Skip to content

chore(backlog): file and triage RPT-010 height precision report - #18

Merged
jmlweb merged 1 commit into
mainfrom
hustle/maintainer/backlog-2026-11-17
Nov 17, 2026
Merged

jmlweb merged 1 commit into
mainfrom
hustle/maintainer/backlog-2026-11-17

Conversation

@jmlweb

@jmlweb jmlweb commented Nov 17, 2026

Copy link
Copy Markdown
Owner

Backlog grooming run, 2026-11-17. Only .backlog/ files change, and every change went through the flowstate CLI.

Checks run on main: pnpm typecheck passes. pnpm test passes (135 passed, 5 skipped). The CI config (ci.yml, deploy.yml) runs typecheck, tests and build.

Reports filed

  • RPT-010 (bug, low): Tower config accepts heights the STL name cannot encode, which breaks the lossless round trip.
    • validateTowerConfig (packages/core/src/tower-config.ts:283-284) puts no precision bound on block or base height.
    • formatNumber (packages/core/src/filename-codec.ts:11) rounds every value to 0.001 mm.
    • Reproduction: blockHeight: 10.0004 passes validation and encodes as h10. It decodes as 10, and towerHeight comes out 91.404 vs 91.4.
    • The web form reaches this path: its only check is validateTowerConfig.
    • The round-trip property test (filename-codec.test.ts:12-13) only draws heights in 0.1 mm steps, so it misses this.
    • Why it qualifies: a missing validation at a boundary that contradicts ADR-0005's documented lossless encode/decode.
    • Why low: at realistic layer heights, tier placement and the height check do not change.

Triage

  • RPT-010 → TSK-017 (P4): low severity maps to P4. The fix is clear: bound height precision in validation, and widen the round-trip arbitrary to the full domain validation accepts.

Pending reports before this run

None.

Idea reviews

None. .backlog/ideas/pending/ is empty.

🤖 Generated with Claude Code

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@jmlweb

jmlweb commented Nov 17, 2026

Copy link
Copy Markdown
Owner Author

🤖 hustle review: approve — Only .backlog/ files change. RPT-010 cites file:line locations and a reproduction with observed output, and the P4 triage to TSK-017 is justified by the low severity. The index count matches.

@jmlweb
jmlweb merged commit f510c40 into main Nov 17, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant