Restore exception handling for `||=` operator on class variables and constants
that was inadvertently removed in commit 0ca48e24f. When reading an undefined
class variable with GETCV opcode raises NameError, the exception handler
catches it and loads false, allowing the assignment to proceed.
Co-authored-by: Claude <noreply@anthropic.com>
pr #6643 fixed node type check but introduced a bug by accessing t->cdr
on a NODE_ARRAY structure. NODE_ARRAY nodes use the elements field, not
cdr. additionally, the fixed rhs path needs to update rhs_reg to point
to where values are actually pushed on the stack.
Co-authored-by: Claude <noreply@anthropic.com>
segment nodes allocated with mrbc_malloc were leaked if gen_string
raised an exception via longjmp. fix by avoiding allocation entirely:
temporarily modify tree structure by saving and clearing cdr pointer,
call gen_string, then restore cdr. no memory is allocated so nothing
leaks even on longjmp.
Co-authored-by: Claude <noreply@anthropic.com>
The bug was in codegen_colon3 which used genop_2(OP_OCLASS, sym)
treating OCLASS as BB format, but OCLASS is B format that only
loads ::Object without a symbol parameter. The fix uses the correct
two-instruction pattern: OCLASS to load Object class, then GETMCNST
to retrieve the constant from it.
Co-authored-by: Claude <noreply@anthropic.com>
cast uint8_t node_type field to enum node_type to satisfy c++ stricter
type checking while maintaining memory efficiency of 1-byte storage.
Co-authored-by: Claude <noreply@anthropic.com>
Remove migration-stage "Phase" and "Group" references from comments,
replacing them with descriptions of actual code organization.
Co-authored-by: Claude <noreply@anthropic.com>
replaced all *_NODE_* accessor macros (e.g., SYM_NODE_VALUE,
INT_NODE_VALUE, CALL_NODE_METHOD) with direct member access using
casting macros (e.g., sym_node(n)->symbol, int_node(n)->value,
call_node(n)->method_name). this eliminates an unnecessary abstraction
layer and improves code readability by making field access explicit.
the accessor macros simply wrapped cast_func(n)->field, providing no
real benefit. direct member access makes it clear what field is being
accessed and reduces macro indirection.
affected files:
- node.h: removed ~100 accessor macro definitions
- codegen.c: replaced 19 macro uses with direct access
- parse.y: replaced 152 macro uses with direct access
Co-authored-by: Claude <noreply@anthropic.com>
replaced unnecessary macro usage with direct struct member access when
struct pointers are already available:
- return_n->args instead of RETURN_NODE_ARGS(return_n)
- yield_n->args instead of YIELD_NODE_ARGS(yield_n)
- for_n->var/iterable/body instead of FOR_NODE_VAR/ITERABLE/BODY(for_n)
- class_n->name/superclass/body instead of CLASS_NODE_* macros
- module_n->name/body instead of MODULE_NODE_NAME/BODY(module_n)
- sclass_n->obj/body instead of SCLASS_NODE_OBJ/BODY(sclass_n)
- hash->pairs instead of HASH_NODE_PAIRS(hash)
- call->method_name/safe_call instead of CALL_NODE_METHOD/SAFE(call)
- array->elements and an->elements instead of ARRAY_NODE_ELEMENTS macro
- splat->value instead of SPLAT_NODE_VALUE macro
improves code readability by removing unnecessary indirection.
Co-authored-by: Claude <noreply@anthropic.com>
Refactored NODE_DSYM to use unified structure directly instead of wrapping
NODE_STR. This eliminates unnecessary allocation and simplifies the AST.
Changes:
- new_dsym() now creates NODE_DSYM directly with mrb_ast_str_node structure
- Parser calls new_dsym(p, n) instead of new_dsym(p, new_str(p, n))
- codegen_dsym() uses gen_string() for proper string generation
- NODE_DSYM dump uses dump_str() for consistent string list handling
- Removed redundant mrb_ast_dsym_node struct definition
This maintains identical functionality while reducing memory overhead
and architectural complexity, with proper string handling to prevent
mrbtest crashes.
Co-authored-by: Claude <noreply@anthropic.com>
Implement NODE_MARG as a dedicated node type for parameter destructuring
to separate it architecturally from general multiple assignment (NODE_MASGN).
