Remove unused typed local opcodes from compiler and VM by frostney · Pull Request #74 · frostney/GocciaScript

frostney · 2026-03-12T14:42:38Z

Summary

Removes 10 typed local opcodes (OP_GET_LOCAL_INT through OP_SET_LOCAL_REF) that had identical implementations to OP_GET_LOCAL/OP_SET_LOCAL — plain register copies with no type-specific behavior.
The optimization decision (choosing OP_ADD_FLOAT vs OP_RT_ADD) comes from ExpressionType at compile time, not from which load opcode was used.
Simplifies the VM dispatch surface (10 fewer case arms in both ExecuteLoop and ExecuteCoreOp) and removes the TypedGetLocalOp/TypedSetLocalOp compiler helpers.
Opcode enum entries preserved in Souffle.Bytecode.pas for .sbc backward compatibility.

Test plan

All JS tests pass in interpreter mode
All JS tests pass in bytecode mode
CI benchmark comparison will confirm no performance change

Made with Cursor

Summary by CodeRabbit

Refactor
- Unified all type-specific local variable loads/stores into single generic local get/set operations across VM, translator, and compiler, removing the specialized variants while preserving behavior.
Chores
- Bumped bytecode/format version to reflect the instruction set change.
Documentation
- Updated docs to describe the unified local access model and where type-specialization now occurs.

The 10 typed local opcodes (OP_GET_LOCAL_INT..OP_SET_LOCAL_REF) had identical implementations to OP_GET_LOCAL/OP_SET_LOCAL — plain register copies with no type-specific behavior. The optimization decision (OP_ADD_FLOAT vs OP_RT_ADD) comes from ExpressionType at compile time, not from the load opcode. Removing these simplifies the dispatch surface without changing any optimization behavior. Changes: - Remove TypedGetLocalOp/TypedSetLocalOp helpers from compiler - Compiler now always emits OP_GET_LOCAL/OP_SET_LOCAL - Remove typed local case arms from VM ExecuteLoop and ExecuteCoreOp - Remove typed local entries from WASM translator - Opcode enum entries preserved for .sbc backward compatibility Made-with: Cursor

coderabbitai · 2026-03-12T14:42:59Z

📝 Walkthrough

Walkthrough

Removes all type-specific local load/store opcodes and related helpers; compiler, Wasm translator, VM, and docs now use only generic OP_GET_LOCAL / OP_SET_LOCAL. Bytecode format version is bumped to v4 to reflect the opcode set change.

Changes

Cohort / File(s)	Summary
VM Execution Engine `souffle/Souffle.VM.pas`	Deleted case handlers for typed local ops (`OP_GET_LOCAL_`, `OP_SET_LOCAL_`) from ExecuteLoop / ExecuteCoreOp; VM now handles only generic `OP_GET_LOCAL` / `OP_SET_LOCAL`.
Bytecode / Opcode defs `souffle/Souffle.Bytecode.pas`	Bumped `SOUFFLE_FORMAT_VERSION` to 4 and removed the ten typed-local opcodes from `TSouffleOpCode` (creates a gap/renumbering in prior opcode space).
WASM Translation `souffle/Souffle.Wasm.Translator.pas`	Consolidated multiple per-type opcode branches into single `OP_GET_LOCAL` / `OP_SET_LOCAL` emission paths in `EmitInstruction`.
Compiler Expressions `units/Goccia.Compiler.Expressions.pas`	Removed `TypedSetLocalOp` and related typed-get/set selection; `CompileIdentifierAccess` and `CompileAssignment` now emit generic local get/set unconditionally.
Documentation `docs/souffle-vm.md`	Removed typed-local opcodes from Tier 1 list; updated explanation to state loads/stores are generic and optimizations rely on typed arithmetic/compare opcodes.

Sequence Diagram(s)

sequenceDiagram
  participant Compiler as Compiler
  participant Translator as WasmTranslator
  participant VM as VM
  Compiler->>Translator: emit OP_GET_LOCAL / OP_SET_LOCAL for a local slot
  Translator->>VM: translate to wasm local.get / local.set (or bytecode stream)
  VM->>VM: Execute generic OP_GET_LOCAL / OP_SET_LOCAL\n(access slot without per-type branch)

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~45 minutes

Possibly related PRs

Strict types in bytecode VM #42: Adds/uses typed local get/set opcodes and helpers that this PR removes or reverses—strongly related at the opcode/compiler level.
Bytecode endianness portability #72: Previously bumped SOUFFLE_FORMAT_VERSION and altered bytecode handling; related due to bytecode format/version changes.
Feature parity between bytecode mode and interpreter mode #39: Earlier work that introduced typed-local get/set opcodes which this PR reverts or modifies.

Poem

🐇
I hopped through ops and left a trail,
Five small paths made into one pale.
One gentle hop, a simpler tune,
Moonlit bytes beneath the rune.
Nibble, nest — the VM's new noon.

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The pull request title accurately and concisely describes the main change: removal of typed local opcodes (OP_GET_LOCAL_INT, OP_SET_LOCAL_INT, etc.) from the compiler and VM, which is the core objective across all modified files.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests
Post copyable unit tests in a comment
Commit unit tests in branch opt/remove-typed-locals

📝 Coding Plan

Generate coding plan for human review comments

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

coderabbitai

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

souffle/Souffle.Wasm.Translator.pas (1)

638-647: ⚠️ Potential issue | 🔴 Critical

Keep the legacy typed-local opcodes mapped here.

Souffle.Bytecode.pas:101-111 still declares OP_GET_LOCAL_INT through OP_SET_LOCAL_REF for old .sbc compatibility. After this change, those legacy opcodes are no longer recognized by the Wasm translator and will fall into the default EmitUnreachable path, so translating persisted bytecode that predates this PR will break.

