SuperLinux Encyclopedia - File-related topics

Misc:

• Recovery of corrupted and removed files
• Removal of lost device files from lost+found
• slocate (FTP dir)  A "locate" command that respects permission bits.
• ClibPDF by FastIO Systems  Generate PDF files directly from C.
• cook  "a tool for constructing files" with some neat-sounding features.
• Copying Files Using Mirror  Jan'98 LG
• FAD file tools  "FAD files are text files which list the attributes of a set of files (owner, group, mode, crc, etc). The FAD tools provide programs for generating, manipulating and analysing file attribute databases."
• File comparision, differencing:
  • Note that if you've just created a file, an immediate comparison of that file might not be reading the file off the disk, but out of a buffer of the OS or disk controller or the drive electronics. The "sync" command will flush the OS buffer only. You could probably flush the other buffers by writting a file that is bigger than the non-OS buffers to the same disk that holds the file of interest and sync'ing. You can write the file with "dd" which can be given a size. You probably have more important things to do though.
  • multiple (FTP file)
  • X-based "diff" programs 
  • Emacs ediff mode  There are probably similar modes.
• File searching:
  • find: A must-learn. You have it. See the man page.
  • glimpse: An advanced gizmo for searching directory trees quickly after building a database.
  • locate: You have it. Requires building of database before use. See the man page.
  • Locus  Full-Text searcher; (something like Glimpse?)
  • slocate (FTP dir)  A "locate" command that respects permission bits.
  • Ask Slashdot: Full Text Search Engines
• File Attributes:
  • These are special attributes that files may have. They are rather new and not often used. Those not described here may not yet be supported by the kernel.
  • Commands: Change with "chattr" and list with "lsattr". See their man pages.
  • Immutable files: These files cannot be modified.
  • Append-only files: Can only be appended to.
  • No-dump files: Ignored by "dump".
  • Zero-on-delete files: These are overwritten with zeros before deletion.
  • Sync files: These have their I/O buffers flushed at the end of each write operation.
• Tips (all assume bash usage):
  • How to create an empty file or empty an existing one: ">|file", or sometimes ">file". (One needn't use "cat >|file". The "|" causes bash to ignore the value of "noclobber" (or "set -C").)
• Erasing files without leaving a trace (not much anyway):
  • See "zero-on-delete" attribute elsewhere on this page.
  • Wipe "for securely erasing files from magnetic media" (over-rights with various patterns)

core files:

A "core" file is created by most program when they crash. They contain debugging information and a complete copy of the program's run-time memory. Commonly, the "gdb" debugger is used to read the core file simply to determine which program created the file and why it crashed. The debugger can be used, if the program which created the core file has the debugging info, to examine program variables and even continue execution of the program. Core files may be freely removed, with the slight risk that someone stupidly names something else "core", like the Linux kernel source tree had a directory by that name.

To inhibit creation of core files: Use bash built-in "ulimit -c 0". (csh has "limit")

To test the most common use:

• Create a core file by running "cat" on a line by itself and then pressing "control-\".
• Run "gdb -c core".
• The gdb program outputs a header and then says what program created the core file and why it crashed. It then prompts the user for a gdb command. Enter "quit" to exit gdb.

File Conversion:

• Dealing with Word documents under Linux (Tutorial by Mike Madore; with links)
• MSWordView "a program that can understand the microsofts word 8 binary file format (office97), it currently converts word into html, which can then be read with a browser."
• juju/juen  "a smart decoder/encoder for uu-, base64-, binhex- or xxencoded data"

File Formats:

• ELF HOWTO   How to install and migrate to the ELF binary file format.
• ENCYCLOPEDIA OF GRAPHICS FILE FORMATS (Boardwatch mag's book review); Might be missing.
• Common filename extensions:
  Note: For "graphical" files, see above "Encyclopedia".
  ATM here means "Adobe Type Manager", PS means "Postscript".
  