This resolves crashes when dumping parameter destructuring nodes and
improves code organization.
Key changes:
- Add NODE_MARG to node type enum
- Create new_marg() function for parameter destructuring
- Consolidate new_masgn() and new_marg() using shared helper
- Fix parameter context checks in lambda_body() to use NODE_MARG only
- Enable shared dumping logic for both NODE_MASGN and NODE_MARG
- Optimize memory management with immediate RHS cleanup
- Combine gen_assignment() cases for code deduplication
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Refactored NODE_MASGN from single lhs field to separate pre/rest/post
fields for cleaner multiple assignment handling. Fixed segfault when
compiling methods with destructured parameters by properly handling
parameter destructuring in lambda_body function.
Co-authored-by: Claude <noreply@anthropic.com>
- Skip no-op splats of empty array literals (`*[]` / zarray) in
call argument generation and array literal codegen.
- Inline non-empty literal splat arrays without inner splats
(e.g. `*[a,b]`) as regular positional args/elements, avoiding
building a temporary array and ARYCAT.
This removes unnecessary `LOADNIL` + `ARRAY 0` + `ARYCAT` sequences
(e.g. `mruby -ve 'p *[]'`) and reduces temporary allocations while
preserving semantics and evaluation order. Falls back to the generic
path when nested splats are present or counts exceed fixed-arity.
No behavior change intended; only codegen improvements.
Co-authored-by: Codex <codex@openai.com>
Now that all cons-list based codegen_* functions have been removed,
rename the gen_*_var functions to use the consistent codegen_* prefix.
This affects 70 functions and improves code clarity by establishing
a single naming convention for all code generation functions.
- gen_scope_var renamed to codegen_scope_node to avoid conflict with codegen_scope type
- All other gen_*_var functions renamed to codegen_* (removing _var suffix)
- Updated all function calls throughout codegen.c
Co-authored-by: Claude <noreply@anthropic.com>
Remove NODE_MATCH enum value, structure definition, accessor macro,
parser dump case, codegen case, and gen_match_var function. This
node type was never actually used in the parser.
Co-authored-by: Claude <noreply@anthropic.com>
Remove the original NODE_REGX node type and related infrastructure,
then rename NODE_DREGX to NODE_REGX to consolidate regex handling
under a single node type.
Changes based on git diff:
- Remove original mrb_ast_regx_node structure with pattern fields
- Remove gen_regx_var() function handling literal regex patterns
- Remove NODE_REGX case from codegen and parser dump
- Rename NODE_DREGX to NODE_REGX for dynamic regex expressions
- Update all related functions and structure references
Co-authored-by: Claude <noreply@anthropic.com>
Remove the last cons list dependency in codegen.c by inlining codegen_regx()
directly into gen_regx_var(). This eliminates the need to create temporary
cons list structures and directly accesses regex pattern, flags, and
encoding from the variable-sized node structure.
Changes:
- Inline codegen_regx() logic into gen_regx_var()
- Remove codegen_regx() function entirely
- Access regex data directly from mrb_ast_regx_node fields
- Eliminate temporary cons list node creation
Co-authored-by: Claude <noreply@anthropic.com>
Remove argc, has_kwargs, has_block, and reserved fields from
mrb_ast_call_node since this information can be determined from the
callargs structure at runtime. Simplify new_call() and call_with_block()
functions to eliminate field analysis during parsing.
Add callargs_empty() helper function to check for empty arguments and
update gen_if_var() to use it instead of accessing removed argc field.
This change reduces memory usage per call node while maintaining full
functionality through runtime analysis of the callargs structure.
Co-authored-by: Claude <noreply@anthropic.com>
Eliminate code duplication in gen_string by using a single loop with
a first-element flag instead of separate first element processing.
The previous structure had ~20 lines of duplicated string literal and
expression processing logic. The refactored version uses a unified loop
that handles concatenation only for non-first elements, reducing code
duplication and improving maintainability.
Functionality remains identical - all string interpolation, regex
patterns, and heredoc processing work correctly.
Co-authored-by: Claude <noreply@anthropic.com>
Rename the overly long and poorly descriptive codegen_cons_list_string()
function to gen_string() which is more concise and follows the existing
naming convention where gen_ prefix indicates code generation functions.