🔧 Minimal fix

-    OP_GET_LOCAL:
+    OP_GET_LOCAL,
+    OP_GET_LOCAL_INT,
+    OP_GET_LOCAL_FLOAT,
+    OP_GET_LOCAL_BOOL,
+    OP_GET_LOCAL_STRING,
+    OP_GET_LOCAL_REF:
     begin
       ABuilder.EmitLocalGet(Bx);
       ABuilder.EmitLocalSet(A);
     end;
-    OP_SET_LOCAL:
+    OP_SET_LOCAL,
+    OP_SET_LOCAL_INT,
+    OP_SET_LOCAL_FLOAT,
+    OP_SET_LOCAL_BOOL,
+    OP_SET_LOCAL_STRING,
+    OP_SET_LOCAL_REF:
     begin
       ABuilder.EmitLocalGet(A);
       ABuilder.EmitLocalSet(Bx);
     end;

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@souffle/Souffle.Wasm.Translator.pas` around lines 638 - 647, The Wasm
translator removed handling for legacy typed-local opcodes (e.g.,
OP_GET_LOCAL_INT .. OP_SET_LOCAL_REF) so older .sbc files will fall through to
EmitUnreachable; restore support by adding case branches mapping each legacy
opcode to the same actions as the generic handlers (use
EmitLocalGet/EmitLocalSet with the appropriate operands), i.e., implement cases
for OP_GET_LOCAL_INT, OP_GET_LOCAL_REF, OP_SET_LOCAL_INT, OP_SET_LOCAL_REF (and
any other legacy OP_GET/OP_SET variants declared in Souffle.Bytecode.pas) that
call ABuilder.EmitLocalGet(...) and ABuilder.EmitLocalSet(...) the same way
OP_GET_LOCAL and OP_SET_LOCAL do so legacy bytecode is recognized.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Outside diff comments:
In `@souffle/Souffle.Wasm.Translator.pas`:
- Around line 638-647: The Wasm translator removed handling for legacy
typed-local opcodes (e.g., OP_GET_LOCAL_INT .. OP_SET_LOCAL_REF) so older .sbc
files will fall through to EmitUnreachable; restore support by adding case
branches mapping each legacy opcode to the same actions as the generic handlers
(use EmitLocalGet/EmitLocalSet with the appropriate operands), i.e., implement
cases for OP_GET_LOCAL_INT, OP_GET_LOCAL_REF, OP_SET_LOCAL_INT, OP_SET_LOCAL_REF
(and any other legacy OP_GET/OP_SET variants declared in Souffle.Bytecode.pas)
that call ABuilder.EmitLocalGet(...) and ABuilder.EmitLocalSet(...) the same way
OP_GET_LOCAL and OP_SET_LOCAL do so legacy bytecode is recognized.

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 0597decc-fbe2-4459-a801-951d00309e7b

📥 Commits

Reviewing files that changed from the base of the PR and between aaa31fe and 76a85e6.

📒 Files selected for processing (3)

souffle/Souffle.VM.pas
souffle/Souffle.Wasm.Translator.pas
units/Goccia.Compiler.Expressions.pas

💤 Files with no reviewable changes (1)

souffle/Souffle.VM.pas

frostney · 2026-03-12T14:48:24Z

TODO: Remove opcode enums entirely - Backwards compatibility is not a priority at the moment and we should up the SBC version number

github-actions · 2026-03-12T14:56:00Z

Suite Timing

Suite	Metric	Interpreted	Bytecode
Tests	Total	3463	3463
Tests	Passed	3422 ✅	3463 ✅
Tests	Skipped	41	0
Tests	Execution	149.7ms	167.9ms
Tests	Engine	306.1ms	683.5ms
Benchmarks	Total	254	254
Benchmarks	Duration	6.66min	10.58min

_{Measured on ubuntu-latest x64.}

github-actions · 2026-03-12T14:56:01Z

Benchmark Results

254 benchmarks

Interpreted: 🟢 46 improved · 🔴 1 regressed · 207 unchanged · avg +4.7%
Bytecode: 🟢 8 improved · 🔴 2 regressed · 244 unchanged · avg +1.1%

arraybuffer.js — Interp: 🟢 2, 12 unch. · avg +3.9% · Bytecode: 14 unch. · avg -2.1%

Benchmark	Interpreted	Δ	Bytecode	Δ
create ArrayBuffer(0)	421,967 → 444,591	+5.4%	146,658 → 144,793	-1.3%
create ArrayBuffer(64)	412,708 → 433,810	+5.1%	142,667 → 142,866	+0.1%
create ArrayBuffer(1024)	317,332 → 330,389	+4.1%	130,112 → 128,653	-1.1%
create ArrayBuffer(8192)	138,729 → 139,907	+0.8%	81,593 → 80,361	-1.5%
slice full buffer (64 bytes)	507,944 → 538,328	+6.0%	384,102 → 365,311	-4.9%
slice half buffer (512 of 1024 bytes)	430,211 → 460,898	🟢 +7.1%	338,592 → 323,968	-4.3%
slice with negative indices	427,289 → 468,260	🟢 +9.6%	364,038 → 350,823	-3.6%
slice empty range	498,101 → 532,289	+6.9%	377,390 → 364,373	-3.4%
byteLength access	1,596,224 → 1,587,867	-0.5%	1,055,008 → 1,036,363	-1.8%
Symbol.toStringTag access	1,192,516 → 1,188,569	-0.3%	545,389 → 540,542	-0.9%
ArrayBuffer.isView	747,110 → 757,938	+1.4%	469,688 → 450,990	-4.0%
clone ArrayBuffer(64)	374,847 → 386,542	+3.1%	318,287 → 311,551	-2.1%
clone ArrayBuffer(1024)	291,518 → 301,207	+3.3%	246,636 → 243,975	-1.1%
clone ArrayBuffer inside object	254,935 → 263,230	+3.3%	156,135 → 155,957	-0.1%

