType m_type;
+ Scalar m_value;
+};
+
/// This class contains one Visit method for each specialized type of
/// DIL AST node. The Visit methods are used to dispatch a DIL AST node to
/// the correct function in the DIL expression evaluator for evaluating that
@@ -195,6 +216,8 @@ class Visitor {
Visit(const ArraySubscriptNode *node) = 0;
virtual llvm::Expected<lldb::ValueObjectSP>
Visit(const BitFieldExtractionNode *node) = 0;
+ virtual llvm::Expected<lldb::ValueObjectSP>
+ Visit(const ScalarLiteralNode *node) = 0;
};
} // namespace lldb_private::dil
diff --git a/lldb/include/lldb/ValueObject/DILEval.h b/lldb/include/lldb/ValueObject/DILEval.h
index 45e29b3ddcd7b..cb2a81d1c7ba1 100644
--- a/lldb/include/lldb/ValueObject/DILEval.h
+++ b/lldb/include/lldb/ValueObject/DILEval.h
@@ -54,6 +54,8 @@ class Interpreter : Visitor {
Visit(const ArraySubscriptNode *node) override;
llvm::Expected<lldb::ValueObjectSP>
Visit(const BitFieldExtractionNode *node) override;
+ llvm::Expected<lldb::ValueObjectSP>
+ Visit(const ScalarLiteralNode *node) override;
// Used by the interpreter to create objects, perform casts, etc.
lldb::TargetSP m_target;
diff --git a/lldb/include/lldb/ValueObject/DILParser.h b/lldb/include/lldb/ValueObject/DILParser.h
index 9eda7bac4a364..2cd8ca3be3c02 100644
--- a/lldb/include/lldb/ValueObject/DILParser.h
+++ b/lldb/include/lldb/ValueObject/DILParser.h
@@ -96,6 +96,8 @@ class DILParser {
std::string ParseIdExpression();
std::string ParseUnqualifiedId();
std::optional<int64_t> ParseIntegerConstant();
+ ASTNodeUP ParseNumericLiteral();
+ ASTNodeUP ParseNumericConstant();
void BailOut(const std::string &error, uint32_t loc, uint16_t err_len);
diff --git a/lldb/source/ValueObject/DILAST.cpp b/lldb/source/ValueObject/DILAST.cpp
index b1cd824c2299e..38215ae18f6ce 100644
--- a/lldb/source/ValueObject/DILAST.cpp
+++ b/lldb/source/ValueObject/DILAST.cpp
@@ -37,4 +37,9 @@ BitFieldExtractionNode::Accept(Visitor *v) const {
return v->Visit(this);
}
+llvm::Expected<lldb::ValueObjectSP>
+ScalarLiteralNode::Accept(Visitor *v) const {
+ return v->Visit(this);
+}
+
} // namespace lldb_private::dil
diff --git a/lldb/source/ValueObject/DILEval.cpp b/lldb/source/ValueObject/DILEval.cpp
index 6f28434c646cd..18cc30d589829 100644
--- a/lldb/source/ValueObject/DILEval.cpp
+++ b/lldb/source/ValueObject/DILEval.cpp
@@ -7,7 +7,9 @@
//===----------------------------------------------------------------------===//
#include "lldb/ValueObject/DILEval.h"
+#include "lldb/Core/Module.h"
#include "lldb/Symbol/CompileUnit.h"
+#include "lldb/Symbol/TypeSystem.h"
#include "lldb/Symbol/VariableList.h"
#include "lldb/Target/RegisterContext.h"
#include "lldb/ValueObject/DILAST.h"
@@ -402,4 +404,30 @@ Interpreter::Visit(const BitFieldExtractionNode *node) {
return child_valobj_sp;
}
+static CompilerType GetBasicTypeFromCU(std::shared_ptr<StackFrame> ctx,
+ lldb::BasicType basic_type) {
+ SymbolContext symbol_context =
+ ctx->GetSymbolContext(lldb::eSymbolContextCompUnit);
+ auto language = symbol_context.comp_unit->GetLanguage();
+
+ symbol_context = ctx->GetSymbolContext(lldb::eSymbolContextModule);
+ auto type_system =
+ symbol_context.module_sp->GetTypeSystemForLanguage(language);
+
+ if (type_system)
+ if (auto compiler_type = type_system.get()->GetBasicTypeFromAST(basic_type))
+ return compiler_type;
+
+ return CompilerType();
+}
+
+llvm::Expected<lldb::ValueObjectSP>
+Interpreter::Visit(const ScalarLiteralNode *node) {
+ CompilerType result_type =
+ GetBasicTypeFromCU(m_exe_ctx_scope, node->GetType());
+ Scalar value = node->GetValue();
+ return ValueObject::CreateValueObjectFromScalar(m_target, value, result_type,
+ "result");
+}
+
} // namespace lldb_private::dil
diff --git a/lldb/source/ValueObject/DILParser.cpp b/lldb/source/ValueObject/DILParser.cpp
index eac41fab90763..91b9d764527b3 100644
--- a/lldb/source/ValueObject/DILParser.cpp
+++ b/lldb/source/ValueObject/DILParser.cpp
@@ -183,6 +183,8 @@ ASTNodeUP DILParser::ParsePostfixExpression() {
// "(" expression ")"
//
ASTNodeUP DILParser::ParsePrimaryExpression() {
+ if (CurToken().Is(Token::numeric_constant))
+ return ParseNumericLiteral();
if (CurToken().IsOneOf(
{Token::coloncolon, Token::identifier, Token::l_paren})) {
// Save the source location for the diagnostics message.