This function generates string bytecode from cons-list structures
containing mixed string literals and expressions for interpolation,
used in string interpolation, regex patterns, and heredocs.
Co-authored-by: Claude <noreply@anthropic.com>
Modernize the parser dump functionality to support the post-NODE_VARIABLE
hybrid AST architecture with both variable-sized nodes and traditional
cons-list nodes.
Co-authored-by: Claude <noreply@anthropic.com>
This removes the NODE_VARIABLE enum and associated wrapper system, updating
the parser and codegen to work directly with variable-sized AST nodes.
Key changes:
- Removed NODE_VARIABLE from node.h enum
- Updated parser functions to handle direct variable-sized nodes
- Fixed codegen() main dispatch to detect variable-sized nodes directly
- Added helper functions for node type detection and header access
- Updated all parser and codegen functions to work with modern AST structure
Co-authored-by: Claude <noreply@anthropic.com>
Updates codegen() function to retrieve filename and line number information
directly from variable-sized node headers instead of assuming traditional
head nodes. Removes dead code for traditional cons-list nodes since all
nodes are now variable-sized.
Co-authored-by: Claude <noreply@anthropic.com>
Eliminates NODE_KW_HASH enum, mrb_ast_kw_hash_node struct, gen_kw_hash_var
function, and related macros. All keyword hash functionality now unified
under NODE_HASH, completing the AST simplification.
Co-authored-by: Claude <noreply@anthropic.com>
Removes the new_kw_hash function entirely and replaces all calls with
new_hash, eliminating the distinction between keyword hashes and regular
hashes in the parser. Updates codegen to handle keyword arguments directly
without intermediate cdr references.
Co-authored-by: Claude <noreply@anthropic.com>
Following the same pattern as the case node upgrade (e0f07c9), this
change eliminates the complex flat array packing approach for hash nodes
in favor of simple cons-list storage. The flat array packing provided
no memory benefit since cons lists aren't recycled, while adding
unnecessary complexity to both allocation and traversal logic.
Changes:
- Simplified mrb_ast_hash_node structure from variable-sized flexible
array to fixed-size structure with cons-list pointer
- Reduced new_hash() from complex 30+ line allocation to simple 4-line
pattern matching array node implementation
- Updated gen_hash_var() to use cons-list iteration instead of
interleaved array access (pairs[i*2] for key, pairs[i*2+1] for value)
- Removed HASH_NODE_LEN macro as length tracking is no longer needed
- Maintains identical functionality while reducing code complexity
Co-authored-by: Claude <noreply@anthropic.com>
Replace complex flat array packing with simple cons-list storage to reduce
memory overhead and code complexity. This continues the compiler simplification
work by reverting array nodes to the original memory-efficient approach.
- Remove len/flags fields from mrb_ast_array_node structure
- Eliminate complex two-pass processing (count + copy) in new_array()
- Replace array indexing with cons-list iteration in gen_array_var()
- Reduce parser code from 30+ lines to 4 lines for array creation
- Maintain full functionality with zero test regressions
Following the same successful pattern used for mrb_ast_case_node upgrade,
this change proves that flat array packing provides no memory benefit
since cons lists aren't recycled, while adding unnecessary complexity.
Co-authored-by: Claude <noreply@anthropic.com>
Replace NODE_KW_REST_ARGS wrapper nodes with direct ** symbol markers to
reduce memory overhead and simplify code structure. This continues the
compiler simplification work by unifying keyword rest arguments with
other node types while maintaining full functionality.
Co-authored-by: Claude <noreply@anthropic.com>
Replace NODE_KW_ARG wrapper with direct (key . value) cons structure,
eliminating unnecessary memory overhead and access indirection.
Changes:
- Remove NODE_KW_ARG node type from enum
- Modify new_kw_arg() to return direct cons instead of wrapped structure
- Update codegen.c to use simplified access patterns (k->car->car, k->car->cdr)
- Fix new_args_tail() to handle simplified keyword argument structure
- Remove NODE_KW_ARG case from parser dump function
This reduces memory usage from 3 cons cells to 1 per keyword argument
while maintaining full functionality and following mruby's design priority
of memory efficiency over complexity.
Co-authored-by: Claude <noreply@anthropic.com>
Replace variable-sized NODE_VARIABLE wrapper with fixed-size struct
allocation, following the same pattern as new_args(). This eliminates
the need for NODE_VARIABLE checking and uses direct casting instead.