arrays.js — Interp: 🟢 4, 15 unch. · avg +3.3% · Bytecode: 19 unch. · avg +1.2%

Benchmark	Interpreted	Δ	Bytecode	Δ
Array.from length 100	14,300 → 14,294	-0.0%	12,842 → 13,137	+2.3%
Array.from 10 elements	220,145 → 237,882	🟢 +8.1%	152,923 → 156,275	+2.2%
Array.of 10 elements	301,512 → 311,977	+3.5%	224,569 → 224,609	+0.0%
spread into new array	328,496 → 353,136	🟢 +7.5%	570,753 → 569,338	-0.2%
map over 50 elements	28,352 → 28,569	+0.8%	22,135 → 22,369	+1.1%
filter over 50 elements	24,269 → 24,438	+0.7%	21,416 → 21,815	+1.9%
reduce sum 50 elements	27,488 → 27,704	+0.8%	19,284 → 19,838	+2.9%
forEach over 50 elements	23,439 → 23,365	-0.3%	23,510 → 23,724	+0.9%
find in 50 elements	35,313 → 35,632	+0.9%	28,036 → 28,445	+1.5%
sort 20 elements	12,140 → 12,514	+3.1%	3,177 → 3,190	+0.4%
flat nested array	108,188 → 122,163	🟢 +12.9%	313,437 → 311,626	-0.6%
flatMap	70,076 → 75,021	🟢 +7.1%	211,685 → 214,567	+1.4%
map inside map (5x5)	20,883 → 22,007	+5.4%	69,693 → 70,007	+0.4%
filter inside map (5x10)	15,877 → 16,371	+3.1%	13,953 → 14,093	+1.0%
reduce inside map (5x10)	19,437 → 19,855	+2.2%	14,014 → 14,176	+1.2%
forEach inside forEach (5x10)	16,597 → 16,718	+0.7%	15,323 → 15,744	+2.8%
find inside some (10x10)	13,895 → 14,070	+1.3%	10,950 → 11,148	+1.8%
map+filter chain nested (5x20)	5,490 → 5,530	+0.7%	4,509 → 4,595	+1.9%
reduce flatten (10x5)	37,914 → 39,618	+4.5%	6,085 → 6,092	+0.1%

async-await.js — Interp: 🟢 3, 3 unch. · avg +6.5% · Bytecode: 6 unch. · avg +0.7%

Benchmark	Interpreted	Δ	Bytecode	Δ
single await	365,057 → 383,655	+5.1%	277,944 → 277,833	-0.0%
multiple awaits	161,054 → 172,812	🟢 +7.3%	119,624 → 119,866	+0.2%
await non-Promise value	831,676 → 876,153	+5.3%	901,554 → 933,453	+3.5%
await with try/catch	356,047 → 372,969	+4.8%	271,580 → 270,621	-0.4%
await Promise.all	48,516 → 52,797	🟢 +8.8%	43,669 → 44,185	+1.2%
nested async function call	179,423 → 192,872	🟢 +7.5%	207,981 → 207,761	-0.1%

classes.js — Interp: 31 unch. · avg +3.8% · Bytecode: 31 unch. · avg +0.5%

Benchmark	Interpreted	Δ	Bytecode	Δ
simple class new	114,658 → 117,532	+2.5%	338,222 → 343,668	+1.6%
class with defaults	91,529 → 95,029	+3.8%	239,266 → 241,228	+0.8%
50 instances via Array.from	5,440 → 5,553	+2.1%	6,384 → 6,413	+0.5%
instance method call	58,082 → 59,380	+2.2%	153,390 → 152,804	-0.4%
static method call	90,225 → 93,139	+3.2%	359,870 → 353,445	-1.8%
single-level inheritance	45,558 → 46,425	+1.9%	152,229 → 149,949	-1.5%
two-level inheritance	38,540 → 39,399	+2.2%	130,358 → 129,590	-0.6%
private field access	57,580 → 59,262	+2.9%	169,516 → 167,187	-1.4%
private methods	62,574 → 64,551	+3.2%	217,406 → 214,511	-1.3%
getter/setter access	64,616 → 66,387	+2.7%	169,354 → 167,464	-1.1%
class decorator (identity)	79,195 → 82,869	+4.6%	54,083 → 52,233	-3.4%
class decorator (wrapping)	45,868 → 47,569	+3.7%	38,099 → 38,547	+1.2%
identity method decorator	56,426 → 59,071	+4.7%	44,999 → 46,257	+2.8%
wrapping method decorator	46,369 → 48,712	+5.1%	38,834 → 38,847	+0.0%
stacked method decorators (x3)	32,389 → 34,380	+6.1%	28,098 → 28,635	+1.9%
identity field decorator	63,344 → 67,150	+6.0%	46,092 → 47,001	+2.0%
field initializer decorator	53,935 → 56,545	+4.8%	41,788 → 42,128	+0.8%
getter decorator (identity)	55,358 → 57,420	+3.7%	40,699 → 41,444	+1.8%
setter decorator (identity)	46,704 → 48,484	+3.8%	35,093 → 36,049	+2.7%
static method decorator	60,399 → 63,150	+4.6%	62,959 → 63,756	+1.3%
static field decorator	69,683 → 73,352	+5.3%	66,959 → 67,018	+0.1%
private method decorator	46,139 → 48,499	+5.1%	37,455 → 38,165	+1.9%
private field decorator	51,511 → 52,905	+2.7%	38,468 → 39,249	+2.0%
plain auto-accessor (no decorator)	86,783 → 91,039	+4.9%	54,192 → 54,787	+1.1%
auto-accessor with decorator	50,330 → 53,095	+5.5%	37,342 → 38,170	+2.2%
decorator writing metadata	41,537 → 43,971	+5.9%	42,152 → 43,169	+2.4%
static getter read	101,736 → 105,054	+3.3%	383,144 → 380,510	-0.7%
static getter/setter pair	77,786 → 80,060	+2.9%	208,226 → 207,423	-0.4%
inherited static getter	58,866 → 60,233	+2.3%	262,035 → 262,683	+0.2%
inherited static setter	63,650 → 65,135	+2.3%	203,868 → 204,849	+0.5%
inherited static getter with this binding	53,213 → 54,490	+2.4%	154,484 → 154,221	-0.2%