            .a:      Archive of the "ar" program (an object library), usually.
            .afm     Adobe Font Manager (?) A Type 1 font in AFM format; ATM-compatible.
            .arj:    Unknown type of commpressed archive. There is an "unarj" decompressor.
            .bash:   A script in the Born-Again-Shell language.
            .bat:    Various kinds of batch command files. (Of M$ heritage.)
            .bdf:    A bitmapped font in BDF format.
            .bmp:    BMP graphical file. (Of M$Win and OS/2 heritage).
            .btm:    Unknown type of executable file.
            .c:      C source code.
            .cc:     C++ source code.
            .cf:     Various kinds of configuration files; usually a text file.
            .cgm:    A non-proprietary vector graphics file format.
            .cmd:    Various kinds of command files.
            .com:    Various kinds of command files. (Of VAX/VMS and M$ heritage.)
            .csh:    A script in the C-Shell language.
            .conf:   Various kinds of configuration files; usually a text file.
            .cpp:    C++ source code.
            .cxx:    C++ source code.
            .doc:    Various kinds of documents, usually Unix text files.
            .dvi:    TeX output file.
            .exe:    Various kinds of executable files; often an M$ OS file.
            .h:      C or C++ source code include files.
            .g3:     G3 Fax file.
            .gif:    GIF graphical file.
            .gmo     A GNU-format Machine Object file. See "gettext" info.
            .gsf:    Ghostscript Font; Type 1 outline font not for
                            ATM (for Ghostscript only? or any PS interpreter?)
            .gz:     Compressed by the gzip program.
            .htm:    Same as .html but usually only as copied from the M$ world.
            .html:   Hyper-Text Markup Language file used by WWW browers.
            .jpg:    JPEG file.
            .ksh:    A script in the Korn-Shell language.
            .l:      unknown
            .lsm:    An entry of the Linux Software Map.
            .lzh:    Commpressed by the lha (LHa/LHarc) program.
            .mo      A Machine Object file. See "gettext" info.
            .mod:    unknown
            .o:      An object (module) file. A compiler output.
            .orig:   Usually a backup of the "original" version of a file.
            .pdf:    Adobe's Page Description Format file.
            .pfa:    Postscript Font A; Type 1 outline font for
                            ATM or any PS interpreter. (M$-world heritage)
            .pfb:    Postscript Font B; Type 1 outline font for
                            ATM or any PS interpreter. (M$-world heritage)
            .pfm     Postscript Font Manager (?) A Type 1 font in PFM format; ATM-compatible.
            .png     Portable Network Graphics (?) (PNG) format file.
            .po      A Portable Object file. See "gettext" info.
            .pot     Temorary Portable Object file. See "gettext" info.
            .pox     A temporary Portable Object file. See "gettext" info.
            .ps:     Adobe's Postscript language (input) file.
            .rar:    unknown; some kind of M$ archive, I suspect.
            .rpm:    Archive used with the "rpm" package management program.
            .rpmsave:  A backup file created by the "rpm" program.
            .S:      unknown
            .s:      unknown
            .s3m:    unknown
            .sa:     unknown
            .sh:     A script in the Bourne shell language.
            .sgml:   Structured Generalized Markup Language (?) file.
            .so:     Shared Object module (linking) library.
            .tar:    An archive of the "tar" program.
            .taz:    Unknown type of commpressed file.
            .tex:    TeX input file.
            .texi:   Input to the texinfo (?) program.
            .tgz:    An archive of the "tar" program compressed with the -z option.
            .tif:    TIFF
            .tiffg3: TIFFG3 Fax file.
            .troff:  Input to the troff (text run-off?) program.
            .ttf:    TrueType Font.
            .txt:    Text file; usually with "^J" (newline) end-of-line markers.
            .voc:    unknown
            .wav:    Sound data of a certain kind.
            .xbm:    XBM graphical file. Monochrome.
            .xpm:    XPM graphical file.
            .xml:    Extensible Markup Language (?) file.
            .y:      unknown
            .zip:    Compressed with one of the "zip" programs.
            .Z:      Compressed with the "compress" program.
            .z:      Compressed with the "gzip" program, usually.
            rc:      Various kinds of configuration files; usually a text file.
                            (No "dot".) Stands for "Resource Configuration". ??
            .:       tbd
          

File Managers:

• Freshmeat: File Managers (text)
• Freshmeat: File Managers (X11)
• Chuck's Linux File Manager Page
• EZFM  A filemanager for X.
• FileRunner:
  • Filerunner
  • FileRunner: A New Tk/Tcl File Manager  Sep'96 LG
• gentoo  Needs gtk lib.
• Moxfm  New version of xfm. Mar'96 LG
• rsync  A file transfer program for making file sets the same on two systems.
• SFM - Simple File Manager
• TkDesk  "Desktop and File Manager"
• X-Files  Nov'96 LG
• XNC - X Northern Commander

End of page.