Changes:
- Remove mrb_ast_var_header from callargs struct
- Use parser_palloc instead of parser_alloc_var for fixed-size allocation
- Update all access points to use direct casting: (struct mrb_ast_callargs*)
- Remove unnecessary backward compatibility code for newly introduced NODE_CALLARGS
Co-authored-by: Claude <noreply@anthropic.com>
Rename mrb_ast_op_asgn_node.operator field to op to avoid conflict with
C++ operator keyword. Update all references including macro definitions
and field access code.
Co-authored-by: Claude <noreply@anthropic.com>
- Remove obsolete NODE_ARGS_TAIL enum value and all references
- Simplify mrb_ast_case_node from variable-sized array back to simple cons-list structure
- Update new_case() function to use original cons-list approach instead of flattening
- Fix infinite loop in gen_case_var() when case statements have no matching clauses
- Improve code readability by renaming pos3 to case_end_jumps in gen_case_var()
- Restore memory-efficient case statement parsing without complex array management
The variable-sized array approach for case nodes provided no memory benefit
since cons lists aren't recycled. This change restores the simpler original
implementation while fixing a critical bug that caused mrbtest to hang
on "register window of calls" test.
Co-authored-by: Claude <noreply@anthropic.com>
NODE_ARG and NODE_LVAR were handled identically in codegen.c, making
the distinction unnecessary. This change:
- Replaces all new_arg() calls with new_xvar(p, sym, NODE_LVAR)
- Removes the new_arg() function entirely
- Removes the unused NODE_ARG enum value
- Updates codegen.c to handle only NODE_LVAR case
The simplification reduces parser complexity while maintaining identical
functionality for argument processing.
Co-authored-by: Claude <noreply@anthropic.com>
This commit completes the transformation of mruby's argument processing from
cons-list based representation to direct struct field access.
Key changes:
- Transform new_args() to return struct mrb_ast_args* instead of cons-list
- Update lambda_body() to use direct struct field access for all argument types
- Fix anonymous keyword rest (**) to use intern_op(pow) marker for proper bytecode generation
- Fix argument forwarding (...) to correctly pass rest_arg to new_args()
- Eliminate mrb_ast_args_tail_node allocation by embedding fields directly in mrb_ast_args
- Update all node structure definitions to use struct mrb_ast_args*
- Remove unused NODE_ARGS enum value since args are now plain C structs
The new approach provides:
- More efficient memory usage by eliminating intermediate cons-list allocations
- Cleaner code generation with direct struct field access
- Proper distinction between anonymous kwrest and no kwrest
- Correct bytecode generation for both anonymous kwrest and argument forwarding
Fixes both anonymous keyword rest (def m(**) end) and argument forwarding
(def a(...) p(...) end) to generate correct bytecode and execute properly.
Co-authored-by: Claude <noreply@anthropic.com>
Inline the 320-line codegen_variable_node() function directly into the
codegen() function to eliminate function call overhead for every
variable-sized AST node processing.
Co-authored-by: Claude <noreply@anthropic.com>
Remove NODE_SCALL and NODE_FCALL node types, consolidating all method calls
into a single NODE_CALL variable-sized node structure. This simplifies the
AST by unifying call semantics while maintaining support for safe navigation
and different call types through node flags.
Key changes:
- Convert call nodes to use variable-sized allocation with call_node structure
- Unify new_call() and new_fcall() to create NODE_CALL nodes consistently
- Replace gen_call() with separate gen_call_var() and gen_call_assign_var()
- Add gen_call_assign_var() for assignment operations like h[k] = v
- Remove legacy call handling from main codegen switch statement
- Preserve argument structure using args pointer instead of unpacking
- Support safe calls, keyword arguments, and blocks in unified structure
This migration maintains backward compatibility while enabling more efficient
call node processing and reduced code duplication in the compiler.
Co-authored-by: Claude <noreply@anthropic.com>
Complete migration of method definition nodes to variable-sized format:
- Convert NODE_DEF and NODE_SDEF from fixed cons-based to variable-sized nodes
- Update parser to create variable-sized def/sdef nodes directly
- Remove old codegen_def and codegen_sdef functions
- Consolidate method setup logic in defn_setup function
- Rename lambda_body_ex to lambda_body after removing wrapper layer
- Update all method definition code generation to use new node structure
This completes the variable-sized node migration for method definitions,
improving memory efficiency and enabling more flexible AST handling.