closures.js — Interp: 🟢 1, 10 unch. · avg +4.7% · Bytecode: 11 unch. · avg +0.5%

Benchmark	Interpreted	Δ	Bytecode	Δ
closure over single variable	123,589 → 131,541	+6.4%	589,065 → 596,786	+1.3%
closure over multiple variables	117,116 → 121,224	+3.5%	389,441 → 392,828	+0.9%
nested closures	121,010 → 128,581	+6.3%	554,552 → 562,213	+1.4%
function as argument	94,190 → 100,071	+6.2%	491,152 → 498,167	+1.4%
function returning function	117,427 → 122,646	+4.4%	588,813 → 600,365	+2.0%
compose two functions	69,896 → 73,501	+5.2%	353,239 → 357,719	+1.3%
fn.call	147,425 → 154,556	+4.8%	137,278 → 137,802	+0.4%
fn.apply	110,315 → 115,461	+4.7%	94,462 → 94,159	-0.3%
fn.bind	132,232 → 141,815	🟢 +7.2%	143,039 → 144,080	+0.7%
recursive sum to 50	11,978 → 12,169	+1.6%	38,927 → 37,563	-3.5%
recursive tree traversal	19,458 → 19,758	+1.5%	58,306 → 58,562	+0.4%

collections.js — Interp: 🟢 2, 10 unch. · avg +2.7% · Bytecode: 12 unch. · avg +1.4%

Benchmark	Interpreted	Δ	Bytecode	Δ
add 50 elements	7,138 → 7,212	+1.0%	5,672 → 5,788	+2.0%
has lookup (50 elements)	89,465 → 89,165	-0.3%	86,313 → 86,656	+0.4%
delete elements	47,513 → 48,297	+1.6%	36,783 → 36,801	+0.0%
forEach iteration	15,686 → 16,214	+3.4%	15,325 → 15,567	+1.6%
spread to array	30,804 → 33,417	🟢 +8.5%	145,515 → 147,574	+1.4%
deduplicate array	40,910 → 42,874	+4.8%	46,658 → 47,415	+1.6%
set 50 entries	5,313 → 5,403	+1.7%	5,789 → 5,772	-0.3%
get lookup (50 entries)	86,692 → 86,393	-0.3%	95,800 → 97,121	+1.4%
has check	129,653 → 128,773	-0.7%	150,478 → 154,162	+2.4%
delete entries	45,408 → 45,760	+0.8%	35,284 → 35,842	+1.6%
forEach iteration	15,502 → 16,265	+4.9%	15,632 → 15,908	+1.8%
keys/values/entries	8,178 → 8,755	🟢 +7.1%	22,373 → 22,890	+2.3%

destructuring.js — Interp: 🟢 10, 12 unch. · avg +6.6% · Bytecode: 22 unch. · avg +1.7%

Benchmark	Interpreted	Δ	Bytecode	Δ
simple array destructuring	417,360 → 434,943	+4.2%	608,533 → 642,663	+5.6%
with rest element	272,645 → 310,199	🟢 +13.8%	480,009 → 483,493	+0.7%
with defaults	421,870 → 434,285	+2.9%	658,700 → 684,238	+3.9%
skip elements	424,552 → 469,026	🟢 +10.5%	705,349 → 722,035	+2.4%
nested array destructuring	173,301 → 189,969	🟢 +9.6%	333,888 → 355,449	+6.5%
swap variables	509,172 → 539,246	+5.9%	990,612 → 1,004,117	+1.4%
simple object destructuring	296,906 → 305,234	+2.8%	504,260 → 529,941	+5.1%
with defaults	355,435 → 372,706	+4.9%	316,614 → 322,273	+1.8%
with renaming	313,351 → 330,594	+5.5%	587,366 → 605,572	+3.1%
nested object destructuring	146,887 → 157,940	🟢 +7.5%	248,565 → 249,173	+0.2%
rest properties	183,921 → 198,324	🟢 +7.8%	225,098 → 223,127	-0.9%
object parameter	91,186 → 95,575	+4.8%	183,291 → 192,437	+5.0%
array parameter	121,036 → 129,371	+6.9%	311,744 → 307,817	-1.3%
mixed destructuring in map	34,535 → 35,798	+3.7%	33,901 → 34,846	+2.8%
forEach with array destructuring	64,945 → 69,777	🟢 +7.4%	138,059 → 136,815	-0.9%
map with array destructuring	65,513 → 71,157	🟢 +8.6%	164,348 → 167,222	+1.7%
filter with array destructuring	68,233 → 74,274	🟢 +8.9%	197,837 → 198,996	+0.6%
reduce with array destructuring	73,697 → 79,316	🟢 +7.6%	183,977 → 181,533	-1.3%
map with object destructuring	78,706 → 81,193	+3.2%	75,319 → 75,525	+0.3%
map with nested destructuring	64,380 → 67,562	+4.9%	62,124 → 61,523	-1.0%
map with rest in destructuring	36,243 → 39,894	🟢 +10.1%	22,437 → 22,443	+0.0%
map with defaults in destructuring	58,054 → 59,984	+3.3%	37,982 → 38,302	+0.8%

fibonacci.js — Interp: 8 unch. · avg +3.6% · Bytecode: 🟢 1, 7 unch. · avg +2.6%

Benchmark	Interpreted	Δ	Bytecode	Δ
recursive fib(15)	327 → 344	+5.2%	1,135 → 1,136	+0.1%
recursive fib(20)	30 → 31	+2.8%	102 → 102	-0.0%
recursive fib(15) typed	335 → 339	+1.3%	1,432 → 1,435	+0.2%
recursive fib(20) typed	30 → 31	+2.2%	127 → 129	+1.8%
iterative fib(20) via reduce	11,971 → 12,238	+2.2%	8,249 → 8,883	🟢 +7.7%
iterator fib(20)	9,109 → 9,625	+5.7%	14,293 → 14,943	+4.5%
iterator fib(20) via Iterator.from + take	14,240 → 14,939	+4.9%	15,931 → 16,799	+5.4%
iterator fib(20) last value via reduce	10,990 → 11,494	+4.6%	12,412 → 12,572	+1.3%

for-of.js — Interp: 7 unch. · avg +4.5% · Bytecode: 7 unch. · avg -1.1%

Benchmark	Interpreted	Δ	Bytecode	Δ
for...of with 10-element array	46,637 → 48,291	+3.5%	151,702 → 147,694	-2.6%
for...of with 100-element array	5,421 → 5,549	+2.3%	21,572 → 20,946	-2.9%
for...of with string (10 chars)	33,808 → 35,885	+6.1%	118,340 → 117,194	-1.0%
for...of with Set (10 elements)	46,767 → 49,489	+5.8%	162,415 → 154,768	-4.7%
for...of with Map entries (10 entries)	29,648 → 31,431	+6.0%	45,352 → 48,237	+6.4%
for...of with destructuring	39,815 → 41,696	+4.7%	69,941 → 70,309	+0.5%
for-await-of with sync array	44,612 → 45,936	+3.0%	128,666 → 124,184	-3.5%

iterators.js — Interp: 🟢 7, 13 unch. · avg +6.8% · Bytecode: 20 unch. · avg +2.2%

