ROADMAP: add browser-target reliability (#609 + #610) to stabilisation tier

[aallan]

Summary

Adds browser-target reliability to the stabilisation tier in ROADMAP.md. Two Conway's Life agent experiments on --target browser surfaced five concrete blockers, and the agent's design memo maps each obstacle to a runtime-only fix path. Promoting #609 (JSPI-driven IO.sleep) and #610 (ANSI subset interpreter) into the stabilisation tier commits to "browser-target Vera is something you'd actually use" before the agent-integration push.

Why

The current stabilisation framing recognises two patterns from the v0.0.120–v0.0.138 bug-killing campaign: scale-dependent bugs (#596 stress harness) and walker-completeness gaps (#597 walker audit). Both are codegen-side concerns. A third pattern is now visible from the same evidence base — browser-target reliability is approximate, not real — and deserves the same treatment.

Specifics, all surfaced by the recent agent experiments:

• Timing: IO.sleep busy-waits and freezes the tab. Animation programs literally can't run. Fix: #609 (JSPI / Asyncify).
• Rendering: ANSI escapes render as literal text. Cursor-addressable terminal programs can't render. Fix: #610 (ANSI subset interpreter, ~200 lines of JS).
• Marshalling: JS can't pass String arguments to Vera functions. Forces all browser programs into a "compute everything upfront" pattern. Fix: #603 (export runtime.mjs helpers).
• Codegen gap A: Five prelude combinators silently skipped on every WASM compile. Bug: #604.
• Codegen gap B: String-returning function call in interpolation produces invalid WASM. Bug: #602.

The first two are the high-leverage ones for the stabilisation tier — they're the difference between "browser-target Vera is theoretical" and "browser-target Vera is something you'd actually use." The other three are tracked in KNOWN_ISSUES.md already.

Changes

• Campaign-pattern list: 2 → 3 patterns. Third pattern documents the browser-target seam with links to the umbrella issue (#608) and the five concrete blockers.
• Stabilisation tier: 3 → 5 items. Added Browser runtime: implement IO.sleep via JSPI (or Asyncify fallback) so animations don't freeze the tab #609 (JSPI sleep) as item 3, Browser runtime: interpret a minimal ANSI escape subset so terminal-style Vera programs render unchanged #610 (ANSI subset interpreter) as item 4, with mechanism notes. Demoted macOS malloc abort during wasmtime cleanup after Ctrl-C-in-host-import #595 (wasmtime-py malloc abort) from item 3 to item 5 — it's contingent on an upstream release we don't control, so it's the longest-tail item.
• Agent-integration tier: renumbered 4/5/6 → 6/7/8 to match.
• "What moves when" gate: "Add coherence safeguards and README narratives #4 starts when v0.0.4: Typed AST layer #1–Add Hello World example and spec design notes #3 are closed" → "ci: bump github/codeql-action from 3 to 4 #6 starts when v0.0.4: Typed AST layer #1–ci: bump actions/setup-python from 5 to 6 #5 are closed".

Test plan

• python scripts/check_doc_counts.py — clean
• No content changes outside ROADMAP.md (single-file edit)
• All issue references (Browser runtime: implement IO.sleep via JSPI (or Asyncify fallback) so animations don't freeze the tab #609, Browser runtime: interpret a minimal ANSI escape subset so terminal-style Vera programs render unchanged #610, etc.) resolve to the correct existing issues

Summary by CodeRabbit

• Documentation
  • Updated the project roadmap to document runtime stability improvements, including scale-dependent bug tracking, diagnostic enhancements, and browser compatibility work. The implementation roadmap has been reorganised with updated priorities for ongoing stabilisation efforts.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 669f5733-3d66-46c9-8716-f1581681fe37

📥 Commits

Reviewing files that changed from the base of the PR and between 20f3ccb and 7f02052.

📒 Files selected for processing (1)

• ROADMAP.md

---

📝 Walkthrough

Walkthrough

ROADMAP.md is updated to document three failure patterns in the runtime-correctness stabilisation campaign and to reprioritise the implementation tiers. The three patterns are enumerated, with the new browser-target reliability pattern expanded with concrete issue references and architectural gaps. The stabilisation-tier table is updated with new browser runtime tasks and a macOS Ctrl-C issue.

Changes

Stabilisation Patterns and Priorities

Layer / File(s)	Summary
Stabilisation Patterns Documentation ROADMAP.md	"What's next" section expands to document three closure-worthy patterns: scale-dependent bugs with permanent VERA_EAGER_GC diagnostic lane, walker-completeness gaps, and browser-target reliability blockers spanning IO timing, ANSI rendering, string marshalling, and codegen gaps.
Stabilisation Tier Prioritisation ROADMAP.md	Implementation order table is updated with new browser runtime reliability entries: IO.sleep via JSPI and ANSI subset interpreter (#609, #610), plus macOS Ctrl-C wasmtime trampoline (#595). These replace earlier items and reorder the near-term checklist.
Agent-Integration Tier Roadmap ROADMAP.md	Agent-integration tier table is reflowed relative to revised stabilisation sequencing, preserving LSP, vera context, and configurable Inference.complete parameters as post-stabilisation priorities.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

Possibly related issues

• Browser runtime: implement IO.sleep via JSPI (or Asyncify fallback) so animations don't freeze the tab #609: Browser runtime IO.sleep via JSPI implementation is explicitly listed as a stabilisation-tier priority in the updated roadmap.
• Browser runtime: interpret a minimal ANSI escape subset so terminal-style Vera programs render unchanged #610: Browser runtime ANSI escape subset interpretation is explicitly listed as a stabilisation-tier priority in the updated roadmap.
• macOS malloc abort during wasmtime cleanup after Ctrl-C-in-host-import #595: macOS wasmtime Ctrl-C trampoline is newly added to the stabilisation-tier implementation order.
• Doc: terminal-vs-browser IO seam in 'write once run anywhere' framing #608: The terminal-vs-browser IO seam framing issue is the conceptual boundary driving the three failure patterns now documented in the roadmap.

Possibly related PRs

• aallan/vera#483: Both PRs modify ROADMAP.md's implementation-order table and short-term priorities sequencing.

Suggested labels

docs

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarises the primary change: adding browser-target reliability items #609 and #610 to the stabilisation tier in ROADMAP.md.
Linked Issues check	✅ Passed	The PR updates ROADMAP.md to document and prioritise browser-target fixes (#609 IO.sleep via JSPI, #610 ANSI interpretation) into the stabilisation tier. Both are directly addressed and correctly referenced in the roadmap changes.
Out of Scope Changes check	✅ Passed	ROADMAP.md is the sole file modified. All changes are scoped to documenting the three stabilisation campaign patterns and promoting browser-target fixes (#609, #610) into the tier. No out-of-scope implementation or unrelated edits detected.
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
• Commit unit tests in branch claude/roadmap-browser-target

---

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

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.92%. Comparing base (20f3ccb) to head (7f02052).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #611   +/-   ##
=======================================
  Coverage   90.92%   90.92%           
=======================================
  Files          59       59           
  Lines       22997    22997           
  Branches      259      259           
=======================================
  Hits        20910    20910           
  Misses       2080     2080           
  Partials        7        7           

Flag	Coverage Δ
javascript	57.36% <ø> (ø)
python	94.74% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[aallan]

Folding into #601 — both PRs originate from the same agent-experiment session and touch independent files (no conflict). The merge-coordination cost of two PRs consistently exceeds the review-clarity benefit. ROADMAP commit cherry-picked onto #601's branch as 757435e.