Functions — github.com/anthropics/claudes-c-compiler

hub / github.com/anthropics/claudes-c-compiler / functions

Functions6,057 in github.com/anthropics/claudes-c-compiler

⨍ Functions 6,057◇ Types & classes 346

Make a path absolute without resolving symlinks. Unlike `std::fs::canonicalize`, this preserves symlinks in the path. This is important for `#include

src/frontend/preprocessor/includes.rs:229

↓ 10 callersFunctionmake_func

Helper to create a minimal IrFunction with given blocks.

src/passes/gvn.rs:1244

↓ 10 callersFunctionmake_module

(func: IrFunction)

src/passes/gvn.rs:1272

↓ 10 callersFunctionmnemonic_size_suffix

Get operand size from mnemonic suffix.

src/backend/x86/assembler/encoder/registers.rs:201

↓ 10 callersMethodoperand_to_vn

Convert an Operand to a VNOperand for hashing. If the value hasn't been assigned a value number yet (e.g. a function parameter or an alloca whose defi

src/passes/gvn.rs:158

↓ 10 callersMethodparse_binary_expr

Parse a left-associative binary expression at the given precedence level. This is the shared core that replaces 10 nearly-identical parsing functions.

src/frontend/parser/expressions.rs:153

↓ 10 callersMethodparse_cast_expr

Parse a cast expression: (type-name)expr, compound literal (type-name){...}, or fall through to unary expression.

src/frontend/parser/expressions.rs:182

↓ 10 callersMethodparse_param_list

Parse a function parameter list: (params...) or (void) or ()

src/frontend/parser/declarators.rs:312

↓ 10 callersMethodparsing_typedef

(&self)

src/frontend/parser/parse.rs:150

↓ 10 callersFunctionphys_reg_name_32

Map a PhysReg index to its x86-64 32-bit sub-register name. Handles both callee-saved (1-5) and caller-saved (10-15) registers.

src/backend/x86/codegen/emit.rs:76

↓ 10 callersMethodrecord_use

Record a use of an alloca (or GEP-derived value) in a block.

src/backend/stack_layout/alloca_coalescing.rs:70

↓ 10 callersMethodresolve_init_field

Resolve which field an initializer targets, with full info about anonymous members. When a designator name is found inside an anonymous struct/union

src/common/types.rs:1029

↓ 10 callersFunctionresolve_numeric_name

Resolve a numeric label reference to its unique name.

src/backend/arm/assembler/mod.rs:152

↓ 10 callersMethodresolve_to_global_name

Resolve a variable name to its global name, checking static local names first.

src/ir/lowering/const_eval_global_addr.rs:45

↓ 10 callersMethodset_thread_local

(&mut self, v: bool)

src/frontend/parser/ast.rs:256

↓ 10 callersFunctionstrip_comment

(line: &str)

src/backend/arm/assembler/parser.rs:1322

↓ 10 callersMethodto_str

Get the path as a string slice (panics if not valid UTF-8).

src/common/temp_files.rs:76

↓ 10 callersFunctionwrite_i64_fast

(buf: &mut String, val: i64)

src/backend/common.rs:756

↓ 9 callersFunctionbuild_cfg

Build predecessor and successor lists from the function's CFG. Returns (preds, succs) as flat adjacency lists (CSR format). Uses only 4 heap allocati

src/ir/analysis.rs:110

↓ 9 callersFunctionbuild_label_map

Build a map from block label to block index.

src/ir/analysis.rs:95

↓ 9 callersMethodcoerce_const_to_type_with_src

Coerce a constant to the target type, using the source expression's type for signedness.

src/ir/lowering/const_eval.rs:624

↓ 9 callersFunctioncompute_dominators

Compute immediate dominators using the Cooper-Harvey-Kennedy algorithm. Returns idom[i] = immediate dominator of block i (idom[0] = 0 for entry). Uses

src/ir/analysis.rs:238

↓ 9 callersMethodemit

Emit a diagnostic: apply warning filtering/promotion, print to stderr, and update counts. For warnings with a `WarningKind`: - If the warning is disa

src/common/error.rs:504

↓ 9 callersMethodemit_array_element_store

Emit a single element store at a given byte offset in an alloca.

src/ir/lowering/lower.rs:1378

↓ 9 callersFunctionenc

(sign: bool, exp: u16, mant: u64)

src/common/long_double.rs:2993

↓ 9 callersMethodencode_bt

Encode bit test instructions (bt, bts, btr, btc).