Benchmark	Interpreted	Δ	Bytecode	Δ
Iterator.from({next}).toArray() — 20 elements	14,380 → 14,903	+3.6%	16,533 → 16,624	+0.6%
Iterator.from({next}).toArray() — 50 elements	6,288 → 6,557	+4.3%	7,139 → 7,416	+3.9%
spread pre-wrapped iterator — 20 elements	10,821 → 11,745	🟢 +8.5%	15,311 → 16,014	+4.6%
Iterator.from({next}).forEach — 50 elements	4,363 → 4,604	+5.5%	5,087 → 5,326	+4.7%
Iterator.from({next}).reduce — 50 elements	4,330 → 4,687	🟢 +8.2%	4,985 → 5,108	+2.5%
wrap array iterator	163,259 → 183,983	🟢 +12.7%	105,774 → 108,595	+2.7%
wrap plain {next()} object	9,798 → 10,576	🟢 +7.9%	12,432 → 12,982	+4.4%
map + toArray (50 elements)	4,444 → 4,823	🟢 +8.5%	5,755 → 5,803	+0.8%
filter + toArray (50 elements)	4,406 → 4,673	+6.1%	5,226 → 5,583	+6.8%
take(10) + toArray (50 element source)	26,078 → 27,577	+5.7%	27,947 → 29,144	+4.3%
drop(40) + toArray (50 element source)	6,260 → 6,612	+5.6%	7,883 → 8,401	+6.6%
chained map + filter + take (100 element source)	8,243 → 8,734	+6.0%	9,382 → 9,642	+2.8%
some + every (50 elements)	2,589 → 2,713	+4.8%	3,358 → 3,346	-0.3%
find (50 elements)	5,654 → 5,948	+5.2%	6,872 → 6,806	-1.0%
array.values().map().filter().toArray()	8,792 → 9,401	+6.9%	9,637 → 9,603	-0.4%
array.values().take(5).toArray()	202,686 → 223,316	🟢 +10.2%	155,073 → 145,468	-6.2%
array.values().drop(45).toArray()	189,768 → 208,947	🟢 +10.1%	141,128 → 146,926	+4.1%
map.entries() chained helpers	10,519 → 11,215	+6.6%	5,262 → 5,322	+1.1%
set.values() chained helpers	18,509 → 19,297	+4.3%	20,166 → 20,760	+2.9%
string iterator map + toArray	13,379 → 14,019	+4.8%	21,524 → 21,530	+0.0%

json.js — Interp: 🟢 1, 19 unch. · avg +3.7% · Bytecode: 🔴 1, 19 unch. · avg -0.0%

Benchmark	Interpreted	Δ	Bytecode	Δ
parse simple object	165,485 → 171,930	+3.9%	139,632 → 147,897	+5.9%
parse nested object	103,682 → 106,902	+3.1%	87,670 → 89,177	+1.7%
parse array of objects	54,238 → 56,765	+4.7%	52,507 → 52,320	-0.4%
parse large flat object	48,866 → 49,393	+1.1%	46,177 → 46,788	+1.3%
parse mixed types	69,296 → 72,395	+4.5%	67,792 → 65,886	-2.8%
stringify simple object	193,976 → 199,106	+2.6%	162,717 → 157,935	-2.9%
stringify nested object	107,097 → 115,023	🟢 +7.4%	88,778 → 85,787	-3.4%
stringify array of objects	59,133 → 62,738	+6.1%	54,272 → 53,137	-2.1%
stringify mixed types	83,991 → 88,124	+4.9%	74,745 → 69,211	🔴 -7.4%
reviver doubles numbers	43,866 → 45,322	+3.3%	45,974 → 45,251	-1.6%
reviver filters properties	36,844 → 38,361	+4.1%	46,039 → 46,687	+1.4%
reviver on nested object	48,886 → 51,193	+4.7%	52,646 → 52,517	-0.2%
reviver on array	28,912 → 29,346	+1.5%	29,696 → 29,801	+0.4%
replacer function doubles numbers	46,660 → 48,100	+3.1%	52,750 → 52,937	+0.4%
replacer function excludes properties	58,249 → 59,520	+2.2%	61,075 → 62,791	+2.8%
array replacer (allowlist)	111,218 → 115,905	+4.2%	97,979 → 98,811	+0.8%
stringify with 2-space indent	94,117 → 97,673	+3.8%	78,834 → 81,317	+3.2%
stringify with tab indent	94,428 → 97,265	+3.0%	79,641 → 81,362	+2.2%
parse then stringify	53,160 → 54,942	+3.4%	51,593 → 51,607	+0.0%
stringify then parse	31,622 → 32,399	+2.5%	30,931 → 30,859	-0.2%

