tomoaki/hakorune
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor(joinir): Phase 189 - Remove hardcoded 'sum' variable, generalize exit PHI connection

Browse Source 
Key improvements:
1. **Eliminate hardcoded variable name**: Replace hardcoded "sum" with generic
   ValueId-based variable_map updates. Add new JoinInlineBoundary constructor
   `new_with_input_and_host_outputs()` for Pattern 3.

2. **Generalize output slot mapping**: Exit PHI result now updates all host_outputs
   entries in variable_map, not just hardcoded "sum". Prepares for future
   multi-carrier patterns.

3. **PHI preservation in blocks**: Fix block finalization to preserve existing
   PHI instructions (from handle_select) instead of overwriting them.

4. **Stable function name→ValueId mapping**: Ensure consistent ValueId
   assignment for tail call detection across multi-function merges.

5. **Enhanced debugging**: Add detailed logging in block converter and meta
   analysis for PHI verification.

Files modified:
- src/mir/builder/control_flow.rs: Remove hardcoded "sum", use boundary outputs
- src/mir/join_ir/lowering/inline_boundary.rs: Add new constructor
- src/mir/join_ir_vm_bridge/joinir_block_converter.rs: Stable mappings, PHI preservation
- src/mir/join_ir_vm_bridge/meta.rs: Debug output for PHI tracking
- src/mir/builder/joinir_id_remapper.rs: PHI value remapping
- src/mir/builder/joinir_inline_boundary_injector.rs: Span preservation

Test status: Pattern 3 (loop_if_phi.hako) still produces correct result (sum=9, RC=9)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: ChatGPT <noreply@openai.com>
Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
nyash-codex
2025-12-05 19:39:54 +09:00
parent e01d1b8a7c
commit 0397df5240
9 changed files with 202 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/mir/optimizer_passes/boxfield.rs
View File

@ -1,3 +1,4 @@
use crate::ast::Span;
use crate::mir::optimizer::MirOptimizer;
use crate::mir::optimizer_stats::OptimizationStats;
use crate::mir::{MirInstruction as I, MirModule};
@ -15,10 +16,15 @@ pub fn optimize_boxfield_operations(
for (_bb_id, block) in &mut function.blocks {
let mut changed = 0usize;
let mut out: Vec<I> = Vec::with_capacity(block.instructions.len());
let mut out_spans: Vec<Span> = Vec::with_capacity(block.instruction_spans.len());
let mut i = 0usize;
while i < block.instructions.len() {
// Get span for current instruction (or unknown if missing)
let span_i = block.instruction_spans.get(i).copied().unwrap_or_else(Span::unknown);
// Look ahead for simple store-followed-by-load on same box/index
if i + 1 < block.instructions.len() {
let span_i1 = block.instruction_spans.get(i + 1).copied().unwrap_or_else(Span::unknown);
match (&block.instructions[i], &block.instructions[i + 1]) {
(
I::BoxCall {
@ -42,10 +48,12 @@ pub fn optimize_boxfield_operations(
// Replace the second with Copy from just-stored value
let src_val = a1[1];
out.push(block.instructions[i].clone());
out_spans.push(span_i);
out.push(I::Copy {
dst: *dst2,
src: src_val,
});
out_spans.push(span_i1);
changed += 1;
i += 2;
continue;
@ -55,10 +63,12 @@ pub fn optimize_boxfield_operations(
}
}
out.push(block.instructions[i].clone());
out_spans.push(span_i);
i += 1;
}
if changed > 0 {
block.instructions = out;
block.instruction_spans = out_spans;
stats.boxfield_optimizations += changed;
}
}