src/backend/x86/assembler/encoder/x87_misc.rs:7

↓ 9 callersFunctionencode_f128

Encode an IEEE 754 binary128 value from sign, binary exponent, and 113-bit mantissa. `binary_exp` is the exponent of the MSB (bit 112) of `mantissa113

src/common/long_double.rs:734

↓ 9 callersFunctionf128_operand_to_arg1

Load an F128 operand into the first argument position with full precision. Three paths: 1. **Constant**: load f128 bytes directly as lo:hi. 2. **Trac

src/backend/f128_softfloat.rs:271

↓ 9 callersMethodid

(&self)

src/frontend/parser/ast.rs:811

↓ 9 callersMethodintern_string_literal

Intern a string literal: add it to the module's .rodata string table and return its unique label.

src/ir/lowering/lower.rs:1058

↓ 9 callersMethodis_defined

Check if a macro is defined.

src/frontend/preprocessor/macro_defs.rs:152

↓ 9 callersMethodis_noreturn

(&self)

src/frontend/parser/ast.rs:90

↓ 9 callersFunctionload_file

( path: &str, objects: &mut Vec<ElfObject>, globals: &mut HashMap<String, GlobalSymbol>, needed_soname

src/backend/x86/linker/input.rs:13

↓ 9 callersMethodload_ptr_to_reg

Load the address represented by a pointer Value into the given register. For alloca values, computes the address; for others, loads the stored pointer

src/backend/arm/codegen/emit.rs:1211

↓ 9 callersFunctionmake_f128_nan

(negative: bool)

src/common/long_double.rs:879

↓ 9 callersMethodoperand_to_ecx

Load an operand into %ecx.

src/backend/i686/codegen/emit.rs:326

↓ 9 callersFunctionparse_long_double_to_f128_bytes

Parse a float string directly to IEEE 754 binary128 (f128) bytes with full 112-bit mantissa precision. Used for long double constants on ARM64/RISC-V

src/common/long_double.rs:713

↓ 9 callersFunctionparse_reg

Parse a register name to our internal ID.

src/backend/riscv/codegen/peephole.rs:120

↓ 9 callersFunctionparse_st_num

Parse x87 register number: "st(0)" -> 0, "st" -> 0, "st(1)" -> 1, etc.

src/backend/x86/assembler/encoder/registers.rs:260

↓ 9 callersFunctionparse_string_literal

String literal parser for assembler directives. Shared by all four assembler backends (x86, i686, ARM, RISC-V) to ensure consistent handling of C/GNU

src/backend/elf/parse_string.rs:22

↓ 9 callersFunctionpromote_allocas

Promote allocas to SSA form with phi insertion. Only promotes scalar allocas that are exclusively loaded/stored (not address-taken by GEP, memcpy, va_

src/ir/mem2reg/promote.rs:42

↓ 9 callersFunctionreg_to_8l

Convert any x86 register name to its 8-bit low variant. Accepts both 64-bit and 32/16/8-bit register names. For `r8`-`r15` (x86-64 only), appends `b`

src/backend/x86_common.rs:136

↓ 9 callersFunctionregister_family

Map i686 register name to family ID.

src/backend/i686/codegen/peephole.rs:124

↓ 9 callersFunctionremap_block

Remap a BlockId by adding an offset.

src/passes/inline.rs:1470

↓ 9 callersMethodresolve_vector_size

Resolve the total vector size in bytes, considering both `vector_size` (total bytes) and `ext_vector_type` (element count). `elem_size` is the sizeof

src/frontend/parser/ast.rs:310

↓ 9 callersMethodreturns_complex_long_double_in_regs

Returns true if the target returns _Complex long double via register pairs rather than via sret hidden pointer. On x86-64: true (COMPLEX_X87 class, re

src/ir/lowering/lower.rs:298

↓ 9 callersMethodtrack_f128_self

(&mut self, value_id: u32)

src/backend/state.rs:371

↓ 9 callersMethodvector_info

For a vector type, returns (element_type, num_elements). Returns None for non-vector types.

src/common/types.rs:1432

↓ 9 callersMethodwith_span

Attach a source span to this diagnostic.

src/common/error.rs:404

↓ 9 callersFunctionx87_mul

Multiply two x87 80-bit extended precision values with full precision.

src/common/long_double.rs:1179

↓ 8 callersFunction_mm_getcsr

Read the MXCSR register */

include/xmmintrin.h:613

↓ 8 callersMethodbuild_full_ctype

Build a full CType from a TypeSpecifier and DerivedDeclarator chain. Delegates to the shared type_builder module for canonical inside-out declarator a

src/ir/lowering/types_ctype.rs:385

↓ 8 callersMethodbyte_size

Return the number of bytes this move size covers.

