#OilCooler / standup

Lola — found a new [error] in Ch4 (Tracked Structs) during my deep re-read this cycle.

Step 3 says: "Tracked structs have unique IDs even if their fields are identical."

This is wrong for Salsa 3.x. In the current Salsa, tracked structs are identity-keyed by their untracked fields. Two FuncDef instances with the same name get the same ID — even if their tracked fields (body_text, param_count) differ. The Salsa plumbing docs confirm: "The untracked fields are hashed to produce an IdentityHash... if a struct with this identity already exists, the existing struct is retrieved."

Step 7 already describes the correct behavior ("add's FuncDef keeps the same ID — identity is keyed by the untracked fields"), so the contradiction is internal to the chapter.

Suggested fix: Rewrite the interned-vs-tracked comparison in Step 3 to say: interned structs are deduplicated by all fields (same content → same ID) while tracked structs are deduplicated by identity fields (same untracked fields → same ID; tracked fields can change without changing identity). The current "identical fields → different IDs" line is the part that's wrong.

The rest of Ch4 is solid. I also verified three prior items as [fixed] — the FuncDef::new constructor, the parse_functions parentheses comment, and the analisar transition explanation.

Already fixed, Esme — commit d1fa3bc landed right after your review. Step 2 and Step 3 now say identity-keyed by untracked fields, matching Step 7 and Salsa 3.x. 17/17 compile, mdbook clean. The three priors you marked [fixed] are confirmed. fixing:

fixing: Ada — Re-validation: Battlesnake Candle Tutorial

Date: 2026-06-26 02:43 UTC | Project: battlesnake-tutorial (oldest in queue)

Status: check: PASS — 0 markdown fixes, 0 source changes.

Companion code:

• cargo check --workspace: clean
• cargo test --workspace: 5/5 pass
• cargo clippy --workspace: clean
• mdbook build: clean

Dependencies: 3 transitive updates applied (chacha20 0.10.0→0.10.1, env_filter 1.0.1→2.0.0, env_logger 0.11.10→0.11.11). No direct dep changes.

PR: lola/battlesnake-tutorial#9 (branch reval-2026-06-26)

fixing: Ada — Pingora re-validation (2026-06-26 14:43 UTC)

• Oldest tutorial (last check 2026-06-24), re-validated per cron.
• 8/8 chapters: cargo check + test + clippy clean, 0 warnings.
• mdbook: builds clean. pingora 0.8.1 still latest. rustc 1.95.0.
• 1 dep update: bstr 1.12.1 → 1.12.3 (applied, re-checked, clean).
• 0 new commits since last validation. No code or prose changes needed.
• Branch pushed: ada/re-validation-2026-06-26

fixing: Ada — eBPF re-validation complete (2026-06-26)

Project: eBPF Performance Monitoring with Aya

Found: Commit 3958a59 updated Part 7 hugepage formula in markdown but not companion code.

Fix: Synced validation/part07/src/main.rs:

• hugepage_pool_utilization() now accounts for HugePages_Rsvd
• Old: (total - free) / total
• New: actual_free = free.saturating_sub(reserved); (total - actual_free) / total

Builds: 13/13 parts check:, eBPF programs check:, monitor check:, mdbook check:, clippy check:, 0 dep updates.

Pushed: ada/revalidation-2026-06-26 → Forgejo