greg/pcompress
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
pcompress/utils/phash/extensions.h

234 lines
8.7 KiB
C
Raw Normal View History

Add extension based file type detection and setting segment data type. Use Bob Jenkins Minimal Perfect Hash to check for known extensions. Use semaphore signaling and direct buffer copy for extraction. Miscellaneous fixes.
2013-11-07 16:18:54 +00:00
/* Generated File. DO NOT EDIT. */
/*
* List of extensions and their types.
*/
#ifndef __EXT_H__
#define __EXT_H__
struct ext_entry {
char *ext;
int type;
int len;
} extlist[] = {
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"c" , TYPE_TEXT, 1},
{"h" , TYPE_TEXT, 1},
{"cc" , TYPE_TEXT, 2},
{"cpp" , TYPE_TEXT, 3},
{"c++" , TYPE_TEXT, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"h++" , TYPE_TEXT, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"hpp" , TYPE_TEXT, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"hxx" , TYPE_TEXT, 3},
{"hh" , TYPE_TEXT, 2},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"txt" , TYPE_TEXT, 3},
{"text" , TYPE_TEXT, 4},
{"html" , TYPE_TEXT|TYPE_MARKUP, 4},
{"htm" , TYPE_TEXT|TYPE_MARKUP, 3},
{"xml" , TYPE_TEXT|TYPE_MARKUP, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"sgml" , TYPE_TEXT|TYPE_MARKUP, 4},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"info" , TYPE_TEXT, 4},
{"svg" , TYPE_TEXT, 3},
{"conf" , TYPE_TEXT, 4},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"cfg" , TYPE_TEXT, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"py" , TYPE_TEXT, 2},
{"rb" , TYPE_TEXT, 2},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"ru" , TYPE_TEXT, 2},
{"rbw" , TYPE_TEXT, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"xpm" , TYPE_TEXT, 3},
{"js" , TYPE_TEXT, 2},
{"jsp" , TYPE_TEXT, 3},
{"pl" , TYPE_TEXT, 2},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"t" , TYPE_TEXT, 1},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"tcl" , TYPE_TEXT, 3},
{"sh" , TYPE_TEXT, 2},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"ksh" , TYPE_TEXT, 3},
{"csh" , TYPE_TEXT, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"php" , TYPE_TEXT, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"php3" , TYPE_TEXT, 4},
{"php4" , TYPE_TEXT, 4},
{"php5" , TYPE_TEXT, 4},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"bat" , TYPE_TEXT, 3},
{"pm" , TYPE_TEXT, 2},
{"r" , TYPE_TEXT, 1},
{"d" , TYPE_TEXT, 1},
{"bas" , TYPE_TEXT, 3},
{"asm" , TYPE_TEXT, 3},
{"go" , TYPE_TEXT, 2},
{"java" , TYPE_TEXT, 4},
{"m4" , TYPE_TEXT, 2},
{"vb" , TYPE_TEXT, 2},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"vba" , TYPE_TEXT, 3},
{"vbs" , TYPE_TEXT, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"xslt" , TYPE_TEXT|TYPE_MARKUP, 4},
{"xsl" , TYPE_TEXT|TYPE_MARKUP, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"xsd" , TYPE_TEXT|TYPE_MARKUP, 3},
{"xs" , TYPE_TEXT, 2},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"yacc" , TYPE_TEXT, 4},
{"lex" , TYPE_TEXT, 3},
{"csv" , TYPE_TEXT, 3},
{"shtml" , TYPE_TEXT|TYPE_MARKUP, 5},
{"xhtml" , TYPE_TEXT|TYPE_MARKUP, 5},
{"xht" , TYPE_TEXT|TYPE_MARKUP, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"tpl" , TYPE_TEXT|TYPE_MARKUP, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"asp" , TYPE_TEXT, 3},
{"aspx" , TYPE_TEXT, 4},
{"rss" , TYPE_TEXT|TYPE_MARKUP, 3},
{"atom" , TYPE_TEXT|TYPE_MARKUP, 4},
{"cgi" , TYPE_TEXT, 3},
{"c#" , TYPE_TEXT, 2},
{"cob" , TYPE_TEXT, 3},
{"ada" , TYPE_TEXT, 3},
{"ini" , TYPE_TEXT, 3},
{"y" , TYPE_TEXT, 1},
{"swg" , TYPE_TEXT, 3},
{"s" , TYPE_TEXT, 1},
{"ps" , TYPE_TEXT, 2},
{"bib" , TYPE_TEXT, 3},
{"lua" , TYPE_TEXT, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"nse" , TYPE_TEXT, 3},
Misclleanous cleanups. Fix some formatting. Increase metadata chunk size slightly.
