MemoSoup » unix

AIX: reset user password

admin — Mon, 30 Jan 2012 13:17:54 +0000

Change the password for the user (as root):

passwd username

Reset the login counter and unlock the user:

chsec -f /etc/security/lastlog -a "unsuccessful_login_count=0" -s username
chuser "account_locked=false" username

Unix: print in color

admin — Sun, 30 Jan 2011 14:46:52 +0000

This simple procedure takes 3 arguments as the parameters:

Text
Foreground color
Background color

It prints the text to the terminal in selected colors.

function print_color {
local text=$1
local fg=$2
local bg=$3
case "$fg" in
red) fg="31m" ;;
green) fg="32m" ;;
yellow) fg="33m" ;;
blue) fg="34m" ;;
white) fg="37m" ;;
black) fg="30m" ;;
*) fg="37m" ;;
esac
case "$bg" in
red) bg="41m" ;;
green) bg="42m" ;;
yellow) bg="43m" ;;
blue) bg="44m" ;;
white) bg="47m" ;;
black) bg="40m" ;;
*) bg="40m" ;;
esac
echo -en "\033[${fg}\033[${bg}${text}\033[0m"
}

Unix: rename files to lowercase

admin — Sun, 30 Jan 2011 14:44:20 +0000

It’s very simple to rename the files to lowercase/uppercase using awk:

Lowercase

ls -1rt | awk '{ printf("mv %s %s\n", $0, tolower($0)) | "sh" } END { close("sh") }'

Uppercase

ls -1rt | awk '{ printf("mv %s %s\n", $0, toupper($0)) | "sh" } END { close("sh") }'

Unix: get the file date

admin — Fri, 18 Sep 2009 15:08:18 +0000

ls -1 | cpio -o | cpio -ivt | awk '{print $NF, $(NF-1), $(NF-4), $(NF-3) }'

Warning: I/O expensive for the large files!

Perl:

@a = localtime((stat($my_file))[9]); $a[4]++;
printf "%02d%02d%02d",@a[5,4,3];

Redirect script output to the log

admin — Sun, 13 Sep 2009 17:29:58 +0000

If the whole output of the complex script should be redirected to the log, the following trick could be used.

if [ "$1" != "-log" ] ; then
$0 -log "$@" 2>&1 | tee the_log_file.$$.log
echo "The log file for the current session: the_log_file.$$.log"
exit 0
fi
shift # remove "-log" parameter

grep in find command: how to display file names

admin — Sat, 12 Sep 2009 22:09:38 +0000

Here is very simple trick to force the grep command to display file name, when it’s used together with find operation.
Just write /dev/null as the “second file”

find . -type f -exec grep somestring {} /dev/null \;

Using pattern lists in Unix

admin — Sat, 12 Sep 2009 18:57:48 +0000

Here is the small reminder about the syntax of the “case” command and the usage of the pattern lists.

#!/bin/ksh
print -n "Please enter the line: "
read line

case "$line" in
?(dog|cat) ) print "zero or one occurrence of any pattern" ;;
*(low|high) ) print "zero or more occurrences of any pattern" ;;
@(duncan|methos) ) print "exactly one occurrence of any pattern" ;;
+(rudolph|blitzen) ) print "one or more occurrence of any pattern" ;;
!(grinch|babay) ) print "everything except patterns" ;;
-@([hH?]) ) print "Some help..." ;;
-v*(erbose) ) print "Some more words..." ;;
*) print "Something else..." ;;
esac

Print the PATH directories in the readable format

admin — Sat, 12 Sep 2009 18:55:53 +0000

echo $PATH| awk -v RS=":" '{ print $0 }'

echo $LD_LIBRARY_PATH |awk -v RS=":" '{ system ( "ls -rltd " $0 ) }'

Warning!
As far as the option ‘-v’ is used, the new awk(nawk in some systems) should be used.

