1b6cee71dSXin LI 2b6cee71dSXin LI#------------------------------------------------------------------------------ 3*ae316d1dSXin LI# $File: msooxml,v 1.23 2024/07/19 18:48:23 christos Exp $ 4b6cee71dSXin LI# msooxml: file(1) magic for Microsoft Office XML 5b6cee71dSXin LI# From: Ralf Brown <ralf.brown@gmail.com> 6b6cee71dSXin LI 7b6cee71dSXin LI# .docx, .pptx, and .xlsx are XML plus other files inside a ZIP 8b6cee71dSXin LI# archive. The first member file is normally "[Content_Types].xml". 9b6cee71dSXin LI# but some libreoffice generated files put this later. Perhaps skip 10b6cee71dSXin LI# the "[Content_Types].xml" test? 11b6cee71dSXin LI# Since MSOOXML doesn't have anything like the uncompressed "mimetype" 12b6cee71dSXin LI# file of ePub or OpenDocument, we'll have to scan for a filename 13b6cee71dSXin LI# which can distinguish between the three types 14b6cee71dSXin LI 1558a0f0d0SEitan Adler0 name msooxml 1658a0f0d0SEitan Adler>0 string word/ Microsoft Word 2007+ 1758a0f0d0SEitan Adler!:mime application/vnd.openxmlformats-officedocument.wordprocessingml.document 18a2dfb722SXin LI!:ext docx 1958a0f0d0SEitan Adler>0 string ppt/ Microsoft PowerPoint 2007+ 2058a0f0d0SEitan Adler!:mime application/vnd.openxmlformats-officedocument.presentationml.presentation 21a2dfb722SXin LI!:ext pptx 2258a0f0d0SEitan Adler>0 string xl/ Microsoft Excel 2007+ 2358a0f0d0SEitan Adler!:mime application/vnd.openxmlformats-officedocument.spreadsheetml.sheet 24a2dfb722SXin LI!:ext xlsx 2543a5ec4eSXin LI>0 string visio/ Microsoft Visio 2013+ 26d38c30c0SXin LI!:mime application/vnd.ms-visio.drawing.main+xml 2743a5ec4eSXin LI>0 string AppManifest.xaml Microsoft Silverlight Application 2843a5ec4eSXin LI!:mime application/x-silverlight-app 2958a0f0d0SEitan Adler 30b6cee71dSXin LI# start by checking for ZIP local file header signature 31b6cee71dSXin LI0 string PK\003\004 32b6cee71dSXin LI!:strength +10 33b6cee71dSXin LI# make sure the first file is correct 342dc4dbb9SEitan Adler>0x1E use msooxml 3543a5ec4eSXin LI>0x1E default x 36a4d6d3b8SXin LI>>0x1E regex \\[Content_Types\\]\\.xml|_rels/\\.rels|docProps|customXml 37b6cee71dSXin LI# skip to the second local file header 38b6cee71dSXin LI# since some documents include a 520-byte extra field following the file 39b6cee71dSXin LI# header, we need to scan for the next header 4043a5ec4eSXin LI>>>(18.l+49) search/6000 PK\003\004 41*ae316d1dSXin LI>>>>&26 use msooxml 42*ae316d1dSXin LI>>>>&26 default x 43b6cee71dSXin LI# now skip to the *third* local file header; again, we need to scan due to a 44b6cee71dSXin LI# 520-byte extra field following the file header 45*ae316d1dSXin LI>>>>>&26 search/6000 PK\003\004 46b6cee71dSXin LI# and check the subdirectory name to determine which type of OOXML 47b6cee71dSXin LI# file we have. Correct the mimetype with the registered ones: 4848c779cdSXin LI# https://technet.microsoft.com/en-us/library/cc179224.aspx 49*ae316d1dSXin LI>>>>>>&26 use msooxml 50*ae316d1dSXin LI>>>>>>&26 default x 5158a0f0d0SEitan Adler# OpenOffice/Libreoffice orders ZIP entry differently, so check the 4th file 52*ae316d1dSXin LI>>>>>>>&26 search/6000 PK\003\004 53*ae316d1dSXin LI>>>>>>>>&26 use msooxml 54a4d6d3b8SXin LI# Some OOXML generators add an extra customXml directory. Check another file. 55*ae316d1dSXin LI>>>>>>>>&26 default x 56*ae316d1dSXin LI>>>>>>>>>&26 search/6000 PK\003\004 57*ae316d1dSXin LI>>>>>>>>>>&26 use msooxml 58*ae316d1dSXin LI>>>>>>>>>>&26 default x 59*ae316d1dSXin LI>>>>>>>>>>>&26 search/6000 PK\003\004 60*ae316d1dSXin LI>>>>>>>>>>>>&26 use msooxml 61*ae316d1dSXin LI>>>>>>>>>>>>&26 default x Microsoft OOXML 62*ae316d1dSXin LI>>>>>>>>>>>&26 default x Microsoft OOXML 63*ae316d1dSXin LI>>>>>>>>>>&26 default x Microsoft OOXML 64a4d6d3b8SXin LI>>>>>>>>>&26 default x Microsoft OOXML 65*ae316d1dSXin LI>>>>>>>>&26 default x Microsoft OOXML 6643a5ec4eSXin LI>>>>>>>&26 default x Microsoft OOXML 67*ae316d1dSXin LI>>>>>>&26 default x Microsoft OOXML 68898496eeSXin LI>>0x1E regex \\[trash\\] 69898496eeSXin LI>>>&26 search/6000 PK\003\004 70898496eeSXin LI>>>>&26 search/6000 PK\003\004 71898496eeSXin LI>>>>>&26 use msooxml 72898496eeSXin LI>>>>>&26 default x 73898496eeSXin LI>>>>>>&26 search/6000 PK\003\004 74898496eeSXin LI>>>>>>>&26 use msooxml 75898496eeSXin LI>>>>>>>&26 default x Microsoft OOXML 76898496eeSXin LI>>>>>>&26 default x Microsoft OOXML 77898496eeSXin LI>>>>>&26 default x Microsoft OOXML 78