Fall back to literal string if page field contains an illegal number

Fixes #114
michel-kraemer · Jan 5, 2022 · c12a33c · c12a33c
1 parent 2c6a4de
commit c12a33c
Show file tree

Hide file tree

Showing 2 changed files with 32 additions and 2 deletions.
diff --git a/citeproc-java/src/main/java/de/undercouch/citeproc/bibtex/PageParser.java b/citeproc-java/src/main/java/de/undercouch/citeproc/bibtex/PageParser.java
@@ -25,8 +25,13 @@ public static PageRange parse(String pages) {
         CommonTokenStream tokens = new CommonTokenStream(lexer);
         InternalPageParser parser = new InternalPageParser(tokens);
         parser.removeErrorListeners(); // do not output errors to console
-        PagesContext ctx = parser.pages();
-        if (ctx.literal == null || ctx.literal.isEmpty() ||
+        PagesContext ctx;
+        try {
+            ctx = parser.pages();
+        } catch (NumberFormatException e) {
+            ctx = null;
+        }
+        if (ctx == null || ctx.literal == null || ctx.literal.isEmpty() ||
                 ctx.exception != null || parser.getNumberOfSyntaxErrors() > 0) {
             // unparsable fall back to literal string
             return new PageRange(pages, null, null, false);

diff --git a/citeproc-java/src/test/java/de/undercouch/citeproc/bibtex/BibTeXConverterTest.java b/citeproc-java/src/test/java/de/undercouch/citeproc/bibtex/BibTeXConverterTest.java
@@ -210,4 +210,29 @@ public void inBookSeries() throws ParseException {
         assertEquals("978-3-030-83014-4", item.getISBN());
         assertEquals("10.1007/978-3-030-83014-4_2", item.getDOI());
     }
+
+    /**
+     * Check if the parser correctly falls back to a literal string if the
+     * 'pages' field contains an illegal number.
+     * See https://github.com/michel-kraemer/citeproc-java/issues/114
+     */
+    @Test
+    public void invalidPageNumber() throws ParseException {
+        String entry = "@Article{baks-2021,\n" +
+                "  author           = {Sandipan Baksi},\n" +
+                "  date             = {2021},\n" +
+                "  journaltitle     = {The Indian Economic {\\&} Social History Review},\n" +
+                "  pages            = {001946462110645},\n" +
+                "  title            = {Science journalism in Hindi in pre-independence India: A study of Hindi periodicals},\n" +
+                "  doi              = {10.1177/00194646211064586},\n" +
+                "  creationdate     = {2022-01-03T11:59:38},\n" +
+                "  modificationdate = {2022-01-03T12:01:49},\n" +
+                "  publisher        = {{SAGE} Publications},\n" +
+                "}";
+        BibTeXDatabase db = new BibTeXParser().parse(new StringReader(entry));
+        BibTeXConverter converter = new BibTeXConverter();
+        Map<String, CSLItemData> items = converter.toItemData(db);
+        CSLItemData item = items.get("baks-2021");
+        assertEquals("001946462110645", item.getPage());
+    }
 }