Evaluate compiler-produced pure Edict programs

[flyingrobots]

Summary

Evaluate exact compiler-produced pure Edict packages under a verified host pin
and explicit decode, execution, allocation, and output budgets.

This PR is stacked on #724. It advances #684 but does not close that issue.
Application nouns and verbs remain opaque compiled data. Echo production code
contains no editing operation, rope algorithm, or application-specific dispatch.

Behavior and authority

The trusted-host API evaluates unsigned values, bounded bytes, records, locals,
field access, integer comparisons, conditionals, and zero-argument authored
helpers. It checks runtime types and authored input constraints, rejects package
substitution, and rejects unsupported syntax even in unselected branches.

The caller must obtain the package identity from an independently verified,
authorized release. A matching digest alone does not establish authorization.
Evaluation has no graph, I/O, callback, clock, or WAL access. The storage budget
measures deterministic semantic storage units, not physical heap allocation.

This is pure computation, not package installation, admission, scheduler
settlement, causal evidence, or end-to-end editing. The complete application
goal remains in flyingrobots/jedit#295.

Validation

The initial exact-package witness failed with UnsupportedProgram. Seven
integration tests and the architecture-independent metering regression now
pass. A separately compiled source mutation changes the result from 1 to 2;
the original verified package pin rejects that changed package.

At 8c725d699241a7e3adee482029031ff6bade25fa, all published-head CI checks passed,
including the Edict provider host contract and independent build comparison.
CodeRabbit skipped its review while this PR was a draft; that is not approval.

cargo test --locked -p warp-core --features trusted_runtime --test edict_pure_evaluation_tests
cargo test --locked -p warp-core --features trusted_runtime --lib edict_pure::
cargo clippy --locked -p warp-core --features trusted_runtime --lib --test edict_pure_evaluation_tests -- -D warnings
bash tests/hooks/test_verify_local.sh

All 162 hook regressions passed. CI and local hook commands explicitly enable
trusted_runtime, preventing a vacuous zero-test success.

Documentation and evidence

The canonical contract is
application-contract-hosting.md.
Exact fixture bytes and public-build reproduction instructions are in
edict-pure-jedit.
The fixture producer commits are Jedit a894c7c4, Edict 3f81f759, and
Echo provider 49e9efb6. The pinned provider bytes were not changed.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Advanced

Run ID: c1a2f4e2-9fb0-4e82-a761-02d7f204152a

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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

• X
• Mastodon
• Reddit
• LinkedIn

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

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.