2014-12-19 13:45:39 +00:00
{"dtd" , TYPE_TEXT|TYPE_MARKUP, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"qml" , TYPE_TEXT|TYPE_MARKUP, 3},
{"fa" , TYPE_TEXT|TYPE_DNA_SEQ, 2},
{"faa" , TYPE_TEXT|TYPE_DNA_SEQ, 3},
{"asn" , TYPE_TEXT|TYPE_MARKUP, 3},
{"ffn" , TYPE_TEXT|TYPE_DNA_SEQ, 3},
{"fna" , TYPE_TEXT|TYPE_DNA_SEQ, 3},
{"frn" , TYPE_TEXT|TYPE_DNA_SEQ, 3},
{"gbk" , TYPE_TEXT, 3},
{"gff" , TYPE_TEXT, 3},
{"ptt" , TYPE_TEXT, 3},
{"rnt" , TYPE_TEXT, 3},
{"val" , TYPE_BINARY, 3},
{"tcc" , TYPE_TEXT, 3},
{"css" , TYPE_TEXT, 3},
{"pod" , TYPE_TEXT, 3},
{"al" , TYPE_TEXT, 2},
{"vim" , TYPE_TEXT, 3},
{"am" , TYPE_TEXT, 2},
{"upp" , TYPE_TEXT, 3},
{"mom" , TYPE_TEXT, 3},
{"tmac" , TYPE_TEXT, 4},
{"pdf" , TYPE_BINARY, 3},
{"tif" , TYPE_BINARY|TYPE_TIFF, 3},
{"tiff" , TYPE_BINARY|TYPE_TIFF, 4},
{"jpg" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_JPEG, 3},
{"jpeg" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_JPEG, 4},
{"mjpeg" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_MJPEG, 5},
{"png" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_GZ, 3},
Tweak some data type settings.
2014-11-19 14:49:16 +00:00
{"mp3" , TYPE_BINARY, 3},
{"wma" , TYPE_BINARY, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"divx" , TYPE_BINARY|TYPE_COMPRESSED, 4},
{"mp4" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_MP4, 3},
Tweak some data type settings.
2014-11-19 14:49:16 +00:00
{"aac" , TYPE_BINARY, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"m4a" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"m4p" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"ofs" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"ofr" , TYPE_BINARY|TYPE_COMPRESSED, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"ogg" , TYPE_BINARY|TYPE_COMPRESSED, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"flac" , TYPE_BINARY|TYPE_FLAC, 4},
{"avi" , TYPE_BINARY|TYPE_AVI, 3},
{"pac" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"gif" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_LZW, 3},
{"jp2" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_JPEG, 3},
{"pjg" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_PACKJPG, 3},
{"gz" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_GZ, 2},
{"tgz" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_GZ, 3},
{"bz2" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_BZ2, 3},
{"tbz2" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_BZ2, 4},
{"zip" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_ZIP, 3},
{"arj" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_ARJ, 3},
{"arc" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_ARC, 3},
{"jar" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_GZ, 3},
{"lz" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_LZ, 2},
{"lzh" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"lha" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"lzma" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_LZMA, 4},
{"lzo" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_LZO, 3},
{"dmg" , TYPE_BINARY, 3},
{"7z" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_LZMA, 2},
{"uha" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"alz" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"ace" , TYPE_BINARY|TYPE_COMPRESSED, 3},
{"rar" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_RAR, 3},
{"xz" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_LZMA, 2},
{"txz" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_LZMA, 3},
{"pmd" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_PPMD, 3},
{"zpaq" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_ZPAQ, 4},
{"xcf" , TYPE_BINARY, 3},
{"mo" , TYPE_BINARY, 2},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"gmo" , TYPE_BINARY, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"pyo" , TYPE_BINARY, 3},
{"pyc" , TYPE_BINARY, 3},
Step 0 of adding WavPack filter - does not work yet. WAV file detection. Rename libarchive dir to be generic.