To check if the new version of awk is installed:

awk 1 /dev/null

The output will be empty for new awk.
You have the old awk, if the output is something like

awk: syntax error near line 1

In this case nawk should be used instead awk

MS-DOS modification: echo %PATH% | awk -v RS=”;” “{ print $0 }”

Parsing script parameters

admin — Sat, 12 Sep 2009 15:26:31 +0000

Quick and dirty parsing procedure for unix shell scripts

parse_command_line ()
{
typeset -i user_opt help_opt password_opt verbose_opt
typeset errmsg

arg_cou=$#

while [ "$#" -gt 0 ]
do
<strong>case "$1" instrong>
-@([U]) ) let user_opt=user_opt+1 ;;
-P ) let password_opt=password_opt+1 ; ask_pass=0 ;;
-option1 ) ;;
-option2 ) ;;
-option3 ) ;;
-@([hH?])) help_opt=1;;
-v*(erbose)) let verbose_opt=verbose_opt+1 ;;
*)
if [ user_opt -eq 1 ];then
USER="$1"; user_opt=0
elif [ password_opt -eq 1 ];then
PASS="$1"; password_opt=0
elif [ verbose_opt -eq 1 ];then
verbose_level="$1"; verbose_opt=0
else
### help_opt=1
CMD=$* ; break
fi
;;
esac
shift
done

if [ arg_cou -eq 0 ] ; then
help_opt=1 ; errmsg="Nothing to do"
fi

for opt in user password ; do
if [ ${opt}_opt -ne 0 ] ; then
help_opt=1 ; errmsg="${errmsg}\n${opt} is absent"
fi
done

if [ verbose_opt -ne 0 ] ; then
verbose_level=$verbose_opt
fi

if [ help_opt -eq 1 ]; then
help "$errmsg"
fi
}

Rev function and comma-separated output

admin — Fri, 11 Sep 2009 23:26:00 +0000

Here is the shell command snippet to display comma-separated output:

ls -lrt | rev | sed 's/\$[0-9][0-9][0-9]\$/\\1,/g' | rev | sed 's/\$[\^0-9]\$,\$[0-9]\$/\\1\\2/g;s/\^,\$[0-9]\$/\\1/g'

Example:
-rw-r—– 1 sybase dba 1,572,872,192 Feb 2 07:09 master.dbf

Rev in awk

#!/bin/ksh
nawk '{ l=length($0) ; for(i=l;i>0;i--) { printf "%s", substr($0,i,1) } ; print "" }'

Rev function (absent on SunOS) :

(Warning! Avoid spaces before #define – SunOS cc compiler doesn’t like them)

#include
#ident "@(#)rev 1.0 02/01/2006"
#define MAX_STRING_SIZE 2000char version[] = "rev v1.0 02/01/2006";

main(argc,argv)
int argc;
char *argv[];
{
char my_src[MAX_STRING_SIZE] ;
char my_dst[MAX_STRING_SIZE] ;
char *retCode ;

char *tmp_s ;
int i;

char *filename;
FILE *fSource ;

/* check command line arguments */
if (argc > 1) {
if( argv[1][0]=='-' || argv[1][0]=='/' ) {
fprintf(stderr, "Usage: %s [file] \n", argv[0]);
exit(1);
}
filename=argv[1] ;
fSource = fopen( filename, "r" ) ;
if ( fSource == NULL ) {
fprintf(stderr, "Can't open %s for reading\n", filename );
exit(1);
}
} else {
/* Using the standard input */
fSource=stdin ;
} /* if (argc > 1) */

while( (retCode = fgets( my_src, MAX_STRING_SIZE, fSource ) )!=NULL ) {

tmp_s=my_dst+MAX_STRING_SIZE-1;
*tmp_s=0;

for( i=0 ; i < MAX_STRING_SIZE && my_src[ i ] != 0 ; i++ ) {
if( my_src[ i ] == '\n' ) {
continue ;
}
tmp_s--;
*tmp_s=my_src[ i ];
}

fprintf( stdout, "%s\n", tmp_s);
}

fclose( fSource ) ;

fflush( stdout );

exit(0);
} /* main */