Index: tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java =================================================================== --- tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java (revision 1213741) +++ tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java (working copy) @@ -70,6 +70,7 @@ assertTypeByData("testPUBLISHER.pub", "application/x-mspublisher"); assertTypeByData("testWORKS.wps", "application/vnd.ms-works"); assertTypeByData("testWORKS2000.wps", "application/vnd.ms-works"); + assertTypeByData("testWORKSSpreadsheet7.0.xlr", "application/vnd.ms-works"); assertTypeByData("testPROJECT2003.mpp", "application/vnd.ms-project"); assertTypeByData("testPROJECT2007.mpp", "application/vnd.ms-project"); Index: tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java =================================================================== --- tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java (revision 1213741) +++ tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java (working copy) @@ -1,5 +1,5 @@ /* - * Licensed to the Apache Software Foundation (ASF) under one or more +* Licensed to the Apache Software Foundation (ASF) under one or more * contributor license agreements. See the NOTICE file distributed with * this work for additional information regarding copyright ownership. * The ASF licenses this file to You under the Apache License, Version 2.0 @@ -27,7 +27,9 @@ import org.apache.tika.Tika; import org.apache.tika.config.TikaConfig; +import org.apache.tika.detect.DefaultDetector; import org.apache.tika.metadata.Metadata; +import org.apache.tika.parser.microsoft.POIFSContainerDetector; /** * @@ -149,6 +151,33 @@ } /** + * Files generated by Works 7.0 Spreadsheet application use the OLE2 + * structure and resemble Excel files (they contain a "Workbook"). They are + * not Excel though. The {@link POIFSContainerDetector} can detect them + * properly. With plain {@link MimeTypes} they are detected as Excel, + * because of the "Workbook" string. It's a problem we discussed in TIKA-806 + * and agreed that we live with that. The policy is that container-based + * detection should trump magic-based detection. It's implemented in + * {@link DefaultDetector} (TIKA-786) and users who don't want to use to + * {@link DefaultDetector} should be aware of it. + * + * @throws Exception + */ + public void testWorks70Detection() throws Exception { + // this is wrong, but we live with that + assertTypeByData("application/vnd.ms-excel", + "testWORKSSpreadsheet7.0.xlr"); + + // this is right, we made x-xlr a subtype of vnd.ms-excel + assertTypeByNameAndData("application/x-xlr", + "testWORKSSpreadsheet7.0.xlr"); + + // with name-only, everything should be all right + assertTypeByName("application/x-xlr", + "testWORKSSpreadsheet7.0.xlr"); + } + + /** * Note - detecting container formats by mime magic is very very * iffy, as we can't be sure where things will end up. * People really ought to use the container aware detection... Index: tika-parsers/src/test/java/org/apache/tika/parser/microsoft/ExcelParserTest.java =================================================================== --- tika-parsers/src/test/java/org/apache/tika/parser/microsoft/ExcelParserTest.java (revision 1213741) +++ tika-parsers/src/test/java/org/apache/tika/parser/microsoft/ExcelParserTest.java (working copy) @@ -192,5 +192,22 @@ input.close(); } } + + public void testWorksSpreadsheet70() throws Exception { + InputStream input = ExcelParserTest.class.getResourceAsStream( + "/test-documents/testWORKSSpreadsheet7.0.xlr"); + try { + Metadata metadata = new Metadata(); + ContentHandler handler = new BodyContentHandler(-1); + ParseContext context = new ParseContext(); + context.set(Locale.class, Locale.US); + new OfficeParser().parse(input, handler, metadata, context); + + String content = handler.toString(); + assertTrue(content.contains("Microsoft Works")); + } finally { + input.close(); + } + } } Index: tika-parsers/src/main/java/org/apache/tika/parser/microsoft/OfficeParser.java =================================================================== --- tika-parsers/src/main/java/org/apache/tika/parser/microsoft/OfficeParser.java (revision 1213741) +++ tika-parsers/src/main/java/org/apache/tika/parser/microsoft/OfficeParser.java (working copy) @@ -69,6 +69,7 @@ POIFSDocumentType.PROJECT.type, POIFSDocumentType.VISIO.type, // Works isn't supported + POIFSDocumentType.XLR.type, // but Works 7.0 Spreadsheet is POIFSDocumentType.OUTLOOK.type, MediaType.application("vnd.ms-excel.sheet.binary.macroenabled.12") ))); @@ -84,6 +85,7 @@ PROJECT("mpp", MediaType.application("vnd.ms-project")), VISIO("vsd", MediaType.application("vnd.visio")), WORKS("wps", MediaType.application("vnd.ms-works")), + XLR("xlr", MediaType.application("x-tika-msworks-spreadsheet")), OUTLOOK("msg", MediaType.application("vnd.ms-outlook")); private final String extension; @@ -186,6 +188,7 @@ new HSLFExtractor(context).parse(root, xhtml); break; case WORKBOOK: + case XLR: Locale locale = context.get(Locale.class, Locale.getDefault()); new ExcelExtractor(context).parse(root, xhtml, locale); break; Index: tika-parsers/src/main/java/org/apache/tika/parser/microsoft/POIFSContainerDetector.java =================================================================== --- tika-parsers/src/main/java/org/apache/tika/parser/microsoft/POIFSContainerDetector.java (revision 1213741) +++ tika-parsers/src/main/java/org/apache/tika/parser/microsoft/POIFSContainerDetector.java (working copy) @@ -71,6 +71,9 @@ /** Microsoft Works */ public static final MediaType WPS = application("vnd.ms-works"); + + /** Microsoft Works Spreadsheet 7.0 */ + public static final MediaType XLR = application("x-tika-msworks-spreadsheet"); /** Microsoft Outlook */ public static final MediaType MSG = application("vnd.ms-outlook"); @@ -133,7 +136,9 @@ */ protected static MediaType detect(Set names) { if (names != null) { - if (names.contains("Workbook")) { + if (names.contains("WksSSWorkBook")) { + return XLR; // a Works 7.0 spreadsheet file + } if (names.contains("Workbook")) { return XLS; } else if (names.contains("EncryptedPackage") && names.contains("EncryptionInfo") && Index: tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml =================================================================== --- tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml (revision 1213741) +++ tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml (working copy) @@ -2810,6 +2810,12 @@ + + + + + +