@@ -370,6 +372,52 @@ std::optional<int64_t> DILParser::ParseIntegerConstant() {
return std::nullopt;
}
+// Parse a numeric_literal.
+//
+// numeric_literal:
+// ? Token::numeric_constant ?
+//
+ASTNodeUP DILParser::ParseNumericLiteral() {
+ Expect(Token::numeric_constant);
+ ASTNodeUP numeric_constant = ParseNumericConstant();
+ if (numeric_constant->GetKind() == NodeKind::eErrorNode) {
+ BailOut(llvm::formatv("Failed to parse token as numeric-constant: {0}",
+ CurToken()),
+ CurToken().GetLocation(), CurToken().GetSpelling().length());
+ return std::make_unique<ErrorNode>();
+ }
+ m_dil_lexer.Advance();
+ return numeric_constant;
+}
+
+static constexpr std::pair<const char *, lldb::BasicType> type_suffixes[] = {
+ {"ull", lldb::eBasicTypeUnsignedLongLong},
+ {"ul", lldb::eBasicTypeUnsignedLong},
+ {"u", lldb::eBasicTypeUnsignedInt},
+ {"ll", lldb::eBasicTypeLongLong},
+ {"l", lldb::eBasicTypeLong},
+};
+
+ASTNodeUP DILParser::ParseNumericConstant() {
+ Token token = CurToken();
+ auto spelling = token.GetSpelling();
+ llvm::StringRef spelling_ref = spelling;
+ lldb::BasicType type = lldb::eBasicTypeInt;
+ for (auto [suffix, t] : type_suffixes) {
+ if (spelling_ref.consume_back_insensitive(suffix)) {
+ type = t;
+ break;
+ }
+ }
+ llvm::APInt raw_value;
+ if (!spelling_ref.getAsInteger(0, raw_value)) {
+ Scalar scalar_value(raw_value);
+ return std::make_unique<ScalarLiteralNode>(token.GetLocation(), type,
+ scalar_value);
+ }
+ return std::make_unique<ErrorNode>();
+}
+
void DILParser::Expect(Token::Kind kind) {
if (CurToken().IsNot(kind)) {
BailOut(llvm::formatv("expected {0}, got: {1}", kind, CurToken()),
diff --git a/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py b/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py
index 0f56057189395..c2ab049a28bc2 100644
--- a/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py
+++ b/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py
@@ -88,7 +88,7 @@ def test_subscript(self):
self.expect(
"frame var '1[2]'",
error=True,
- substrs=["Unexpected token"],
+ substrs=["subscripted value is not an array or pointer"],
)
# Base should not be a pointer to void
diff --git a/lldb/test/API/commands/frame/var-dil/basics/Indirection/TestFrameVarDILIndirection.py b/lldb/test/API/commands/frame/var-dil/basics/Indirection/TestFrameVarDILIndirection.py
index 38c72131d797c..28eba4f1a70bc 100644
--- a/lldb/test/API/commands/frame/var-dil/basics/Indirection/TestFrameVarDILIndirection.py
+++ b/lldb/test/API/commands/frame/var-dil/basics/Indirection/TestFrameVarDILIndirection.py
@@ -35,7 +35,7 @@ def test_frame_var(self):
self.expect(
"frame variable '*1'",
error=True,
- substrs=["Unexpected token: <'1' (numeric_constant)>"],
+ substrs=["dereference failed: not a pointer, reference or array type"],
)
self.expect(
"frame variable '*val'",
>From 8a04a676f24666c238840f8e568b9276ad3a86f7 Mon Sep 17 00:00:00 2001
From: Ilia Kuklin <ikuklin at accesssoftek.com>
Date: Fri, 18 Jul 2025 21:52:53 +0500
Subject: [PATCH 2/3] Lex and parse floating numbers; split number tokens to
avoid double parsing
---
lldb/docs/dil-expr-lang.ebnf | 3 ++
lldb/include/lldb/ValueObject/DILLexer.h | 3 +-
lldb/include/lldb/ValueObject/DILParser.h | 3 +-
lldb/source/ValueObject/DILLexer.cpp | 54 +++++++++++++++----
lldb/source/ValueObject/DILParser.cpp | 37 +++++++++++--
.../TestFrameVarDILArraySubscript.py | 2 +-
.../frame/var-dil/expr/Arithmetic/Makefile | 3 ++
.../Arithmetic/TestFrameVarDILArithmetic.py | 30 +++++++++++
.../frame/var-dil/expr/Arithmetic/main.cpp | 3 ++
lldb/unittests/ValueObject/DILLexerTests.cpp | 21 +++++---
10 files changed, 134 insertions(+), 25 deletions(-)
create mode 100644 lldb/test/API/commands/frame/var-dil/expr/Arithmetic/Makefile
create mode 100644 lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py
create mode 100644 lldb/test/API/commands/frame/var-dil/expr/Arithmetic/main.cpp
diff --git a/lldb/docs/dil-expr-lang.ebnf b/lldb/docs/dil-expr-lang.ebnf
index 783432dabd6db..da1796d936c6a 100644
--- a/lldb/docs/dil-expr-lang.ebnf
+++ b/lldb/docs/dil-expr-lang.ebnf
@@ -31,6 +31,9 @@ identifier = ? C99 Identifier ? ;
integer_literal = ? Integer constant: hexademical, decimal, octal, binary ? ;
+numeric_literal = ? Integer constant: hexademical, decimal, octal, binary ?
