paperlint 3.1.2 → 3.1.3

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "paperlint",
-  "version": "3.1.2",
+  "version": "3.1.3",
   "description": "A linter for scientific papers written in LaTeX: catches mistakes before you submit to a conference or journal.",
   "type": "module",
   "license": "MIT",

package/skills/verify-citations/scripts/verify-cites.mjs

@@ -156,13 +156,23 @@ export function normalizeArxiv(arxiv) {
     .trim();
 }
 
+/** arXiv registers a DataCite DOI for every paper: `10.48550/arXiv.<id>`. */
+const ARXIV_DOI = /^10\.48550\/arxiv\.(.+)$/i;
+
 /** Shallow copy of a citation with doi/arxiv identifiers normalized. Applied at the
  *  verify boundary so BOTH JSON input and .bib-parsed input are clean before any
- *  network / authority lookup. */
+ *  network / authority lookup.
+ *
+ *  An arXiv DOI also yields the arXiv id it names, when the entry has none of its own: without
+ *  it arXiv is asked only by title, and only for an entry with no DOI, so a registry that maps
+ *  the DOI to the wrong work goes uncontradicted (OpenAlex did, for SWE-bench, on 2026-09-27). */
 export function normalizeIdentifiers(citation) {
   const c = { ...citation };
   if (c.doi) c.doi = normalizeDoi(c.doi);
   if (c.arxiv) c.arxiv = normalizeArxiv(c.arxiv);
+  // Decided on the NORMALIZED id: a blank field normalizes to "" and must not block the DOI's id.
+  const fromDoi = c.doi ? ARXIV_DOI.exec(c.doi)?.[1] : undefined;
+  if (!c.arxiv && fromDoi) c.arxiv = fromDoi;
   return c;
 }
 

package/skills/verify-citations/scripts/verify-cites.net.test.mjs

@@ -271,6 +271,41 @@ test("a corrected arXiv id is asked again even when the DOI is unchanged", async
   );
 });
 
+// Guards: an arXiv DOI (10.48550/arXiv.<id>) names an arXiv id, so arXiv is asked by that id.
+// Measured 2026-09-27 on a real bibliography: OpenAlex maps 10.48550/arXiv.2310.06770 (SWE-bench)
+// to an unrelated work; with Crossref and Semantic Scholar refusing and arXiv asked only by title
+// "when there is no DOI", nothing contradicted it and the paper came out `exists: false`.
+test("an arXiv DOI is asked of arXiv by its id, whose match outweighs a registry's wrong record", async () => {
+  const calls = fakeFetch({
+    [CR]: () => json(429),
+    [OA]: () =>
+      json(200, {
+        id: "W1",
+        display_name: "GardenBench",
+        publication_year: 2023,
+      }),
+    [S2]: () => json(429),
+    [AX]: () => text(200, feed(TITLE, 2017)),
+    [DOI]: () => json(200, { responseCode: 1 }),
+  });
+  const v = await verifyCitationLive(
+    { id: "c", doi: "10.48550/arXiv.1706.03762", title: TITLE, year: "2017" },
+    { cache: {} },
+  );
+  assert.deepEqual(
+    {
+      verdict: v.verdict,
+      matched: v.matched_db,
+      arxiv: calls.filter((u) => u.startsWith(AX)),
+    },
+    {
+      verdict: "true",
+      matched: "arxiv",
+      arxiv: [`${AX}?id_list=1706.03762&max_results=1`],
+    },
+  );
+});
+
 // Guards: NVD's "not found" must not be cached. A CVE cited before NVD indexes it would otherwise
 // stay "fabricated" in the committed cache forever; only positive existence is kept, as for doi.org.
 test("an NVD not-found answer is not cached, so a later build asks again", async () => {

package/skills/verify-citations/scripts/verify-cites.test.mjs

@@ -738,6 +738,20 @@ console.log("\n[S4] .bib parser edge cases:");
 
 // ── Edges of the pure layer: the defensive arms, each with the input that reaches it ──────────
 
+// Guards: the arXiv id is read out of an arXiv DOI whenever the entry has no USABLE arXiv id, so
+// a blank `arxiv` field (JSON input) does not keep arXiv from being asked.
+test.each([
+  [{ doi: "10.48550/arXiv.2310.06770" }, "2310.06770"],
+  [{ doi: "10.48550/arXiv.2310.06770", arxiv: " " }, "2310.06770"],
+  [
+    { doi: "10.48550/arXiv.2310.06770", arxiv: "arXiv:2401.00001" },
+    "2401.00001",
+  ],
+  [{ doi: "10.1145/3548606" }, undefined],
+])("normalizeIdentifiers(%o).arxiv is %s", (citation, arxiv) => {
+  assert.equal(normalizeIdentifiers(citation).arxiv || undefined, arxiv);
+});
+
 test("levenshtein, normalizeTitle and titleIdentity take a missing string as empty", () => {
   assert.deepEqual(
     [