fix: address PR review comments for opensmith tooling

[Copilot]

Addresses review feedback from PR #5 on the OpenSmith parser front-end scaffold.

opensmithgen.c

• Strip trailing -- from comment bodies (was keeping comment -- instead of comment)
• Add raw_len field to ast_node_t for byte-accurate reconstruction—handles embedded NUL bytes that strlen() would truncate
• Skip empty final literal segments to avoid polluting AST with zero-length nodes
• Defensive null check in reconstruct() loop

// Before: body contains delimiter artifacts
"body": "This is a comment --"

// After: clean comment text
"body": "This is a comment"

Shell security/correctness

• opensmith_parity.sh: Fix shell injection—pass fixture as separate argv instead of sh -c "$ENGINE $fixture"
• opensmith_frontend_check.sh: Check -x (executable) not -f (exists)
• opensmith_corpus.sh: Fix nested_zip_prefix grep pattern to handle JSON whitespace ("key": "value" vs "key":"value")

Makefile

• opensmithgen-ape target: Add cosmocc detection with fallback to ~/.cosmocc/bin/cosmocc and actionable error message (matches existing ape: target)

No Python removed—scripts were already POSIX sh.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a67ab5ecc5

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Preserve command-template semantics for --engine

Invoking the engine as "$ENGINE" "$fixture" now treats the entire --engine string as a single executable path, so command templates with arguments (including the documented ENGINE="cat {input}" flow in docs/OPENSMITH_PARITY.md) fail with 127 for every fixture instead of actually running the tool. This is a regression from the previous {input} substitution behavior and breaks parity runs that need an interpreter + script or any extra flags.

Useful? React with 👍 / 👎.

[ludoplex]

refactor with previous {input} substitution behavior

[Copilot]

Pull request overview

Addresses follow-up review feedback for the OpenSmith front-end tooling by improving parser roundtrip fidelity and tightening shell harness correctness/security.

Changes:

• Update opensmithgen AST handling to strip comment delimiter artifacts and reconstruct using stored byte lengths (avoids strlen() truncation issues).
• Harden/correct OpenSmith shell harnesses (engine execution, executable checks, JSON whitespace handling in lock parsing).
• Improve opensmithgen-ape Makefile target with cosmocc detection + actionable error output.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tools/opensmithgen/opensmithgen.c	Adds raw_len for byte-accurate reconstruction and cleans up comment/literal parsing.
scripts/opensmith_parity.sh	Changes engine execution to avoid shell injection by passing fixture as a separate argv.
scripts/opensmith_frontend_check.sh	Validates tool is executable (-x) rather than merely present (-f).
scripts/opensmith_corpus.sh	Makes nested_zip_prefix extraction tolerant to JSON whitespace.
Makefile	Adds cosmocc detection/fallback for opensmithgen-ape builds.

---

You can also share your feedback on Copilot code review. Take the survey.