+ | ? Floating constant ? ;
+
register = "$" ? Register name ? ;
nested_name_specifier = type_name "::"
diff --git a/lldb/include/lldb/ValueObject/DILLexer.h b/lldb/include/lldb/ValueObject/DILLexer.h
index 9c1ba97680253..a9f01785c6c20 100644
--- a/lldb/include/lldb/ValueObject/DILLexer.h
+++ b/lldb/include/lldb/ValueObject/DILLexer.h
@@ -28,11 +28,12 @@ class Token {
arrow,
coloncolon,
eof,
+ floating_constant,
identifier,
+ integer_constant,
l_paren,
l_square,
minus,
- numeric_constant,
period,
r_paren,
r_square,
diff --git a/lldb/include/lldb/ValueObject/DILParser.h b/lldb/include/lldb/ValueObject/DILParser.h
index 2cd8ca3be3c02..90df109337dcf 100644
--- a/lldb/include/lldb/ValueObject/DILParser.h
+++ b/lldb/include/lldb/ValueObject/DILParser.h
@@ -97,7 +97,8 @@ class DILParser {
std::string ParseUnqualifiedId();
std::optional<int64_t> ParseIntegerConstant();
ASTNodeUP ParseNumericLiteral();
- ASTNodeUP ParseNumericConstant();
+ ASTNodeUP ParseIntegerLiteral();
+ ASTNodeUP ParseFloatingPointLiteral();
void BailOut(const std::string &error, uint32_t loc, uint16_t err_len);
diff --git a/lldb/source/ValueObject/DILLexer.cpp b/lldb/source/ValueObject/DILLexer.cpp
index eaefaf484bc18..00f9a0c515461 100644
--- a/lldb/source/ValueObject/DILLexer.cpp
+++ b/lldb/source/ValueObject/DILLexer.cpp
@@ -28,16 +28,18 @@ llvm::StringRef Token::GetTokenName(Kind kind) {
return "coloncolon";
case Kind::eof:
return "eof";
+ case Kind::floating_constant:
+ return "floating_constant";
case Kind::identifier:
return "identifier";
+ case Kind::integer_constant:
+ return "integer_constant";
case Kind::l_paren:
return "l_paren";
case Kind::l_square:
return "l_square";
case Kind::minus:
return "minus";
- case Kind::numeric_constant:
- return "numeric_constant";
case Kind::period:
return "period";
case Kind::r_paren:
@@ -72,12 +74,39 @@ static std::optional<llvm::StringRef> IsWord(llvm::StringRef expr,
static bool IsNumberBodyChar(char ch) { return IsDigit(ch) || IsLetter(ch); }
-static std::optional<llvm::StringRef> IsNumber(llvm::StringRef expr,
- llvm::StringRef &remainder) {
- if (IsDigit(remainder[0])) {
- llvm::StringRef number = remainder.take_while(IsNumberBodyChar);
- remainder = remainder.drop_front(number.size());
- return number;
+static std::optional<llvm::StringRef> IsNumber(llvm::StringRef &remainder,
+ bool &isFloat) {
+ llvm::StringRef::iterator cur_pos = remainder.begin();
+ if (*cur_pos == '.') {
+ auto next_pos = cur_pos + 1;
+ if (next_pos == remainder.end() || !IsDigit(*next_pos))
+ return std::nullopt;
+ }
+ if (IsDigit(*(cur_pos)) || *(cur_pos) == '.') {
+ while (IsNumberBodyChar(*cur_pos))
+ cur_pos++;
+
+ if (*cur_pos == '.') {
+ isFloat = true;
+ cur_pos++;
+ while (IsNumberBodyChar(*cur_pos))
+ cur_pos++;
+
+ // Check if there's an exponent
+ char prev_ch = *(cur_pos - 1);
+ if (prev_ch == 'e' || prev_ch == 'E' || prev_ch == 'p' ||
+ prev_ch == 'P') {
+ if (*(cur_pos) == '+' || *(cur_pos) == '-') {
+ cur_pos++;
+ while (IsNumberBodyChar(*cur_pos))
+ cur_pos++;
+ }
+ }
+ }
+
+ llvm::StringRef number = remainder.substr(0, cur_pos - remainder.begin());
+ if (remainder.consume_front(number))
+ return number;
}
return std::nullopt;
}
@@ -106,9 +135,12 @@ llvm::Expected<Token> DILLexer::Lex(llvm::StringRef expr,
return Token(Token::eof, "", (uint32_t)expr.size());
uint32_t position = cur_pos - expr.begin();
- std::optional<llvm::StringRef> maybe_number = IsNumber(expr, remainder);
- if (maybe_number)
- return Token(Token::numeric_constant, maybe_number->str(), position);
+ bool isFloat = false;
+ std::optional<llvm::StringRef> maybe_number = IsNumber(remainder, isFloat);
+ if (maybe_number) {
+ auto kind = isFloat ? Token::floating_constant : Token::integer_constant;
+ return Token(kind, maybe_number->str(), position);
+ }
std::optional<llvm::StringRef> maybe_word = IsWord(expr, remainder);
if (maybe_word)
return Token(Token::identifier, maybe_word->str(), position);
diff --git a/lldb/source/ValueObject/DILParser.cpp b/lldb/source/ValueObject/DILParser.cpp
index 91b9d764527b3..eee4a7bc4d363 100644
--- a/lldb/source/ValueObject/DILParser.cpp
+++ b/lldb/source/ValueObject/DILParser.cpp
@@ -179,11 +179,12 @@ ASTNodeUP DILParser::ParsePostfixExpression() {
// Parse a primary_expression.