jsx.jsx — Interp: 🟢 1, 20 unch. · avg +4.9% · Bytecode: 21 unch. · avg +2.4%

Benchmark	Interpreted	Δ	Bytecode	Δ
simple element	197,025 → 206,793	+5.0%	605,346 → 624,992	+3.2%
self-closing element	205,553 → 214,886	+4.5%	635,877 → 666,774	+4.9%
element with string attribute	165,085 → 174,618	+5.8%	442,393 → 451,178	+2.0%
element with multiple attributes	143,704 → 150,053	+4.4%	384,981 → 397,355	+3.2%
element with expression attribute	157,327 → 166,541	+5.9%	427,678 → 436,255	+2.0%
text child	200,218 → 206,912	+3.3%	608,325 → 640,881	+5.4%
expression child	195,291 → 202,887	+3.9%	607,063 → 610,191	+0.5%
mixed text and expression	184,034 → 195,057	+6.0%	553,482 → 570,749	+3.1%
nested elements (3 levels)	74,180 → 77,149	+4.0%	242,235 → 246,028	+1.6%
sibling children	55,637 → 57,100	+2.6%	180,272 → 187,485	+4.0%
component element	139,209 → 148,426	+6.6%	419,867 → 424,312	+1.1%
component with children	86,877 → 89,604	+3.1%	269,652 → 274,362	+1.7%
dotted component	115,823 → 123,318	+6.5%	307,709 → 310,491	+0.9%
empty fragment	205,976 → 218,480	+6.1%	639,262 → 666,250	+4.2%
fragment with children	54,414 → 57,074	+4.9%	181,754 → 183,381	+0.9%
spread attributes	104,404 → 109,945	+5.3%	107,496 → 107,679	+0.2%
spread with overrides	91,515 → 95,474	+4.3%	80,665 → 82,120	+1.8%
shorthand props	152,216 → 159,819	+5.0%	406,761 → 423,625	+4.1%
nav bar structure	24,990 → 26,979	🟢 +8.0%	78,081 → 79,277	+1.5%
card component tree	29,988 → 30,765	+2.6%	84,717 → 85,571	+1.0%
10 list items via Array.from	13,709 → 14,477	+5.6%	23,116 → 23,730	+2.7%

numbers.js — Interp: 11 unch. · avg +2.2% · Bytecode: 🟢 1, 10 unch. · avg +0.9%

Benchmark	Interpreted	Δ	Bytecode	Δ
integer arithmetic	565,026 → 525,727	-7.0%	1,415,209 → 1,534,647	🟢 +8.4%
floating point arithmetic	570,650 → 589,851	+3.4%	1,703,496 → 1,675,065	-1.7%
number coercion	181,026 → 190,082	+5.0%	136,645 → 138,290	+1.2%
toFixed	99,498 → 101,903	+2.4%	210,827 → 212,458	+0.8%
toString	151,738 → 156,706	+3.3%	688,101 → 689,562	+0.2%
valueOf	214,218 → 224,099	+4.6%	947,730 → 916,146	-3.3%
toPrecision	138,875 → 145,397	+4.7%	401,864 → 404,265	+0.6%
Number.isNaN	287,739 → 293,758	+2.1%	172,909 → 177,186	+2.5%
Number.isFinite	286,785 → 285,463	-0.5%	170,646 → 170,540	-0.1%
Number.isInteger	277,273 → 278,484	+0.4%	186,714 → 183,356	-1.8%
Number.parseInt and parseFloat	233,096 → 245,362	+5.3%	152,970 → 157,227	+2.8%

objects.js — Interp: 🟢 1, 6 unch. · avg +5.5% · Bytecode: 🟢 1, 6 unch. · avg +3.9%

Benchmark	Interpreted	Δ	Bytecode	Δ
create simple object	459,785 → 473,727	+3.0%	817,028 → 846,158	+3.6%
create nested object	210,981 → 224,358	+6.3%	356,929 → 367,926	+3.1%
create 50 objects via Array.from	8,904 → 9,423	+5.8%	8,324 → 8,465	+1.7%
property read	612,233 → 627,553	+2.5%	667,680 → 655,795	-1.8%
Object.keys	282,082 → 297,325	+5.4%	205,089 → 213,481	+4.1%
Object.entries	97,841 → 109,343	🟢 +11.8%	63,293 → 66,476	+5.0%
spread operator	185,021 → 191,497	+3.5%	178,075 → 198,508	🟢 +11.5%

promises.js — Interp: 🟢 12 · avg +11.0% · Bytecode: 🟢 1, 11 unch. · avg +1.9%

Benchmark	Interpreted	Δ	Bytecode	Δ
Promise.resolve(value)	534,648 → 580,648	🟢 +8.6%	345,981 → 355,236	+2.7%
new Promise(resolve => resolve(value))	181,775 → 203,276	🟢 +11.8%	167,776 → 169,678	+1.1%
Promise.reject(reason)	521,616 → 566,023	🟢 +8.5%	344,290 → 347,643	+1.0%
resolve + then (1 handler)	162,375 → 176,222	🟢 +8.5%	148,724 → 158,716	+6.7%
resolve + then chain (3 deep)	62,610 → 69,132	🟢 +10.4%	69,875 → 70,168	+0.4%
resolve + then chain (10 deep)	19,888 → 22,306	🟢 +12.2%	23,712 → 22,206	-6.4%
reject + catch + then	89,403 → 99,077	🟢 +10.8%	87,240 → 93,514	🟢 +7.2%
resolve + finally + then	78,877 → 87,270	🟢 +10.6%	76,892 → 76,902	+0.0%
Promise.all (5 resolved)	29,979 → 34,089	🟢 +13.7%	26,793 → 26,938	+0.5%
Promise.race (5 resolved)	31,810 → 35,910	🟢 +12.9%	28,170 → 28,573	+1.4%
Promise.allSettled (5 mixed)	25,378 → 28,229	🟢 +11.2%	22,043 → 23,458	+6.4%
Promise.any (5 mixed)	29,805 → 33,740	🟢 +13.2%	26,450 → 26,798	+1.3%

strings.js — Interp: 🔴 1, 10 unch. · avg +2.7% · Bytecode: 🟢 2, 9 unch. · avg +2.3%

