/*******************************************************************************
*
* text.c
*        
* AUTHOR:
*   Dan Harkless                                                    <opensource@
*   http://harkless.org/dan/software/                              harkless.org>
* 
* COPYRIGHT:
*   This file is Copyright (C) 2003 by Dan Harkless, and is released under the
*   GNU General Public License <http://www.gnu.org/copyleft/gpl.html>.
*
* USAGE:
*   % text <file1> [<file2>...]
*
* DESCRIPTION:                                                               
*   It's really annoying when you're doing:
*
*       grep PATTERN *
*
*   and the current directory includes binary files and/or subdirectories,
*   resulting in the output of long lines filled with garbage characters,
*   ringing bells, and often the freeze-up of your terminal.
*
*   Proper modern greps like GNU grep (which just outputs "Binary file FILENAME
*   matches") don't suffer from this problem, but in case you're using a vendor
*   version that does, you can use this program to "screen" files before they're
*   passed off to grep for processing.  If you do:
*
*       grep PATTERN `text *`
*
*   grep will only look for your Pattern in "text" files in the local directory.
*
*   I put "text" in quotes because this program's algorithm is fast but
*   imperfect -- it only checks whether the first character in each file is a
*   printable character.  This is fine for general use, because the most common
*   binaries files most people will come across in a directory start with
*   non-printable "magic numbers", but proprietary data files could well start
*   with a printable character.  In the future I may add an option to check the
*   first N bytes of each file for non-printable characters.
*                                                                            
*   To facilitate use of this command in scripts, 'text' prints "/dev/null" if
*   it finds no text files, so grep won't hang waiting for you to type a CTRL-D.
*                                                                            
*   Note that of course you can use 'text' with other commands besides
*   [ef]?grep.
*                                                                            
* DATE        MODIFICATION
* ==========  ==================================================================
* 2003-10-09  Noted that GNU grep doesn't suffer the problem this script fixes.
* 2003-06-25  Ignore setocale() failure if ENOENT; for use on minimal installs.
* 2003-04-23  Need setlocale(LC_ALL, "") for isgraph() to work for non-"C"
*             locales.  Doing this increases the chances for false positives,
*             however -- we might want to add an option to disable the call.
*             In the meantime, `env LANG=C text *` works for that.
* 1999-04-09  Get our program name from argv[0] rather than hardcoding as text.
* 1998-10-22  Directories can start with a printable character -- use stat() and
*             only consider regular files.
* 1995-08-30  Original.
*                                                                            
*******************************************************************************/


#include <ctype.h>     /* for isgraph() and isspace() */
#include <errno.h>     /* for errno */
#include <locale.h>    /* for setlocale() */
#include <stdio.h>     /* for fprintf(), etc. */
#include <stdlib.h>    /* for malloc() */
#include <string.h>    /* for strrchr(), etc. */
#include <sys/stat.h>  /* for stat(), etc. */


#define FALSE 0
#define TRUE  1
typedef unsigned char  boolean;


int  main(int  argc, char**  argv) {
    boolean   found_at_least_one_text_file = FALSE;
    char*     last_slash = strrchr(argv[0], '/');
    char*     our_program_name;
    int       i;

    if (last_slash == NULL)
	our_program_name = argv[0];
    else
	our_program_name = last_slash + 1;
    
    if (setlocale(LC_ALL, "") == NULL && errno != ENOENT) {
        fprintf(stderr, "%s: setlocale(): %s.\n",
                our_program_name, strerror(errno));
	return EXIT_FAILURE;
    }

    if (argc < 2) {
        fprintf(stderr, "Usage: %s <file1> [<file2>...]\n", our_program_name);
        return EXIT_FAILURE;
    }

    for (i = 1; i < argc; i++) {
	char         c;
	struct stat  file_stat;
	
	if (stat(argv[i], &file_stat) < 0)
	    fprintf(stderr, "%s: %s.\n", argv[i], strerror(errno));
	else if (S_ISREG(file_stat.st_mode)) {
	    FILE*  file = fopen(argv[i], "r");
	    
	    if (file == NULL)
		fprintf(stderr, "%s: %s.\n", argv[i], strerror(errno));
	    else {
		fread(&c, 1, 1, file);
		if (isgraph(c) || isspace(c)) {
		    if (found_at_least_one_text_file)
			putchar(' ');
		    
		    fputs(argv[i], stdout);
		    
		    found_at_least_one_text_file = TRUE;
		}
		fclose(file);
	    }
	}
    }
    
    if (!found_at_least_one_text_file)
	fputs("/dev/null", stdout);
    
    putchar('\n');
    
    return EXIT_SUCCESS;
}