//
// primary_expression:
+// numeric_literal
// id_expression
// "(" expression ")"
//
ASTNodeUP DILParser::ParsePrimaryExpression() {
- if (CurToken().Is(Token::numeric_constant))
+ if (CurToken().IsOneOf({Token::integer_constant, Token::floating_constant}))
return ParseNumericLiteral();
if (CurToken().IsOneOf(
{Token::coloncolon, Token::identifier, Token::l_paren})) {
@@ -348,6 +349,7 @@ void DILParser::BailOut(const std::string &error, uint32_t loc,
m_dil_lexer.ResetTokenIdx(m_dil_lexer.NumLexedTokens() - 1);
}
+// FIXME: Remove this once subscript operator uses ScalarLiteralNode.
// Parse a integer_literal.
//
// integer_literal:
@@ -375,11 +377,15 @@ std::optional<int64_t> DILParser::ParseIntegerConstant() {
// Parse a numeric_literal.
//
// numeric_literal:
-// ? Token::numeric_constant ?
+// ? Token::integer_constant ?
+// ? Token::floating_constant ?
//
ASTNodeUP DILParser::ParseNumericLiteral() {
- Expect(Token::numeric_constant);
- ASTNodeUP numeric_constant = ParseNumericConstant();
+ ASTNodeUP numeric_constant;
+ if (CurToken().Is(Token::integer_constant))
+ numeric_constant = ParseIntegerLiteral();
+ else
+ numeric_constant = ParseFloatingPointLiteral();
if (numeric_constant->GetKind() == NodeKind::eErrorNode) {
BailOut(llvm::formatv("Failed to parse token as numeric-constant: {0}",
CurToken()),
@@ -398,7 +404,7 @@ static constexpr std::pair<const char *, lldb::BasicType> type_suffixes[] = {
{"l", lldb::eBasicTypeLong},
};
-ASTNodeUP DILParser::ParseNumericConstant() {
+ASTNodeUP DILParser::ParseIntegerLiteral() {
Token token = CurToken();
auto spelling = token.GetSpelling();
llvm::StringRef spelling_ref = spelling;
@@ -418,6 +424,27 @@ ASTNodeUP DILParser::ParseNumericConstant() {
return std::make_unique<ErrorNode>();
}
+ASTNodeUP DILParser::ParseFloatingPointLiteral() {
+ Token token = CurToken();
+ auto spelling = token.GetSpelling();
+ llvm::StringRef spelling_ref = spelling;
+ spelling_ref = spelling;
+ lldb::BasicType type = lldb::eBasicTypeDouble;
+ llvm::APFloat raw_float(llvm::APFloat::IEEEdouble());
+ if (spelling_ref.consume_back_insensitive("f")) {
+ type = lldb::eBasicTypeFloat;
+ raw_float = llvm::APFloat(llvm::APFloat::IEEEsingle());
+ }
+ auto StatusOrErr = raw_float.convertFromString(
+ spelling_ref, llvm::APFloat::rmNearestTiesToEven);
+ if (!errorToBool(StatusOrErr.takeError())) {
+ Scalar scalar_value(raw_float);
+ return std::make_unique<ScalarLiteralNode>(token.GetLocation(), type,
+ scalar_value);
+ }
+ return std::make_unique<ErrorNode>();
+}
+
void DILParser::Expect(Token::Kind kind) {
if (CurToken().IsNot(kind)) {
BailOut(llvm::formatv("expected {0}, got: {1}", kind, CurToken()),
diff --git a/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py b/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py
index c2ab049a28bc2..1d0340160f5e4 100644
--- a/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py
+++ b/lldb/test/API/commands/frame/var-dil/basics/ArraySubscript/TestFrameVarDILArraySubscript.py
@@ -66,7 +66,7 @@ def test_subscript(self):
self.expect(
"frame var 'int_arr[1.0]'",
error=True,
- substrs=["expected 'r_square', got: <'.'"],
+ substrs=["failed to parse integer constant: <'1.0' (floating_constant)>"],
)
# Base should be a "pointer to T" and index should be of an integral type.
diff --git a/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/Makefile b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/Makefile
new file mode 100644
index 0000000000000..99998b20bcb05
--- /dev/null
+++ b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/Makefile
@@ -0,0 +1,3 @@
+CXX_SOURCES := main.cpp
+
+include Makefile.rules
diff --git a/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py
new file mode 100644
index 0000000000000..883781873a525
--- /dev/null
+++ b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py
@@ -0,0 +1,30 @@
+"""
+Test DIL arithmetic.
