Index: tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java =================================================================== --- tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java (revision 1213232) +++ 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 1213232) +++ 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 @@ -128,7 +128,7 @@ public void testOLE2Detection() throws Exception { // These have the properties block near the start, so our mime // magic will spot them - assertTypeByData("application/vnd.ms-excel", "testEXCEL.xls"); + assertTypeByData("application/x-tika-msoffice", "testEXCEL.xls"); // This one quite legitimately doesn't have its properties block // as one of the first couple of entries @@ -136,6 +136,7 @@ assertTypeByData("application/x-tika-msoffice", "testWORD.doc"); assertTypeByData("application/x-tika-msoffice", "testPPT.ppt"); + assertTypeByData("application/x-tika-msoffice", "testWORKSSpreadsheet7.0.xlr"); // By name + data: @@ -146,6 +147,8 @@ // ones to their actual type assertTypeByNameAndData("application/vnd.ms-powerpoint", "testPPT.ppt"); assertTypeByNameAndData("application/msword", "testWORD.doc"); + + assertTypeByNameAndData("application/vnd.ms-works", "testWORKSSpreadsheet7.0.xlr"); } /** 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 1213232) +++ tika-parsers/src/main/java/org/apache/tika/parser/microsoft/POIFSContainerDetector.java (working copy) @@ -133,7 +133,9 @@ */ protected static MediaType detect(Set names) { if (names != null) { - if (names.contains("Workbook")) { + if (names.contains("WksSSWorkBook")) { + return WPS; // a Works 7.0 spreadsheet file + } if (names.contains("Workbook")) { return XLS; } else if (names.contains("EncryptedPackage") && names.contains("EncryptionInfo") && Index: tika-core/src/test/java/org/apache/tika/mime/MimeDetectionTest.java =================================================================== --- tika-core/src/test/java/org/apache/tika/mime/MimeDetectionTest.java (revision 1213232) +++ tika-core/src/test/java/org/apache/tika/mime/MimeDetectionTest.java (working copy) @@ -174,13 +174,35 @@ /** * Tests that when we repeatedly test the detection of a document - * that can be detected with Mime Magic, that we consistently - * detect it correctly. See TIKA-391 for more details. + * that can be detected with Mime Magic, that we consistently + * detect it correctly. See TIKA-391 for more details. + * + * In TIKA-806 I removed the patterns, which allowed the test.xls + * file to be detected based on the magic only. Since then, the + * expected media type for this file is application/x-tika-msoffice + * when detected without the RESOURCE_NAME hint, and the normal + * Excel media type when detected with the hint. */ public void testMimeMagicStability() throws IOException { + for(int i=0; i<100; i++) { - testFile("application/vnd.ms-excel", "test.xls"); + InputStream in = getClass().getResourceAsStream("test.xls"); + assertNotNull("Test stream: [test.xls] is null!", in); + if (!in.markSupported()) { + in = new java.io.BufferedInputStream(in); + } + try { + Metadata metadata = new Metadata(); + String mime = this.mimeTypes.detect(in, metadata).toString(); + assertEquals("test.xls is not properly detected: detected.", "application/x-tika-msoffice", mime); + + //Add resource name and test again + metadata.set(Metadata.RESOURCE_NAME_KEY, "test.xls"); + mime = this.mimeTypes.detect(in, metadata).toString(); + assertEquals("test.xls is not properly detected after adding resource name.", "application/vnd.ms-excel", mime); + } finally { + in.close(); + } } } - } 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 1213232) +++ tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml (working copy) @@ -264,9 +264,6 @@ - - - @@ -1213,9 +1210,6 @@ - - - @@ -1283,11 +1277,6 @@ <_comment>Microsoft Powerpoint Presentation - - - - - @@ -1359,6 +1348,7 @@ +