feat: add /pace dashboard tracking artifact output vs Year 1 target

[byte-the-bot]

Review Brief

What changed and why

Adds a public /pace dashboard that reads blog post and note frontmatter to track artifact output toward two Year 1 targets (48 Battlesnake, 12 Other posts by June 1 2027). Introduces a typed Track enum field on both BlogFrontMatter and notes::FrontMatter, a new server-side handler with pure stat helpers (in_window, compute_track_stats, render_track_section), and backfills track: battlesnake on 4 existing posts. No DB writes, no migrations — everything recomputed on each request from AppState.

Architecture decisions and trade-offs

• Track is a closed enum with serde(rename_all = "lowercase") — unknown values like track: typo hard-error at parse time rather than silently defaulting. Intentional per spec ("typed, not free-form").
• Window dates use static LazyLock<NaiveDate> rather than const — NaiveDate::from_ymd_opt is not const fn in chrono 0.4.41; const here would be a compile error (E0015).
• projected_finish is suppressed for the first 28 days (PROJECTION_MIN_DAYS) to avoid misleading projections when the sample is tiny (e.g., one post → ~365 projected).
• Tailwind classes are full literal strings, never format!-constructed — required so the Tailwind purge scanner includes them in target/tailwind.css.
• Track iteration uses an array literal [Track::Battlesnake, Track::Other], not vec![…] — the useless_vec pedantic lint forbids the heap-allocated form.

Risk assessment

• Blast radius: Low. The track field is additive and defaults to Other — no existing frontmatter breaks. Backfill is 4 files, all pre-dating WINDOW_START, so they affect no live count. New route only.
• Confidence level: High. Pure helper functions have comprehensive unit tests covering all edge cases from the plan.
• Rollback safety: Clean revert — no schema changes, no migrations, no data mutations.

Spot-check suggestions

• server/src/http_server/pages/pace.rs:render_track_section — verify the expected-by-now tick (div at left: {tick_pct:.2}%) is a sibling of the fill bar, not a child; both should be children of the outer bg-neutral-700 rounded container.
• posts/src/blog.rs:Track — confirm #[serde(rename_all = "lowercase")] makes battlesnake (lowercase) the valid value, and all 4 backfilled files use exactly track: battlesnake (not track: Battlesnake).
• server/src/http_server/pages/pace.rs:collect_artifacts — notes.posts is accessed as a direct field; verify NotePosts.posts is pub (not pub(crate) or private) from the server crate's perspective.
• server/src/http_server/pages/pace.rs:compute_track_stats — artifacts passed in are already in_window-filtered; confirm the per-track filter(|a| a.track == track) is applied for both the count and the track_artifacts vec, so compute_track_stats_filters_only_requested_track is testing real behavior.

What the agent verified

• cargo fmt --check ✅
• cargo clippy --all-targets --all-features --workspace --tests ✅ (clean under pedantic deny)
• cargo test -p posts ✅
• All new pace, track, and nav_includes_pace_link tests ✅
• Pre-existing #[sqlx::test] DB-backed tests fail locally (Postgres creds mismatch) — unrelated to this change, expected to pass in CI

---

Summary

Implements the public /pace dashboard (task BLOG-84ffd942c4f94ead): a cumulative-target-with-graceful-failure progress tracker for Year 1 artifact output, through June 1 2027.

The dashboard reads existing blog post and note frontmatter, classifies each published artifact into one of two tracks (battlesnake or other), and renders server-side stats per track:

• count vs target (Battlesnake 48, Other 12)
• expected-by-now marker on the progress bar (elapsed_fraction * target)
• on/off-pace verdict (green / amber)
• need-per-week to still hit target (featured prominently)
• weeks left to the window end
• projected finish — greyed and suppressed until ~week 4 (PROJECTION_MIN_DAYS = 28), since it lies early
• chronological artifact list (date + link, newest first) doubling as public proof-of-work

Everything is recomputed on each request from AppState — no DB writes, no migrations, no manual entry.

Changes

• posts/src/blog.rs — new Track enum + track frontmatter field (defaults to other)
• posts/src/notes.rs — mirrored track field on note frontmatter
• server/src/http_server/pages/pace.rs — new dashboard page (handler, pure stat helpers, render, unit tests)
• route /pace, module registration, and a "Pace" nav link (with regression test)
• backfilled track: battlesnake on the four existing Battlesnake posts

Testing

• cargo fmt --check ✅
• cargo clippy --all-targets --all-features --workspace --tests ✅ (clean)
• cargo test -p posts ✅
• New pace, track, and nav_includes_pace_link tests ✅

Pre-existing #[sqlx::test] DB-backed tests (sponsors, memories, thread_processor, thread builder) fail locally only due to the sandbox's Postgres credentials (password authentication failed for user "coreyja") — unrelated to this change, which touches no SQL. They should pass in CI.

🤖 Generated with Claude Code