+"""
+
+import lldb
+from lldbsuite.test.lldbtest import *
+from lldbsuite.test.decorators import *
+from lldbsuite.test import lldbutil
+
+
+class TestFrameVarDILArithmetic(TestBase):
+ NO_DEBUG_INFO_TESTCASE = True
+
+ def test_arithmetic(self):
+ self.build()
+ lldbutil.run_to_source_breakpoint(
+ self, "Set a breakpoint here", lldb.SBFileSpec("main.cpp")
+ )
+
+ self.runCmd("settings set target.experimental.use-DIL true")
+
+ # Check number parsing
+ self.expect_var_path("1", value="1", type="int")
+ self.expect_var_path("1ull", value="1", type="unsigned long long")
+ self.expect_var_path("0b10", value="2", type="int")
+ self.expect_var_path("010", value="8", type="int")
+ self.expect_var_path("0x10", value="16", type="int")
+ self.expect_var_path("1.0", value="1", type="double")
+ self.expect_var_path("1.0f", value="1", type="float")
+ self.expect_var_path("0x1.2p+3f", value="9", type="float")
diff --git a/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/main.cpp b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/main.cpp
new file mode 100644
index 0000000000000..c9bd8afb0d71d
--- /dev/null
+++ b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/main.cpp
@@ -0,0 +1,3 @@
+int main(int argc, char **argv) {
+ return 0; // Set a breakpoint here
+}
diff --git a/lldb/unittests/ValueObject/DILLexerTests.cpp b/lldb/unittests/ValueObject/DILLexerTests.cpp
index f65034c1dbea3..ff5234ac64451 100644
--- a/lldb/unittests/ValueObject/DILLexerTests.cpp
+++ b/lldb/unittests/ValueObject/DILLexerTests.cpp
@@ -151,22 +151,31 @@ TEST(DILLexerTests, IdentifiersTest) {
Token token = lexer.GetCurrentToken();
EXPECT_TRUE(token.IsNot(Token::identifier));
EXPECT_TRUE(token.IsOneOf({Token::eof, Token::coloncolon, Token::l_paren,
- Token::r_paren, Token::numeric_constant}));
+ Token::r_paren, Token::integer_constant}));
}
}
TEST(DILLexerTests, NumbersTest) {
// These strings should lex into number tokens.
- std::vector<std::string> valid_numbers = {"123", "0x123", "0123", "0b101"};
+ std::vector<std::string> valid_integers = {"123", "0x123", "0123", "0b101"};
+ std::vector<std::string> valid_floats = {
+ "1.2", ".2", "2.f", "0x1.2", "0x.2", ".2e1f",
+ "2.e+1f", "0x1.f", "0x1.2p1", "0x1.p-1f", "0x1.2p+3f"};
// The lexer can lex these strings, but they should not be numbers.
- std::vector<std::string> invalid_numbers = {"", "x123", "b123"};
+ std::vector<std::string> invalid_numbers = {"", "x123", "b123", "a.b"};
- for (auto &str : valid_numbers) {
+ for (auto &str : valid_integers) {
SCOPED_TRACE(str);
EXPECT_THAT_EXPECTED(ExtractTokenData(str),
llvm::HasValue(testing::ElementsAre(
- testing::Pair(Token::numeric_constant, str))));
+ testing::Pair(Token::integer_constant, str))));
+ }
+ for (auto &str : valid_floats) {
+ SCOPED_TRACE(str);
+ EXPECT_THAT_EXPECTED(ExtractTokenData(str),
+ llvm::HasValue(testing::ElementsAre(
+ testing::Pair(Token::floating_constant, str))));
}
// Verify that none of the invalid numbers come out as numeric tokens.
for (auto &str : invalid_numbers) {
@@ -175,7 +184,7 @@ TEST(DILLexerTests, NumbersTest) {
EXPECT_THAT_EXPECTED(maybe_lexer, llvm::Succeeded());
DILLexer lexer(*maybe_lexer);
Token token = lexer.GetCurrentToken();
- EXPECT_TRUE(token.IsNot(Token::numeric_constant));
+ EXPECT_TRUE(token.IsNot(Token::integer_constant));
EXPECT_TRUE(token.IsOneOf({Token::eof, Token::identifier}));
}
}
>From 27466c3529741d86982e4ebd29764c6812a6c264 Mon Sep 17 00:00:00 2001
From: Ilia Kuklin <ikuklin at accesssoftek.com>
Date: Tue, 5 Aug 2025 21:30:15 +0500
Subject: [PATCH 3/3] Automatically pick integer literal type
---
lldb/include/lldb/ValueObject/DILAST.h | 24 ++++-
lldb/include/lldb/ValueObject/DILEval.h | 5 +
lldb/source/ValueObject/DILEval.cpp | 98 +++++++++++++++++--
lldb/source/ValueObject/DILParser.cpp | 43 ++++----
.../Arithmetic/TestFrameVarDILArithmetic.py | 19 +++-
llvm/include/llvm/ADT/StringRef.h | 2 +
llvm/lib/Support/StringRef.cpp | 6 +-
7 files changed, 153 insertions(+), 44 deletions(-)
diff --git a/lldb/include/lldb/ValueObject/DILAST.h b/lldb/include/lldb/ValueObject/DILAST.h
