This is the record of an abcd-managed repository

abcdBetav0.16.2 · 2026-11-06

Development record

The write boundary now refuses a memory page whose filename carries a hard-fail secret, but a store written before that change is not repaired and the read side cannot report it. The MR001 lint scans stored text and the registry bytes, and a token embedded in a page name such as topic_auth_ghp_ is hidden from it by the same missing word boundary the write-side fix had to work around: underscore is a word character, so an anchored token pattern finds no boundary after the type and domain prefix and the span never matches. A legacy dirty filename therefore sits in the committed tree, in index.md, in log.md and in the registry back-link, and no instrument reports it. The write-side fix deliberately left this alone to keep its commit to one record. Fix: give the read-side scan the same component-wise treatment the write side uses, splitting a page name into its type, domain and slug parts before matching rather than relying on a boundary the naming grammar cannot provide. Detector: a store seeded with a page whose name embeds a token must be reported by the read-side lint, and an ordinary store must stay clean. Surfaced by the fix for iss-2609020321100138.

issueresolvedminoriss-2609090642035097

The write boundary now refuses a memory page whose filename carries a hard-fail secret, but a store written before that change is not repaired and the read side cannot report it. The MR001 lint scans stored text and the registry bytes, and a token embedded in a page name such as topic_auth_ghp_ is hidden from it by the same missing word boundary the write-side fix had to work around: underscore is a word character, so an anchored token pattern finds no boundary after the type and domain prefix and the span never matches. A legacy dirty filename therefore sits in the committed tree, in index.md, in log.md and in the registry back-link, and no instrument reports it. The write-side fix deliberately left this alone to keep its commit to one record. Fix: give the read-side scan the same component-wise treatment the write side uses, splitting a page name into its type, domain and slug parts before matching rather than relying on a boundary the naming grammar cannot provide. Detector: a store seeded with a page whose name embeds a token must be reported by the read-side lint, and an ordinary store must stay clean. Surfaced by the fix for iss-2609020321100138.

Grounds

  • pursued: sharing filenameHardFailKinds between judgeFilename and the lint makes a name the write side would refuse exactly a name the lint reports; a seeded store whose page name embeds a token going unreported, or an ordinary prose-shaped name drawing a page-name finding, would show it wrong

Frontmatter

idiss-2609090642035097
typeissue
lifecycleresolved
severityminor

mentions1