Author: nick
Date: Thu May 10 11:37:13 2012
New Revision: 1336610
URL: http://svn.apache.org/viewvc?rev=1336610&view=rev
Log:
TIKA-913 Mime Magic for PE, PE32 and PE64 executables
Modified:
tika/trunk/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml
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=1336610&r1=1336609&r2=1336610&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
Thu May 10 11:37:13 2012
@@ -2677,13 +2677,83 @@
<mime-type type="application/x-msclip">
<glob pattern="*.clp"/>
</mime-type>
+
<mime-type type="application/x-msdownload">
<glob pattern="*.exe"/>
<glob pattern="*.dll"/>
<glob pattern="*.com"/>
<glob pattern="*.bat"/>
<glob pattern="*.msi"/>
+ <magic priority="50">
+ <match value="MZ" type="string" offset="0"/>
+ </magic>
+ </mime-type>
+ <mime-type type="application/x-msdownload;format=pe">
+ <sub-class-of type="application/x-msdownload"/>
+ <magic priority="55">
+ <!-- Technically the header offset is stored at 0x3c, and isn't a -->
+ <!-- constant, but it's almost always set to start at 0x80 or 0xf0 -->
+ <match value="PE\000\000" type="string" offset="128"/>
+ <match value="PE\000\000" type="string" offset="240"/>
+ </magic>
+ </mime-type>
+ <!-- the PE header should be PEx00x00 then a two byte machine type -->
+ <mime-type type="application/x-msdownload;format=pe32">
+ <sub-class-of type="application/x-msdownload;format=pe"/>
+ <magic priority="60">
+ <match value="PE\000\000" type="string" offset="128">
+ <match value="0x014c" type="little16" offset="132"/>
+ </match>
+ <match value="PE\000\000" type="string" offset="240">
+ <match value="0x014c" type="little16" offset="244"/>
+ </match>
+ </magic>
+ </mime-type>
+ <mime-type type="application/x-msdownload;format=pe64">
+ <sub-class-of type="application/x-msdownload;format=pe"/>
+ <magic priority="60">
+ <match value="PE\000\000" type="string" offset="128">
+ <match value="0x8664" type="little16" offset="132"/>
+ </match>
+ <match value="PE\000\000" type="string" offset="240">
+ <match value="0x8664" type="little16" offset="244"/>
+ </match>
+ </magic>
</mime-type>
+ <mime-type type="application/x-msdownload;format=pe-itanium">
+ <sub-class-of type="application/x-msdownload;format=pe"/>
+ <magic priority="60">
+ <match value="PE\000\000" type="string" offset="128">
+ <match value="0x0200" type="little16" offset="132"/>
+ </match>
+ <match value="PE\000\000" type="string" offset="240">
+ <match value="0x0200" type="little16" offset="244"/>
+ </match>
+ </magic>
+ </mime-type>
+ <mime-type type="application/x-msdownload;format=pe-armLE">
+ <sub-class-of type="application/x-msdownload;format=pe"/>
+ <magic priority="60">
+ <match value="pe\000\000" type="string" offset="128">
+ <match value="0x01c0" type="little16" offset="132"/>
+ </match>
+ <match value="pe\000\000" type="string" offset="240">
+ <match value="0x01c0" type="little16" offset="244"/>
+ </match>
+ </magic>
+ </mime-type>
+ <mime-type type="application/x-msdownload;format=pe-arm7">
+ <sub-class-of type="application/x-msdownload;format=pe"/>
+ <magic priority="60">
+ <match value="pe\000\000" type="string" offset="128">
+ <match value="0x01c4" type="little16" offset="132"/>
+ </match>
+ <match value="pe\000\000" type="string" offset="240">
+ <match value="0x01c4" type="little16" offset="244"/>
+ </match>
+ </magic>
+ </mime-type>
+
<mime-type type="application/x-msmediaview">
<glob pattern="*.mvb"/>
<glob pattern="*.m13"/>