index a174e28ea5c06..a0e5909a8c6a7 100644
--- a/lldb/include/lldb/ValueObject/DILAST.h
+++ b/lldb/include/lldb/ValueObject/DILAST.h
@@ -181,22 +181,36 @@ class BitFieldExtractionNode : public ASTNode {
class ScalarLiteralNode : public ASTNode {
public:
- ScalarLiteralNode(uint32_t location, lldb::BasicType type, Scalar value)
- : ASTNode(location, NodeKind::eScalarLiteralNode), m_type(type),
- m_value(value) {}
+ ScalarLiteralNode(uint32_t location, Scalar value, uint32_t radix,
+ bool is_unsigned, bool is_long, bool is_longlong)
+ : ASTNode(location, NodeKind::eScalarLiteralNode), m_value(value),
+ m_radix(radix), m_is_unsigned(is_unsigned), m_is_long(is_long),
+ m_is_longlong(is_longlong) {}
+
+ ScalarLiteralNode(uint32_t location, Scalar value, bool is_float)
+ : ASTNode(location, NodeKind::eScalarLiteralNode), m_value(value),
+ m_is_float(is_float) {}
llvm::Expected<lldb::ValueObjectSP> Accept(Visitor *v) const override;
- lldb::BasicType GetType() const { return m_type; }
Scalar GetValue() const & { return m_value; }
+ uint32_t GetRadix() const { return m_radix; }
+ bool IsUnsigned() const { return m_is_unsigned; }
+ bool IsLong() const { return m_is_long; }
+ bool IsLongLong() const { return m_is_longlong; }
+ bool IsFloat() const { return m_is_float; }
static bool classof(const ASTNode *node) {
return node->GetKind() == NodeKind::eScalarLiteralNode;
}
private:
- lldb::BasicType m_type;
Scalar m_value;
+ uint32_t m_radix;
+ bool m_is_unsigned;
+ bool m_is_long;
+ bool m_is_longlong;
+ bool m_is_float;
};
/// This class contains one Visit method for each specialized type of
diff --git a/lldb/include/lldb/ValueObject/DILEval.h b/lldb/include/lldb/ValueObject/DILEval.h
index cb2a81d1c7ba1..22a6c5bd0af9a 100644
--- a/lldb/include/lldb/ValueObject/DILEval.h
+++ b/lldb/include/lldb/ValueObject/DILEval.h
@@ -57,6 +57,11 @@ class Interpreter : Visitor {
llvm::Expected<lldb::ValueObjectSP>
Visit(const ScalarLiteralNode *node) override;
+ llvm::Expected<CompilerType>
+ PickLiteralType(lldb::TypeSystemSP type_system,
+ std::shared_ptr<ExecutionContextScope> ctx,
+ const ScalarLiteralNode *literal);
+
// Used by the interpreter to create objects, perform casts, etc.
lldb::TargetSP m_target;
llvm::StringRef m_expr;
diff --git a/lldb/source/ValueObject/DILEval.cpp b/lldb/source/ValueObject/DILEval.cpp
index 18cc30d589829..ee667e9306ca5 100644
--- a/lldb/source/ValueObject/DILEval.cpp
+++ b/lldb/source/ValueObject/DILEval.cpp
@@ -404,8 +404,7 @@ Interpreter::Visit(const BitFieldExtractionNode *node) {
return child_valobj_sp;
}
-static CompilerType GetBasicTypeFromCU(std::shared_ptr<StackFrame> ctx,
- lldb::BasicType basic_type) {
+static lldb::TypeSystemSP GetTypeSystemFromCU(std::shared_ptr<StackFrame> ctx) {
SymbolContext symbol_context =
ctx->GetSymbolContext(lldb::eSymbolContextCompUnit);
auto language = symbol_context.comp_unit->GetLanguage();
@@ -414,20 +413,103 @@ static CompilerType GetBasicTypeFromCU(std::shared_ptr<StackFrame> ctx,
auto type_system =
symbol_context.module_sp->GetTypeSystemForLanguage(language);
+ if (type_system)
+ return *type_system;
+
+ return lldb::TypeSystemSP();
+}
+
+static CompilerType GetBasicType(lldb::TypeSystemSP type_system,
+ lldb::BasicType basic_type) {
if (type_system)
if (auto compiler_type = type_system.get()->GetBasicTypeFromAST(basic_type))
return compiler_type;
- return CompilerType();
+ CompilerType empty_type;
+ return empty_type;
+}
+
+llvm::Expected<CompilerType>
+Interpreter::PickLiteralType(lldb::TypeSystemSP type_system,
+ std::shared_ptr<ExecutionContextScope> ctx,
+ const ScalarLiteralNode *literal) {
+ Scalar scalar = literal->GetValue();
+ if (scalar.GetType() == Scalar::e_float) {
+ if (literal->IsFloat())
+ return GetBasicType(type_system, lldb::eBasicTypeFloat);
+ return GetBasicType(type_system, lldb::eBasicTypeDouble);
+ } else if (scalar.GetType() == Scalar::e_int) {
+ // Binary, Octal, Hexadecimal and literals with a U suffix are allowed to be
+ // an unsigned integer.
+ bool unsigned_is_allowed =
+ literal->IsUnsigned() || literal->GetRadix() != 10;
+
+ // Try int/unsigned int.