src/backend/x86/codegen/peephole/types.rs:150

↓ 8 callersFunctionclassify_line

Parse one assembly line into a `LineInfo`.

src/backend/x86/codegen/peephole/types.rs:201

↓ 8 callersMethodcoerce_to

Coerce this constant to match a target IrType (assumes signed source for int-to-float).

src/ir/constants.rs:550

↓ 8 callersMethodemit_branch_to_block

(&mut self, block: BlockId)

src/backend/riscv/codegen/emit.rs:502

↓ 8 callersMethodemit_compound_field_init

Emit a single field initializer in compound (relocation-aware) mode.

src/ir/lowering/global_init_compound_struct.rs:501

↓ 8 callersMethodemit_f128_classify_libcall

Emit a call to a libc FP classification function for F128 (long double) arguments. Returns the call result as a Value. The libc functions take a long

src/ir/lowering/expr_builtins_fpclass.rs:32

↓ 8 callersMethodemit_placeholder

Emit placeholder bytes for a deferred value (symbol diff, etc.).

src/backend/elf/writer_base.rs:591

↓ 8 callersMethodemit_store_at_offset

Emit a GEP + Store: store `val` at `base + byte_offset` with the given type.

src/ir/lowering/lower.rs:1126

↓ 8 callersMethodemit_store_result

(&mut self, dest: &Value)

src/backend/i686/codegen/emit.rs:1274

↓ 8 callersMethodemit_sub_struct_to_compound

Emit a sub-struct's initialization into compound elements, choosing between compound (relocation-aware) and byte-level approaches based on whether any

src/ir/lowering/global_init_compound_struct.rs:417

↓ 8 callersMethodencode_alu

(&mut self, ops: &[Operand], mnemonic: &str, alu_op: u8)

src/backend/i686/assembler/encoder/gp_integer.rs:433

↓ 8 callersMethodencode_avx_3op_3a_imm8

Encode AVX 3-operand in 0F3A map with imm8

src/backend/x86/assembler/encoder/avx.rs:742

↓ 8 callersMethodencode_avx_shift

Encode AVX shift instructions (imm8 form or xmm form)

src/backend/x86/assembler/encoder/avx.rs:978

↓ 8 callersFunctionencode_fcvt_from_int

(operands: &[Operand], funct7: u32, rs2: u32)

src/backend/riscv/assembler/encoder/float.rs:131

↓ 8 callersFunctionencode_fcvt_int

(operands: &[Operand], funct7: u32, rs2: u32)

src/backend/riscv/assembler/encoder/float.rs:116

↓ 8 callersFunctionencode_fma

(operands: &[Operand], opcode: u32, fmt: u32)

src/backend/riscv/assembler/encoder/float.rs:175

↓ 8 callersFunctionencode_fp_arith

(operands: &[Operand], opcode: u32)

src/backend/arm/assembler/encoder/fp_scalar.rs:68

↓ 8 callersMethodencode_inc_dec

Encode INC/DEC using Group 5 opcode (0xFE/0xFF), not Group 3 (0xF6/0xF7).

src/backend/x86/assembler/encoder/gp_integer.rs:733

↓ 8 callersFunctionencode_neon_ld_st_dispatch

Encode NEON LD1 (vector load, multiple structures) Dispatch LD/ST1-4: choose between "multiple structures" and "single structure (element)" encoding.

src/backend/arm/assembler/encoder/neon.rs:889

↓ 8 callersFunctionencode_neon_qshrn

── NEON shift right narrow saturating (SQSHRN/UQSHRN/SQRSHRN/UQRSHRN) ─

src/backend/arm/assembler/encoder/neon.rs:1496

↓ 8 callersFunctionencode_neon_three_diff_narrow

── NEON ADDHN/RADDHN/SUBHN/RSUBHN ────────────────────────────────────── Three-different narrowing high: Format: 0 Q U 01110 size 1 Rm opcode 00 Rn Rd

src/backend/arm/assembler/encoder/neon.rs:1514

↓ 8 callersFunctionencode_neon_two_misc_narrow

Encode NEON two-register miscellaneous narrowing: UQXTN, SQXTN, XTN Format: 0 Q U 01110 size 10000 opcode 10 Rn Rd

src/backend/arm/assembler/encoder/neon.rs:206

↓ 8 callersMethodencode_shift

(&mut self, ops: &[Operand], mnemonic: &str, shift_op: u8)

src/backend/x86/assembler/encoder/gp_integer.rs:756

↓ 8 callersFunctionencode_shift_imm

(operands: &[Operand], funct3: u32, funct6: u32)

