Bound public parsing resource use and add scale regressions

[coderabbitai]

Summary

Bound public parsing resource use and add regression coverage for large input, deep conditional nesting, and output growth.

Rationale

The CLI limits source reads to 16 MiB, but public parse_source accepts unbounded byte slices. Nested facts clone full condition ancestry. The repository has no realistic 10 MiB regression or output-growth measurement to prevent resource regressions.

Affected areas

• src/application.rs
• src/domain/
• src/adapters/source.rs
• src/adapters/makefile.rs
• Parser and integration test modules
• Resource-limit documentation

Required changes

• Enforce a source-size bound at the public parsing boundary, or define and enforce equivalent bounds for condition ancestry and report/output growth.
• Keep CLI and library behaviour consistent for bounded input.
• Reduce or cap repeated condition-ancestry allocation where required by the selected limit.
• Add a realistic approximately 10 MiB input regression.
• Add a deep-nesting regression.
• Define measurable time and memory limits, or use deterministic bounded-output assertions that detect unbounded growth.
• Document public resource limits and failure behaviour.

Acceptance criteria

• parse_source cannot process unbounded input without an explicit documented resource contract.
• Large input at the supported limit has regression coverage.
• Deep conditional nesting has regression coverage.
• Tests detect excessive output or ancestry growth.
• Tests specify and enforce measurable resource limits or deterministic bounded-growth limits.
• CLI diagnostics and existing source-size error behaviour remain stable.
• Documentation states the public parsing and output resource limits.

Backlinks

• Originating PR: (adr-0001) Parse one GNU Makefile into versioned JSON facts #6
• Source request: The GitHub API did not expose a permalink for the request comment at issue creation time.
• Requester: @leynos

[coderabbitai]

🔗 Related PRs

leynos/mdtablefix#343 - Preserve fenced code block contents during wrapping (#329) [closed]
leynos/whitaker#238 - Map candidate spans to ra_ap_syntax nodes and extract AST feature vectors (7.3.1) [merged]
#6 - (adr-0001) Parse one GNU Makefile into versioned JSON facts [open]
leynos/mdtablefix#407 - Share canonical frontmatter boundary (#361) [merged]
leynos/whitaker#310 - Emit bounded AST lowering metrics (7.3.2) (#308) [open]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!