Published status of HHY semantics, Runtime, performance, and engineering quality with reproducible CI measurements.
Release summary
Dimension
Question
Current conclusion
Language baseline
Are core semantics stable?
Pipe, Value, Stream, Error, and core callable contracts are frozen
Runtime health
Are resource, memory, and cancellation boundaries reliable?
Resource limits, GC stress, sanitizers, fuzzing, and explicit ownership governance are present
Performance
Is performance measurable and controlled?
Fixed workloads, five-sample medians, machine-readable evidence, and blocking budgets are established
Engineering governance
Are changes auditable?
Four-platform CI, layered gates, version consistency, and release evidence form a closed loop
Data at a glance
Signal
Result
Evidence basis
Current formal release
v1.3.10
Four platform archives, per-asset SHA-256, and SHA256SUMS
Execution engines
Bytecode default / AST fallback
Full dual-engine suite and machine-readable decision
Core callables
96
Runtime Callable Contract Registry
Continuous-verification platforms
4
macOS arm64, Linux arm64, Linux x86_64, Windows x86_64
Final CI engine gates
All pass
1M CPU 0.3695; short task 1.0088; sustained JSON 1.0207
Profiler overhead
1.0269× / +2.786 ms
Nine samples; limits 1.35× and 12 ms
Bytecode cache
Not admitted
Five workloads × 21 samples; compile+verify is not a major cost
Complete practical projects
6
AST/Bytecode end-to-end acceptance with stable exit status
Overall baseline and compatibility
Baseline
Stable commitment
Verification
Language semantics
No second Pipe, Stream, or Error model
Specification examples, Parser/Checker fixtures, and valid-program regression
Callable contracts
Names, arity, effect, lazy, cancellable, and threading metadata are machine-readable
Contract Registry JSON and 96-contract consistency checks
Diagnostics
CLI text and JSON/LSP share the Core checking path
Diagnostic schema and LSP protocol tests
Extension boundary
Third-party capabilities prefer the Process Extension Protocol
Manifest integrity, Protocol 1, and official-extension acceptance
C ABI
Runtime internals are not currently a public ABI
Reconsider only when real integrations prove the process protocol insufficient
The current formal baseline is v1.3.10. Bytecode is default; the AST evaluator remains the semantic oracle and is selectable through --engine ast or HHY_ENGINE=ast. Compiler-produced Stream Kernels must pass their independent Verifier; dynamic and unknown shapes fall back losslessly to general Bytecode.
v1.2.2 release and extension status
Capability
Current state
Acceptance result
Extension distribution
Ed25519-signed Registry and deterministic resolution
Tampering, unknown sources, and dependency conflicts fail closed
Reproducible environment
Lockfile and content-addressed offline cache
The same lock yields the same graph and rebuilds offline
Safe change
Transactional install, upgrade, and explicit rollback
Failed upgrades preserve the old environment
HTML 0.2.0
Lexbor, CSS selectors, and single-parse multi-field projection
Malformed HTML, hard limits, truncation, and structured errors pass on four platforms
Protocol decision
Retain the bounded synchronous batch API
No real evidence requires stream credit, cross-call cancellation, or opaque handles
Final v1.3.10 CI evidence comes from commit 4ddc8c3 on GitHub Actions Ubuntu 24.04: schema-2 paired/interleaved engine benchmarks plus independent Profiler and cache-decision artifacts. Ratios are Bytecode/AST wall time; lower than one means Bytecode is faster.
Gate
Measured
Limit
Result
1M CPU
0.3695×
0.90×
Pass
Short task
1.0088×
1.25×
Pass
Sustained JSON/I/O
1.0207×
1.10×
Pass
Profiler overhead
1.0269× / +2.786 ms
1.35× / +12 ms
Pass
Cache-admission workload
Compile+verify median
Cold-run median
Share
Hello
0.0097 ms
5.860 ms
0.1649%
Advanced Flow
0.0269 ms
5.916 ms
0.4546%
Stdlib
0.0286 ms
7.003 ms
0.4086%
Sustained JSON
0.0089 ms
33.518 ms
0.0266%
Core Flow 1M
0.0105 ms
166.285 ms
0.0063%
v1.3.10 six-runtime same-machine rerun
Rerun on 2026-09-01 on macOS 26.6.2 arm64 with HHY 1.3.10, PHP 8.5.10, Go 1.27.0, Python 3.14.7, Lua 5.5.1, and OpenJDK 26.0.2.1. The fixed task maps, filters, stable-distincts, materializes, and counts one million integers; all six implementations validate output 333334. After two warmups, two independent rounds use seven deterministically shuffled and interleaved fresh processes per runtime. Wall time includes process startup; Go and Java are precompiled and compile time is excluded.
Implementation
Version
Round 1 median
Round 2 median
Combined range
Go
1.27.0
7.995 ms
7.969 ms
7.622–15.325 ms
Lua
5.5.1
18.368 ms
17.957 ms
17.582–19.822 ms
PHP
8.5.10
43.174 ms
43.549 ms
42.588–49.033 ms
Java
OpenJDK 26.0.2.1
49.394 ms
48.153 ms
46.995–53.066 ms
HHY Bytecode
1.3.10
55.297 ms
53.404 ms
51.027–78.702 ms
Python
3.14.7
81.747 ms
86.459 ms
79.993–87.774 ms
Comparison
Round 1
Round 2
Interpretation
HHY / PHP
1.28×
1.23×
HHY uses about 23%–28% more wall time on this task
HHY / Java
1.12×
1.11×
Java is slightly faster including fresh-JVM startup
HHY / Python
0.68×
0.62×
HHY takes less wall time on this task
HHY / Lua
3.01×
2.97×
Lua is faster on this integer loop
HHY / Go
6.92×
6.70×
Precompiled Go remains substantially faster
Governance conclusion and watch list
Overall status: v1.3.10 is formally released; v1.3.7–v1.3.10 semantic, three-path, Profiler, resource, and cache-governance gates are complete.
Engine policy: Bytecode is default; AST remains permanently available as the semantic oracle, differential-test engine, and --engine ast emergency fallback.
Performance conclusion: final CI reports 0.3695 for the 1M CPU ratio, 1.0088 for short tasks, 1.0207 for sustained JSON, and 1.0269× Profiler overhead; all gates pass.
Cache conclusion: compile+verify is not a major cold-run cost; no cache is introduced until new data and a complete threat model trigger review.
Cross-language conclusion: this same-machine fixed task has Go/Lua/PHP/Java ahead of HHY and HHY ahead of Python; it cannot be generalized to all workloads.
Update rule: synchronize this report whenever the release baseline, measurement method, engine/cache decision, or overall risk conclusion changes.