Benchmark	Interpreted	Δ	Bytecode	Δ
string concatenation	415,547 → 413,655	-0.5%	429,996 → 428,276	-0.4%
template literal	700,755 → 640,100	🔴 -8.7%	651,790 → 704,492	🟢 +8.1%
string repeat	394,779 → 410,886	+4.1%	953,332 → 990,507	+3.9%
split and join	133,762 → 136,850	+2.3%	310,333 → 322,407	+3.9%
indexOf and includes	161,214 → 169,873	+5.4%	602,735 → 605,301	+0.4%
toUpperCase and toLowerCase	244,261 → 254,388	+4.1%	623,424 → 710,207	🟢 +13.9%
slice and substring	152,705 → 162,184	+6.2%	674,670 → 660,927	-2.0%
trim operations	183,345 → 188,378	+2.7%	743,080 → 746,297	+0.4%
replace and replaceAll	202,365 → 211,487	+4.5%	668,675 → 663,836	-0.7%
startsWith and endsWith	129,931 → 136,711	+5.2%	512,834 → 502,075	-2.1%
padStart and padEnd	191,810 → 200,234	+4.4%	579,858 → 581,845	+0.3%

typed-arrays.js — Interp: 🟢 2, 20 unch. · avg +3.7% · Bytecode: 🟢 2, 🔴 1, 19 unch. · avg +0.3%

Benchmark	Interpreted	Δ	Bytecode	Δ
new Int32Array(0)	311,345 → 333,988	🟢 +7.3%	128,971 → 129,048	+0.1%
new Int32Array(100)	288,309 → 307,709	+6.7%	123,596 → 124,549	+0.8%
new Int32Array(1000)	169,338 → 177,923	+5.1%	65,577 → 57,935	🔴 -11.7%
new Float64Array(100)	259,755 → 276,797	+6.6%	110,870 → 107,321	-3.2%
Int32Array.from([...])	179,577 → 184,149	+2.5%	155,138 → 153,236	-1.2%
Int32Array.of(1, 2, 3, 4, 5)	308,565 → 319,460	+3.5%	234,078 → 242,991	+3.8%
sequential write 100 elements	3,625 → 3,712	+2.4%	13,213 → 12,925	-2.2%
sequential read 100 elements	3,730 → 3,814	+2.3%	10,228 → 10,173	-0.5%
Float64Array write 100 elements	3,372 → 3,417	+1.3%	12,534 → 12,144	-3.1%
fill(42)	45,601 → 45,905	+0.7%	43,936 → 43,954	+0.0%
slice()	204,057 → 211,834	+3.8%	186,189 → 184,633	-0.8%
map(x => x * 2)	8,199 → 8,486	+3.5%	7,459 → 7,801	+4.6%
filter(x => x > 50)	8,292 → 8,596	+3.7%	7,859 → 7,988	+1.6%
reduce (sum)	7,880 → 8,177	+3.8%	6,744 → 6,954	+3.1%
sort()	177,194 → 172,286	-2.8%	151,059 → 155,240	+2.8%
indexOf()	450,459 → 461,409	+2.4%	361,542 → 354,664	-1.9%
reverse()	332,846 → 341,527	+2.6%	271,068 → 268,327	-1.0%
create view over existing buffer	389,343 → 415,980	+6.8%	138,274 → 137,664	-0.4%
subarray()	449,450 → 483,421	🟢 +7.6%	352,010 → 350,014	-0.6%
set() from array	587,777 → 606,744	+3.2%	233,585 → 250,844	🟢 +7.4%
for-of loop	5,101 → 5,243	+2.8%	18,429 → 17,635	-4.3%
spread into array	17,648 → 18,792	+6.5%	46,499 → 53,163	🟢 +14.3%

_{Measured on ubuntu-latest x64. Changes within ±7% are considered insignificant.}

The typed local opcodes (56-65) are no longer emitted by the compiler and had no case arms in the VM. Remove the dead enum entries entirely and bump SOUFFLE_FORMAT_VERSION from 3 to 4. Update documentation to reflect that local access always uses OP_GET_LOCAL/OP_SET_LOCAL. Made-with: Cursor

coderabbitai

🧹 Nitpick comments (1)

souffle/Souffle.Bytecode.pas (1)

98-106: Consider adding a comment documenting the removed typed local opcodes (56-65).

Similar to the comment on line 104 about opcode 68, a brief note about the removed typed local opcodes would help future maintainers understand the gap between OP_CONCAT = 55 and OP_NEW_BLUEPRINT = 66.

📝 Optional documentation comment

     // ── Core: String ──
     OP_CONCAT        = 55,  // ABC   R[A] := String(R[B]) + String(R[C])
 
+    // opcodes 56-65 removed: were OP_GET_LOCAL_INT through OP_SET_LOCAL_REF (typed local opcodes)
+    // use OP_GET_LOCAL/OP_SET_LOCAL instead — type optimization is via arithmetic/compare opcodes
+
     // ── Core: Blueprint ──
     OP_NEW_BLUEPRINT = 66,  // ABx   R[A] := new Blueprint(name=Constants[Bx])

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@souffle/Souffle.Bytecode.pas` around lines 98 - 106, Add a brief comment
between OP_CONCAT and OP_NEW_BLUEPRINT in Souffle.Bytecode.pas documenting that
opcode slots 56–65 were removed (the typed-local opcodes) and why/what replaced
them—mirroring the style of the existing comment for the removed
OP_BLUEPRINT_METHOD (68); reference the surrounding symbols OP_CONCAT,
OP_NEW_BLUEPRINT and the removed range 56-65 so future maintainers understand
the numeric gap.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@souffle/Souffle.Bytecode.pas`:
- Around line 98-106: Add a brief comment between OP_CONCAT and OP_NEW_BLUEPRINT
in Souffle.Bytecode.pas documenting that opcode slots 56–65 were removed (the
typed-local opcodes) and why/what replaced them—mirroring the style of the
existing comment for the removed OP_BLUEPRINT_METHOD (68); reference the
surrounding symbols OP_CONCAT, OP_NEW_BLUEPRINT and the removed range 56-65 so
future maintainers understand the numeric gap.

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: e4ac18b4-dba1-46fb-8d54-825d0a8a1239

📥 Commits

Reviewing files that changed from the base of the PR and between 8963103 and 8afab47.

