feat: add commit_files for multi-path commits and fix repo-relative path resolution for nested config dirs

[Scott McMaster (scottmcmaster)]

Summary

This is to improve/harden the git layer to make it easier and less bug-prone supporting nested config dirs, based on this comment from Alex Shabalin (@alex-sparus) here: #655 (comment)

It also proactively adds the commit_files function which was in the PR that inspired said comment.

The main thing to look at is the repo_relative_path method -- the correctness of its logic as well as its call sites.

Test Plan

New unit tests.

Docs

• Docs updated (companion PR in darkmatter/nixmac-web: #___)
• No docs update needed

[Scott McMaster (scottmcmaster)]

• feat: add commit_files for multi-path commits and fix repo-relative path resolution for nested config dirs #679 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[darkmatter]

🎨 Storybook preview

Open Storybook preview

Updated for 69e2e8d

---

⚠️ Detected UI changes (5)

These stories' HTML snapshots changed. I've added screenshots + links to the changed stories below. Review them carefully then accept the changes to regenerate baselines and include them in this PR:

Flows/Evolve › Playground

Flows/Evolve › 1. Begin (idle)

Flows/Evolve › 2. Evolving (progress)

Flows/Evolve › Evolving With Error Event

Flows/Evolve › 3. Review (changes generated)

---

Accept UI changes

• Click here to accept these changes

Alternatively, you can run bun run test:update-snapshots locally to re-generate the baselines and then push the changes to this PR.

What does this do?

The screenshots above show UI changes detected by the Storybook
snapshot tests run on this PR. Each image is the rendered output of
a Storybook story from the code in this PR branch; the snapshot
test compared it against the committed baseline in
__snapshots__/ and flagged the difference.

Checking the box tells the darkmatter[bot] to regenerate the
baselines from this PR's current code and commit them directly to
this branch. The new baselines become the source of truth for
future runs — only accept after confirming the visual changes are
intentional.

Comparison baseline: the committed __snapshots__/ files on this
PR branch (carried forward from develop). Accept updates them in
place on this branch.

[darkmatteragent]

Review — APPROVE

69e2e8d2c82f · 0 findings

Approve: solid hardening — every producer/consumer boundary checks out. - All app callers pass the git toplevel plus repo-relative filenames, which repo_relative_path reproduces unchanged; the nested-config-dir case is fixed as intended. - Per-hunk commit/discard keeps hunk granularity through find_hunk_by_hash; no sibling-row mutation exists. - Escape/symlink paths fail closed before mutation, and the index.write() reorder preserves intent-to-add entries, with new tests locking both in.

No findings.

[darkmatteragent]

Reviewer · darkmatter agent

VERDICT APPROVE

Approve: solid hardening — every producer/consumer boundary checks out. - All app callers pass the git toplevel plus repo-relative filenames, which `repo_relative_path` reproduces unchanged; the nested-config-dir case is fixed as intended. - Per-hunk commit/discard keeps hunk granularity through `find_hunk_by_hash`; no sibling-row mutation exists. - Escape/symlink paths fail closed before mutation, and the `index.write()` reorder preserves intent-to-add entries, with new tests locking both in.

No findings.

reviewed: 69e2e8d — 2026-09-04T07:53Z verdict: approve findings: 0

Reviewer darkmatter agent

Run details

Metadata	Value
Agent	darkmatter/agent/reviewer
Session	pr:darkmatter/nixmac#679
Kind	pr
Runtime	platform-runtime
Configured model	litellm/glm-5.2-fp8
Target	darkmatter/nixmac#679
Revision	69e2e8d2c82fe2af0ff501d8f13b678545dc336c
Verdict	approve
Reviewed at	2026-09-04T07:53Z
Logs	Open Braintrust session

[github-actions]

	Warnings
⚠️	No Linear issue ID found in this PR's title, description, or branch name (expected something like ENG-123). Add one so this work is traceable in Linear, or add #no-linear to the PR description to acknowledge it's intentionally untracked.

📋 PR Overview

Lines changed	324 (+299 / -25)
Files	0 added, 2 modified, 0 deleted
Draft / WIP	no
Has Test Plan	yes
Linear issue	no
No Test Plan Needed	no
New UI components	no
New Storybook stories	no
New Rust modules	no
New TS source files	no
New tests	no
package.json touched	no
Cargo.toml touched	no
Infra / CI touched	no

🔬 Coverage

Report	Lines	Statements	Functions	Branches
apps/native/coverage/coverage-summary.json	34.3%	33.8%	29.0%	29.2%

Generated by 🚫 dangerJS against 69e2e8d

[Alex Shabalin (alex-sparus)]

Severity 2 — not blocking

commit_hunk still has the write-before-bail shape this PR fixes in commit_files: further down (line 550) the index is written before the nothing to commit check. It's practically unreachable today (find_hunk_by_hash just found the hunk in the live diff), but the lookup→commit window is a check-then-act race, and when the bail does fire it clobbers intent-to-add entries the same way. Same two-line move as in commit_files?

[Alex Shabalin (alex-sparus)]

Severity 2 — not blocking

Now that paths here are dir-relative, restore_hunk/commit_hunk are only correct with dir = repo root: find_hunk_by_hash and file_diff_contents still match workdir-relative names, so a future caller passing a nested config dir would look a hunk up under one convention and resolve the file under the other. Worth a one-line comment on these two functions ("callers must pass the repo root until the query side is dir-relative too") so the first nested-dir caller finds out at review time instead of at runtime?

[Juanpe Bolívar (arximboldi)]

Oh damn! I guess I missed #655 (comment) before writing my review comment in #655 ... I think this probably solves my issues! Looking forward to re-try once this is in :)