src/backend/riscv/assembler/encoder/base.rs:251

↓ 8 callersFunctionencode_shift_imm_w

(operands: &[Operand], funct3: u32, funct7: u32)

src/backend/riscv/assembler/encoder/base.rs:274

↓ 8 callersMethodencode_sse_rr_rm

(&mut self, ops: &[Operand], load_opcode: &[u8], store_opcode: &[u8])

src/backend/i686/assembler/encoder/sse.rs:11

↓ 8 callersMethodencode_sse_shift

(&mut self, ops: &[Operand], reg_opcode: &[u8], imm_ext: u8, imm_opcode: &[u8])

src/backend/x86/assembler/encoder/sse.rs:325

↓ 8 callersMethodencode_sse_shift

(&mut self, ops: &[Operand], _reg_opcode: &[u8], imm_ext: u8, imm_opcode: &[u8])

src/backend/i686/assembler/encoder/sse.rs:159

↓ 8 callersMethodencode_suffixless_alu

(&mut self, ops: &[Operand], alu_op: u8)

src/backend/x86/assembler/encoder/x87_misc.rs:337

↓ 8 callersFunctionf128_bytes_to_f64

Convert f128 bytes to f64 (lossy narrowing).

src/common/long_double.rs:906

↓ 8 callersFunctionf128_bytes_to_u64

Convert f128 bytes to u64 (for constant folding).

src/common/long_double.rs:1106

↓ 8 callersFunctionfind_conflicting_phis

Determine which phi copies on a given predecessor edge need temporaries. A copy `dest_i = src_i` needs a temporary if `src_i` is the destination of a

src/ir/mem2reg/phi_eliminate.rs:199

↓ 8 callersFunctionfind_natural_loops

Find all natural loops in the CFG. A natural loop is defined by a back edge (tail -> header) where the header dominates the tail. The loop body is th

src/passes/loop_analysis.rs:22

↓ 8 callersFunctionfirst_field_designator

Extract the field name from the first designator of an initializer item. Returns `None` if the item has no designators or the first is not a Field.

src/ir/lowering/global_init_helpers.rs:19

↓ 8 callersMethodget_pointer_elem_size_from_expr

Get the element size for a pointer expression (for scaling in pointer arithmetic). For `int *p`, returns 4. For `char *s`, returns 1.

src/ir/lowering/pointer_analysis.rs:286

↓ 8 callersMethodinit_has_addr_exprs

Check if an initializer contains address expressions (recursive).

src/ir/lowering/global_init.rs:1456

↓ 8 callersMethodis_func_ptr_variable

Check if an identifier is a function pointer variable rather than a direct function. A local variable always shadows a known function of the same name

src/ir/lowering/expr_calls.rs:57

↓ 8 callersMethodis_inline

(&self)

src/frontend/parser/ast.rs:79

↓ 8 callersFunctionis_null_pointer_constant

Check if an AST expression is a null pointer constant per C11 6.3.2.3p3: "An integer constant expression with the value 0, or such an expression cast

src/common/const_arith.rs:467

↓ 8 callersFunctionis_reg64

Is this a 64-bit GP register?

src/backend/x86/assembler/encoder/registers.rs:98

↓ 8 callersMethodlower_and_cast_init_expr

Lower an expression and cast it to the target type (e.g., int to float).

src/ir/lowering/stmt.rs:1056

↓ 8 callersMethodlower_complex_to_bool

Convert a complex value (given as a pointer to {real, imag}) to _Bool. Returns (real != 0) || (imag != 0) per C11 6.3.1.2.

src/ir/lowering/complex.rs:52

↓ 8 callersMethodlower_va_list_pointer

Get a pointer to the va_list struct from an expression. Used by va_start, va_end, va_copy builtins. Target-dependent behavior (same logic as lower_va

src/ir/lowering/expr_access.rs:1421

↓ 8 callersFunctionmake_f128_infinity

(negative: bool)

src/common/long_double.rs:875

↓ 8 callersFunctionmake_func_with_blocks

(blocks: Vec<BasicBlock>)

src/passes/narrow.rs:649

↓ 8 callersMethodmark_operand_escaped

Mark an operand's value as escaped (no use-block recording).

src/backend/stack_layout/alloca_coalescing.rs:89

↓ 8 callersFunctionnarrow_function

Narrow operations in a single function.

src/passes/narrow.rs:57

↓ 8 callersMethodoperand_to_rax_rdx

Load a 128-bit operand into %rax (low) and %rdx (high).

src/backend/x86/codegen/emit.rs:636

← previous next →401–500 of 6,057, ranked by callers