📒 Files selected for processing (2)

docs/souffle-vm.md
souffle/Souffle.Bytecode.pas

- Rewrite optimization-log.md with correct PR numbers for all 22 PRs since #73 (previous version had multiple wrong PR-to-description mappings) - Add missing PRs: #73, #74, #75, #76, #77, #79, #80, #81, #82, #83, #85, #88, #89, plus infrastructure PRs #90, #92, #98 - Fix GC section: Collect-after-each-file is deliberate, not a workaround; any GC revisit needs fresh analysis - Fix minor items: link to actual issues #103 and #104, correct naming convention context (shared units should NOT use Goccia.* prefix) - Create docs/decision-log.md with 1-3 sentence summaries linking to detailed documentation - Link fpc-generics-performance.md and fpc-dispatch-performance.md from AGENTS.md, code-style.md, design-decisions.md, and optimization-log.md Made-with: Cursor

…105) * Convert spike PDFs to Markdown, update VM docs, add optimization log - Convert all 4 PDF spike documents (string performance, hashmap performance, generics performance, dispatch performance) to Markdown for better AI agent parsing - Fix stale TSouffleValue documentation: 26 bytes → 16 bytes, string[23] → string[13] across souffle-vm.md, design-decisions.md - Update test count to 3,501 across all docs (AGENTS.md, README.md, souffle-vm.md, design-decisions.md) - Add svkString to TSouffleValue kinds list in AGENTS.md - Add THashMap optimization details to AGENTS.md component table - Add docs/optimization-log.md with full VM optimization history, experiment results, remaining work, and key insights for fresh context pickup - Add hashmap spike update note documenting TScopeMap removal, TOrderedStringMap refactoring, and THashMap hash optimization Made-with: Cursor * Fix optimization log PR numbers, add decision log, link all spikes - Rewrite optimization-log.md with correct PR numbers for all 22 PRs since #73 (previous version had multiple wrong PR-to-description mappings) - Add missing PRs: #73, #74, #75, #76, #77, #79, #80, #81, #82, #83, #85, #88, #89, plus infrastructure PRs #90, #92, #98 - Fix GC section: Collect-after-each-file is deliberate, not a workaround; any GC revisit needs fresh analysis - Fix minor items: link to actual issues #103 and #104, correct naming convention context (shared units should NOT use Goccia.* prefix) - Create docs/decision-log.md with 1-3 sentence summaries linking to detailed documentation - Link fpc-generics-performance.md and fpc-dispatch-performance.md from AGENTS.md, code-style.md, design-decisions.md, and optimization-log.md Made-with: Cursor * Address PR review comments: fix ratios, bytes/chars, stale refs - Fix generics spike ratio values (arithmetic errors from original PDF) - Change "13 characters" to "13 bytes" in souffle-vm.md (FPC string[N] is byte-counted), add UTF-8 note - Mark TScopeMap as removed/historical in hashmap spike document - Fix Key Insight #4 PR reference (was #88 async/await, not boolean) - Remove stale GOCCIA_EXT_EVAL_CLASS / FPendingClasses / decorator bridge references from AGENTS.md, souffle-vm.md, design-decisions.md (decorators are now compiled natively via BEGIN/APPLY/FINISH opcodes) - Fix "44 second" → "44-second" hyphenation in string spike Made-with: Cursor * Expand TScopeMap history in hashmap spike with full timeline The spike originally recommended TScopeMap for scope bindings based on micro-benchmarks at N≤10. After implementation (PR #66), real-world profiling showed the assumption about scope sizes was incorrect — global and bridged scopes far exceed 20 bindings, causing a 2.7× regression. The update section now tells the complete timeline: original recommendation → implementation → profiling discovery → revert. Made-with: Cursor * Add grill-me interview skill Introduce a new 'grill-me' agent skill that relentlessly interviews users about a plan or design, exploring each branch of the decision tree until a shared understanding is reached and dependencies are resolved. The skill prefers to inspect the codebase when that can answer a question. Updated skills lock to include the new skill.

coderabbitai Bot reviewed Mar 12, 2026

View reviewed changes

frostney added the implement-feedback label Mar 12, 2026

frostney added 2 commits March 12, 2026 22:49

Merge branch 'main' into opt/remove-typed-locals

8963103

coderabbitai Bot reviewed Mar 12, 2026

View reviewed changes

frostney merged commit ac221f2 into main Mar 13, 2026
9 checks passed

frostney deleted the opt/remove-typed-locals branch March 13, 2026 08:06

coderabbitai Bot mentioned this pull request Mar 13, 2026

Promote 10 hot opcodes into main VM dispatch loop #81

Merged

2 tasks

coderabbitai Bot mentioned this pull request Apr 9, 2026

Add bytecode VM profiling support #237

Merged

frostney added the internal Refactoring, CI, tooling, cleanup label Apr 9, 2026

coderabbitai Bot mentioned this pull request May 3, 2026

Fix performance regressions from #496 #508

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Remove unused typed local opcodes from compiler and VM#74

Remove unused typed local opcodes from compiler and VM#74
frostney merged 3 commits intomainfrom
opt/remove-typed-locals

frostney commented Mar 12, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

coderabbitai Bot commented Mar 12, 2026 •

edited

Loading

Walkthrough

Changes

Sequence Diagram(s)

Estimated code review effort

Possibly related PRs

Poem

Uh oh!

coderabbitai Bot left a comment

Uh oh!

frostney commented Mar 12, 2026

Uh oh!

github-actions Bot commented Mar 12, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Mar 12, 2026 •

edited

Loading

Uh oh!

coderabbitai Bot left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

frostney commented Mar 12, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Test plan

Summary by CodeRabbit

Uh oh!

coderabbitai Bot commented Mar 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Sequence Diagram(s)

Estimated code review effort

Possibly related PRs

Poem

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

frostney commented Mar 12, 2026

Uh oh!

github-actions Bot commented Mar 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Suite Timing

Uh oh!

github-actions Bot commented Mar 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Benchmark Results

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

frostney commented Mar 12, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented Mar 12, 2026 •

edited

Loading

github-actions Bot commented Mar 12, 2026 •

edited

Loading

github-actions Bot commented Mar 12, 2026 •

edited

Loading