Author: nick Date: Mon Dec 13 02:41:59 2010 New Revision: 1045006 URL: http://svn.apache.org/viewvc?rev=1045006&view=rev Log: Apply patch from TIKA-570 from Benson Margulies - stricter BMP detection and unit test
Added: tika/trunk/tika-parsers/src/test/resources/test-documents/testBMPfp.txt Modified: tika/trunk/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml tika/trunk/tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java Modified: tika/trunk/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml URL: http://svn.apache.org/viewvc/tika/trunk/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml?rev=1045006&r1=1045005&r2=1045006&view=diff ============================================================================== --- tika/trunk/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml (original) +++ tika/trunk/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml Mon Dec 13 02:41:59 2010 @@ -3125,7 +3125,17 @@ <acronym>BMP</acronym> <_comment>Windows bitmap</_comment> <magic priority="50"> - <match value="BM" type="string" offset="0" /> + <match value="BM" type="string" offset="0"> + <match value="0x0100" type="string" offset="26"> + <match value="0x0000" type="string" offset="28"/> + <match value="0x0100" type="string" offset="28"/> + <match value="0x0400" type="string" offset="28"/> + <match value="0x0800" type="string" offset="28"/> + <match value="0x1000" type="string" offset="28"/> + <match value="0x1800" type="string" offset="28"/> + <match value="0x2000" type="string" offset="28"/> + </match> + </match> </magic> <glob pattern="*.bmp"/> <glob pattern="*.dib"/> Modified: tika/trunk/tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java URL: http://svn.apache.org/viewvc/tika/trunk/tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java?rev=1045006&r1=1045005&r2=1045006&view=diff ============================================================================== --- tika/trunk/tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java (original) +++ tika/trunk/tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java Mon Dec 13 02:41:59 2010 @@ -191,6 +191,8 @@ public class TestMimeTypes extends TestC assertTypeByName("image/x-ms-bmp", "x.BMP"); assertTypeByName("image/x-ms-bmp", "x.dib"); assertTypeByName("image/x-ms-bmp", "x.DIB"); + //false positive check -- contains part of BMP signature + assertType("text/plain", "testBMPfp.txt"); } public void testPnmDetection() throws Exception { Added: tika/trunk/tika-parsers/src/test/resources/test-documents/testBMPfp.txt URL: http://svn.apache.org/viewvc/tika/trunk/tika-parsers/src/test/resources/test-documents/testBMPfp.txt?rev=1045006&view=auto ============================================================================== --- tika/trunk/tika-parsers/src/test/resources/test-documents/testBMPfp.txt (added) +++ tika/trunk/tika-parsers/src/test/resources/test-documents/testBMPfp.txt Mon Dec 13 02:41:59 2010 @@ -0,0 +1,3 @@ +BMW to Make Hybrid Sports Carâ© + By CHRISTOPH RAUWALD . +LEIPZIG, GermanyâGerman car maker BMW AG said Friday it will start series production of a new plug-in hybrid sports car in 2013, to be based on the Vision EfficientDynamics Concept car shown at the Frankfurt auto show in September last year. Chief Executive Norbert Reithofer said the car will be produced in Germany but didn't provide details on the price. The BMW Vision EfficientDynamics Concept car is a sporty plug-in, full hybrid with a turbo-diesel engine, four seats and upward-pivoting doors. BMW executive board member Klaus Draeger told reporters he expects to achieve "a significant sales volume" with the new high-performance sports car. Asked whether annual sales could exceed 1,000 vehicles, Mr. Draeger said, "You said this and I'm not saying this is wrong." In March, Mr. Reithofer indicated that the concept car was set to make it into series production. "I like the car. And you know what it means when I say I like the carâit means I will drive it. It's not j ust a concept car," he told analysts during a presentation in Munich. The car will be designed for sale in all major global markets, which according to Mr. Draeger might require offering a gasoline engine instead of the prototype's three-cylinder diesel engine. Diesel cars account for roughly half of the European market, but are significantly less popular in the U.S. and hardly present at all in China. Mr. Draeger declined to comment on the vehicle's price tag, but noted that in order to achieve substantial sales volumes the price mustn't be too high. He said the same goes for BMW's planned Megacity Vehicle. A price tag of â¬60,000 ($85,242) or more would certainly limit potential sales volumes, he said.