[kurtnordstrom]

…ys above highest imported value

[Copilot]

🟡 Changes recommended

Critical ID-collision risks remain because concurrent imports can regress the sequence and hyphenated prefixes silently bypass sequence advancement.

2 open issues

• Advance sequences for hyphenated request ID prefixes · New
• Serialize concurrent imported-ID sequence advances · New

What changed in this PR

This PR updates mod-rs migration handling to normalize request-ID prefixes and advance CrossLink’s HRID counters when importing borrowing requests.

Changes:

• Remove punctuation and whitespace from exported request-ID prefixes.
• Accept PREFIX-N HRIDs and advance the corresponding broker sequence during import.
• Update migration documentation and regression tests.

File	Description
migration/​scripts.md	Documents prefix normalization and import ordering.
migration/​export-crosslink-open-patron-requests.sql	Allows numeric HRIDs with a hyphenated separator.
migration/​export-crosslink-directory.sql	Normalizes prefixes before validation and export.
migration/​export-crosslink-directory-field-mapping.md	Updates prefix mapping rules.
directory/​test/​request_id_prefix_test.go	Tests normalized prefixes and invalid values.
broker/​sqlc/​pr_schema.sql	Adds the sequence-advancement function to the schema.
broker/​sqlc/​import_query.sql	Exposes sequence advancement through SQLC.
broker/​migrations/​080_advance_hrid.up.sql	Installs the sequence-advancement function.
broker/​migrations/​080_advance_hrid.down.sql	Removes the function on rollback.
broker/​import/​test/​patron_request_test.go	Tests borrowing sequence advancement and lending exclusion.
broker/​import/​test/​export_hrid_preflight_test.go	Updates accepted and rejected HRID cases.
broker/​import/​db/​repo.go	Parses imported borrowing HRIDs and advances their sequences.

_{Review effort: Balanced}

[Copilot]

🔵 Needs a closer look

Locale-dependent character stripping can allow non-ASCII prefixes to bypass validation, producing inconsistent migration results across database locales.

No open issues 🎉

1 previously missed issue

In code that hasn't changed since last review

Make punctuation and whitespace removal locale-independent

migration/​export-crosslink-directory.sql:334

PostgreSQL's [:punct:] and [:space:] classes are locale-dependent, not ASCII-only. Under a Unicode-aware locale, a prefix such as ABC— can have its non-ASCII dash removed and pass the subsequent alphanumeric validation, whereas a C-locale database rejects it. This makes migration results depend on the source database's locale and bypasses the documented rejection of non-ASCII characters. Apply the C collation to this replacement so only ASCII punctuation and whitespace are removed.

2 resolved since last review

• Serialize concurrent imported-ID sequence advances
• Advance sequences for hyphenated request ID prefixes

_{Review effort: Balanced}

[Copilot]

🔵 Needs a closer look

Human assessment is needed because a Go-formatting finding passed the severity filter but was omitted from the published review.

No open issues 🎉

_{Review effort: Balanced}