Co-authored-by: Claude <noreply@anthropic.com>
Clean up function names by removing unnecessary _var suffixes for
consistency with other assignment functions.
Co-authored-by: Claude <noreply@anthropic.com>
created lambda_body_ex that takes locals, args, and body as separate
parameters instead of a cons structure. this eliminates complex cons
cell navigation and makes the interface cleaner for variable-sized
nodes. updated all call sites (gen_def_var, gen_sdef_var, gen_block_var,
gen_lambda_var) to use lambda_body_ex directly.
Co-authored-by: Claude <noreply@anthropic.com>
Convert NODE_LVAR and NODE_ARG from cons-list to variable-sized nodes.
Remove new_lvar wrapper and use new_xvar directly. Update parser
semantic functions and cleanup gen_assignment. Rename codegen_lvar
to gen_lvar for consistency.
Co-authored-by: Claude <noreply@anthropic.com>
Remove conditional logic from new_op_asgn() that created traditional
cons-list nodes when var_nodes_enabled was false. Now always creates
variable-sized nodes using struct mrb_ast_op_asgn_node.
Also remove traditional NODE_OP_ASGN codegen path and unused
codegen_op_asgn function, completing the migration to variable-sized
nodes for all operator assignment patterns.
Co-authored-by: Claude <noreply@anthropic.com>
Remove conditional logic from new_asgn() and inline new_asgn_var() helper
function for cleaner implementation. Assignment expressions maintain proper
value semantics while using more efficient memory allocation.
Changes:
- Remove var_nodes_enabled conditional in new_asgn()
- Inline new_asgn_var() logic directly into new_asgn()
- Remove new_asgn_var() function and declaration
- Remove NODE_ASGN case from main codegen() switch
- Update gen_asgn_var() to use direct struct field access
- Remove traditional codegen_asgn() function
Co-authored-by: Claude <noreply@anthropic.com>
Update codegen_op_asgn() to handle variable-sized nodes wrapped in NODE_VARIABLE
instead of assuming traditional cons-list format. Remove obsolete traditional
node type checks since NODE_CONST and NODE_CVAR now always use variable-sized
nodes.
The ||= operator generates special exception-handling bytecode for undefined
constant/class variable detection that requires checking the node type to
apply proper optimization.
Co-authored-by: Claude <noreply@anthropic.com>
Remove conditional logic from new_const() and inline new_const_var() helper
function for cleaner implementation. Update codegen to handle NODE_CONST
in both variable-sized access and assignment contexts.
Changes:
- Remove var_nodes_enabled conditional in new_const()
- Inline new_const_var() logic directly into new_const()
- Remove new_const_var() function and declaration
- Remove NODE_CONST case from main codegen() switch
- Add NODE_CONST support in gen_assignment() for variable-sized nodes
- Inline codegen_const() logic into gen_const_var()
- Remove traditional codegen_const() function
Co-authored-by: Claude <noreply@anthropic.com>
Remove var_nodes_enabled conditional from new_nvar() and traditional
NODE_NVAR case from codegen. All numbered parameter operations now use
unified variable-sized node handling.
Co-authored-by: Claude <noreply@anthropic.com>
Remove remaining NODE_GVAR/IVAR/CVAR cases from codegen switch statements
since parser now always creates variable-sized nodes. All variable operations
now use unified NODE_VARIABLE handling.
Co-authored-by: Claude <noreply@anthropic.com>
Remove traditional NODE_BLOCK_ARG support from main codegen() switch and
eliminate synthetic node creation in gen_block_arg_var. The function now
handles the variable-sized node directly without creating temporary
traditional nodes on the stack.
This completes the NODE_BLOCK_ARG migration by removing the dual handling
pattern while maintaining the gen_block_arg_var function for better code
organization and readability.
Co-authored-by: Claude <noreply@anthropic.com>
remove NODE_POSTEXE from main codegen function and inline gen_postexe_var
into codegen_variable_node. remove unused codegen_postexe function since
NODE_POSTEXE is now only a marker like NODE_ARGS
Co-authored-by: Claude <noreply@anthropic.com>
Yukihiro "Matz" Matsumoto