+ uint64_t int_byte_size = 0;
+ if (auto temp = GetBasicType(type_system, lldb::eBasicTypeInt)
+ .GetByteSize(ctx.get()))
+ int_byte_size = *temp;
+ unsigned int_size = int_byte_size * CHAR_BIT;
+ llvm::APInt apint = scalar.GetAPSInt();
+ if (!literal->IsLong() && !literal->IsLongLong() &&
+ apint.isIntN(int_size)) {
+ if (!literal->IsUnsigned() && apint.isIntN(int_size - 1))
+ return GetBasicType(type_system, lldb::eBasicTypeInt);
+ if (unsigned_is_allowed)
+ return GetBasicType(type_system, lldb::eBasicTypeUnsignedInt);
+ }
+ // Try long/unsigned long.
+ uint64_t long_byte_size = 0;
+ if (auto temp = GetBasicType(type_system, lldb::eBasicTypeLong)
+ .GetByteSize(ctx.get()))
+ long_byte_size = *temp;
+ unsigned long_size = long_byte_size * CHAR_BIT;
+ if (!literal->IsLongLong() && apint.isIntN(long_size)) {
+ if (!literal->IsUnsigned() && apint.isIntN(long_size - 1))
+ return GetBasicType(type_system, lldb::eBasicTypeLong);
+ if (unsigned_is_allowed)
+ return GetBasicType(type_system, lldb::eBasicTypeUnsignedLong);
+ }
+ // Try long long/unsigned long long.
+ uint64_t long_long_byte_size = 0;
+ if (auto temp = GetBasicType(type_system, lldb::eBasicTypeLongLong)
+ .GetByteSize(ctx.get()))
+ long_long_byte_size = *temp;
+ unsigned long_long_size = long_long_byte_size * CHAR_BIT;
+ if (apint.isIntN(long_long_size)) {
+ if (!literal->IsUnsigned() && apint.isIntN(long_long_size - 1))
+ return GetBasicType(type_system, lldb::eBasicTypeLongLong);
+ // If we still couldn't decide a type, we probably have something that
+ // does not fit in a signed long long, but has no U suffix. Also known as:
+ //
+ // warning: integer literal is too large to be represented in a signed
+ // integer type, interpreting as unsigned [-Wimplicitly-unsigned-literal]
+ //
+ return GetBasicType(type_system, lldb::eBasicTypeUnsignedLongLong);
+ }
+ return llvm::make_error<DILDiagnosticError>(
+ m_expr,
+ "integer literal is too large to be represented in any integer type",
+ literal->GetLocation());
+ }
+ return llvm::make_error<DILDiagnosticError>(
+ m_expr, "unable to create a const literal", literal->GetLocation());
}
llvm::Expected<lldb::ValueObjectSP>
Interpreter::Visit(const ScalarLiteralNode *node) {
- CompilerType result_type =
- GetBasicTypeFromCU(m_exe_ctx_scope, node->GetType());
- Scalar value = node->GetValue();
- return ValueObject::CreateValueObjectFromScalar(m_target, value, result_type,
- "result");
+ auto type_system = GetTypeSystemFromCU(m_exe_ctx_scope);
+ if (type_system) {
+ auto type = PickLiteralType(type_system, m_exe_ctx_scope, node);
+ if (type) {
+ Scalar scalar = node->GetValue();
+ return ValueObject::CreateValueObjectFromScalar(m_target, scalar, *type,
+ "result");
+ } else
+ return type.takeError();
+ }
+ return llvm::make_error<DILDiagnosticError>(
+ m_expr, "unable to create a const literal", node->GetLocation());
}
} // namespace lldb_private::dil
diff --git a/lldb/source/ValueObject/DILParser.cpp b/lldb/source/ValueObject/DILParser.cpp
index eee4a7bc4d363..35eb6d62b7ba4 100644
--- a/lldb/source/ValueObject/DILParser.cpp
+++ b/lldb/source/ValueObject/DILParser.cpp
@@ -396,30 +396,26 @@ ASTNodeUP DILParser::ParseNumericLiteral() {
return numeric_constant;
}
-static constexpr std::pair<const char *, lldb::BasicType> type_suffixes[] = {
- {"ull", lldb::eBasicTypeUnsignedLongLong},
- {"ul", lldb::eBasicTypeUnsignedLong},
- {"u", lldb::eBasicTypeUnsignedInt},
- {"ll", lldb::eBasicTypeLongLong},
- {"l", lldb::eBasicTypeLong},
-};
-
ASTNodeUP DILParser::ParseIntegerLiteral() {
Token token = CurToken();
auto spelling = token.GetSpelling();
llvm::StringRef spelling_ref = spelling;
- lldb::BasicType type = lldb::eBasicTypeInt;
- for (auto [suffix, t] : type_suffixes) {
- if (spelling_ref.consume_back_insensitive(suffix)) {
- type = t;
- break;
- }
- }
+
+ auto radix = llvm::getAutoSenseRadix(spelling_ref);
+ bool is_unsigned = false, is_long = false, is_longlong = false;
+ if (spelling_ref.consume_back_insensitive("ll"))
+ is_longlong = true;
+ if (spelling_ref.consume_back_insensitive("l"))
+ is_long = true;
+ if (spelling_ref.consume_back_insensitive("u"))
+ is_unsigned = true;
+
llvm::APInt raw_value;
- if (!spelling_ref.getAsInteger(0, raw_value)) {