2014-09-14 18:26:38 +00:00
{"wav" , TYPE_BINARY|TYPE_WAV, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"tta" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_AUDIO_COMPRESSED, 3},
{"wv" , TYPE_BINARY|TYPE_COMPRESSED, 2},
{"swf" , TYPE_BINARY, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"svgz" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_GZ, 4},
{"odt" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_COMPRESSED_ZIP, 3},
{"3dm" , TYPE_BINARY, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"chm" , TYPE_BINARY, 3},
{"CHM" , TYPE_BINARY, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"svn" , TYPE_BINARY, 3},
Regenerate extensions hash.
2014-09-12 11:30:36 +00:00
{"ppm" , TYPE_BINARY|TYPE_PNM, 3},
{"pbm" , TYPE_BINARY|TYPE_PNM, 3},
{"pgm" , TYPE_BINARY|TYPE_PNM, 3},
{"pnm" , TYPE_BINARY|TYPE_PNM, 3},
{"ppn" , TYPE_BINARY|TYPE_COMPRESSED|TYPE_PACKPNM, 3},
Improve file sorting algorithm. Add more file extension names. Fix data type mask size.
2014-10-27 13:53:03 +00:00
{"mk" , TYPE_TEXT, 2},
{"diff" , TYPE_TEXT, 4},
{"po" , TYPE_TEXT, 2},
{"pot" , TYPE_TEXT, 3},
{"in" , TYPE_TEXT, 2},
{"ac" , TYPE_TEXT, 2},
{"guess" , TYPE_TEXT, 5},
{"sub" , TYPE_TEXT, 3},
{"rpath" , TYPE_TEXT, 5},
{"texi" , TYPE_TEXT, 4},
{"valgrind" , TYPE_TEXT, 8},
{"gperf" , TYPE_TEXT, 5},
{"latex" , TYPE_TEXT, 5},
{"f77" , TYPE_TEXT, 3},
{"f90" , TYPE_TEXT, 3},
{"f95" , TYPE_TEXT, 3},
{"groovy" , TYPE_TEXT, 6},
{"ebuild" , TYPE_TEXT, 6},
{"rex" , TYPE_TEXT, 3},
{"rexx" , TYPE_TEXT, 4},
{"scala" , TYPE_TEXT, 5},
{"xaml" , TYPE_TEXT|TYPE_MARKUP, 4},
{"yaml" , TYPE_TEXT|TYPE_MARKUP, 4},
{"tex" , TYPE_TEXT, 3},
{"rebol" , TYPE_TEXT, 5},
{"reb" , TYPE_TEXT, 3},
{"perl" , TYPE_TEXT, 4},
{"pas" , TYPE_TEXT, 3},
{"p6" , TYPE_TEXT, 2},
{"z80" , TYPE_TEXT, 3},
{"scm" , TYPE_TEXT, 3},
{"ss" , TYPE_TEXT, 2},
{"ml" , TYPE_TEXT, 2},
{"ml4" , TYPE_TEXT, 3},
{"mli" , TYPE_TEXT, 3},
{"mm" , TYPE_TEXT, 2},
{"m3" , TYPE_TEXT, 2},
{"lisp" , TYPE_TEXT, 4},
{"kdebuild-1" , TYPE_TEXT, 10},
{"hs" , TYPE_TEXT, 2},
{"gemspec" , TYPE_TEXT, 7},
{"fs" , TYPE_TEXT, 2},
{"coffee" , TYPE_TEXT, 6},
{"e" , TYPE_TEXT, 1},
{"cu" , TYPE_TEXT, 2},
{"awk" , TYPE_TEXT, 3},
{"xls" , TYPE_BINARY, 3},
{"xlw" , TYPE_BINARY, 3},
{"qt" , TYPE_BINARY, 2},
{"charset" , TYPE_TEXT, 7},
{"sed" , TYPE_TEXT, 3},
{"mailmap" , TYPE_TEXT, 7},
{"sin" , TYPE_BINARY, 3},
Add couple of mozilla file extensions. Check for files > INT64_T when sorting. Makefile targets to help development.
2014-11-15 13:47:33 +00:00
{"xul" , TYPE_TEXT, 3},
{"xpt" , TYPE_BINARY, 3},
Add extension based file type detection and setting segment data type. Use Bob Jenkins Minimal Perfect Hash to check for known extensions. Use semaphore signaling and direct buffer copy for extraction. Miscellaneous fixes.
2013-11-07 16:18:54 +00:00
};
#endif
Reference in a new issue Copy permalink