+ if (!spelling_ref.getAsInteger(radix, raw_value)) {
Scalar scalar_value(raw_value);
- return std::make_unique<ScalarLiteralNode>(token.GetLocation(), type,
- scalar_value);
+ return std::make_unique<ScalarLiteralNode>(token.GetLocation(),
+ scalar_value, radix, is_unsigned,
+ is_long, is_longlong);
}
return std::make_unique<ErrorNode>();
}
@@ -428,19 +424,20 @@ ASTNodeUP DILParser::ParseFloatingPointLiteral() {
Token token = CurToken();
auto spelling = token.GetSpelling();
llvm::StringRef spelling_ref = spelling;
- spelling_ref = spelling;
- lldb::BasicType type = lldb::eBasicTypeDouble;
+
+ bool is_float = false;
llvm::APFloat raw_float(llvm::APFloat::IEEEdouble());
if (spelling_ref.consume_back_insensitive("f")) {
- type = lldb::eBasicTypeFloat;
+ is_float = true;
raw_float = llvm::APFloat(llvm::APFloat::IEEEsingle());
}
+
auto StatusOrErr = raw_float.convertFromString(
spelling_ref, llvm::APFloat::rmNearestTiesToEven);
if (!errorToBool(StatusOrErr.takeError())) {
Scalar scalar_value(raw_float);
- return std::make_unique<ScalarLiteralNode>(token.GetLocation(), type,
- scalar_value);
+ return std::make_unique<ScalarLiteralNode>(token.GetLocation(),
+ scalar_value, is_float);
}
return std::make_unique<ErrorNode>();
}
diff --git a/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py
index 883781873a525..57a636ebb0829 100644
--- a/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py
+++ b/lldb/test/API/commands/frame/var-dil/expr/Arithmetic/TestFrameVarDILArithmetic.py
@@ -20,11 +20,20 @@ def test_arithmetic(self):
self.runCmd("settings set target.experimental.use-DIL true")
# Check number parsing
- self.expect_var_path("1", value="1", type="int")
- self.expect_var_path("1ull", value="1", type="unsigned long long")
- self.expect_var_path("0b10", value="2", type="int")
- self.expect_var_path("010", value="8", type="int")
- self.expect_var_path("0x10", value="16", type="int")
self.expect_var_path("1.0", value="1", type="double")
self.expect_var_path("1.0f", value="1", type="float")
self.expect_var_path("0x1.2p+3f", value="9", type="float")
+ self.expect_var_path("1", value="1", type="int")
+ self.expect_var_path("1u", value="1", type="unsigned int")
+ self.expect_var_path("0b1l", value="1", type="long")
+ self.expect_var_path("01ul", value="1", type="unsigned long")
+ self.expect_var_path("0o1ll", value="1", type="long long")
+ self.expect_var_path("0x1ULL", value="1", type="unsigned long long")
+ self.expect_var_path("0xFFFFFFFFFFFFFFFF", value="18446744073709551615")
+ self.expect(
+ "frame var '0xFFFFFFFFFFFFFFFFF'",
+ error=True,
+ substrs=[
+ "integer literal is too large to be represented in any integer type"
+ ],
+ )
diff --git a/llvm/include/llvm/ADT/StringRef.h b/llvm/include/llvm/ADT/StringRef.h
index 0ced1c0379a3b..16aca4d45892d 100644
--- a/llvm/include/llvm/ADT/StringRef.h
+++ b/llvm/include/llvm/ADT/StringRef.h
@@ -38,6 +38,8 @@ namespace llvm {
LLVM_ABI bool getAsSignedInteger(StringRef Str, unsigned Radix,
long long &Result);
+ LLVM_ABI unsigned getAutoSenseRadix(StringRef &Str);
+
LLVM_ABI bool consumeUnsignedInteger(StringRef &Str, unsigned Radix,
unsigned long long &Result);
LLVM_ABI bool consumeSignedInteger(StringRef &Str, unsigned Radix,
diff --git a/llvm/lib/Support/StringRef.cpp b/llvm/lib/Support/StringRef.cpp
index dc758785e40d5..b6a2f8aeadccf 100644
--- a/llvm/lib/Support/StringRef.cpp
+++ b/llvm/lib/Support/StringRef.cpp
@@ -385,7 +385,7 @@ size_t StringRef::count(StringRef Str) const {
return Count;
}
-static unsigned GetAutoSenseRadix(StringRef &Str) {
+unsigned llvm::getAutoSenseRadix(StringRef &Str) {
if (Str.empty())
return 10;
@@ -410,7 +410,7 @@ bool llvm::consumeUnsignedInteger(StringRef &Str, unsigned Radix,
unsigned long long &Result) {
// Autosense radix if not specified.
if (Radix == 0)
- Radix = GetAutoSenseRadix(Str);
+ Radix = getAutoSenseRadix(Str);
// Empty strings (after the radix autosense) are invalid.
if (Str.empty()) return true;
@@ -509,7 +509,7 @@ bool StringRef::consumeInteger(unsigned Radix, APInt &Result) {
// Autosense radix if not specified.
if (Radix == 0)
- Radix = GetAutoSenseRadix(Str);
+ Radix = getAutoSenseRadix(Str);
assert(Radix > 1 && Radix <= 36);
More information about the llvm-commits
mailing list