#!/bin/bash
## Originally...
# Copyright (c) 2004 Matthias S. Benkmann <article AT winterdrache DOT de>
# You may do everything with this code except misrepresent its origin.
# PROVIDED `AS IS' WITH ABSOLUTELY NO WARRANTY OF ANY KIND!

# Copyright (C) 2014 Steve Youngs <steve@steveyoungs.com>
#  Many updates/tweaks --SY.

# The following list should contain the mount points of all
# filesystems that are to be scanned as a space-separated list within
# parentheses. / will usually be in this list and if you have /usr on
# a separate partition, it will also be in this list. Mount points
# whose filesystems are special, such as procfs or sysfs must not be
# in this list. While a simple find on those special filesystems
# should be harmless, operations such as "-exec grep something" are
# NOT SAFE and may have HARMFUL SIDE-EFFECTS, especially when
# performed as root.

## Bastard settings
# fs_to_scan=(/ /opt /usr /var)

fs_to_scan=(/)

## NOTE: if any of the directories listed in fs_to_scan contain
## non-UNIX filesystems (MS-DOS, CD-ROM etc) you need to set $NOLEAF
## here to `-noleaf'.  But only do so if you really need to as it
## comes with a significant slow down on the find.
# NOLEAF='-noleaf'
NOLEAF=

# Files with a path prefix found in the following list are ignored. As
# the main function of this script is to help you find files that
# contain hardwired paths to /tools or other unwanted references to
# your build system, you will usually prune any directories that don't
# contain files of interest, such as /tools (whose files naturally
# refer to /tools) and your package users' home directories (which may
# also test positive if you have unpacked and configured sources lying
# around).
#
# NOTE: If a directory you want to prune is on a separate filesystem
# (separate from those listed in fs_to_scan) you don't need to list it
# here because of the -xdev option used in the find command.

## Bastard settings
# prune_prefixes=(\
#	/{,*/{,*/}}lost+found \
#       /root \
#	/opt/pgsql/data \
#	/opt/sql-ledger/{spool,templates,users,css} \
#       /etc/apache/ssl.key \
#	/etc/audisp/plugins.d \
#       /etc/cups/ssl \
#       /etc/firewall \
#	/etc/polkit-1/rules.d \
#       /etc/skel \
#       /etc/ssl/private \
#	/etc/sudoers.d \
#	/var/lib/{sasl,sudo,net-snmp,udisks{,2},NetworkManager} \
#	/var/log \
#	/usr/share/polkit-1/rules.d \
#	/var/tmp \
#       /var/{cache,chroot,run,snmp,spool} \
#       /var/lib/{sshd,nfs,spamassassin,pulse}) #NO TRAILING SLASHES!!!!

prune_prefixes=(/home /usr/src /dev /tools) #NO TRAILING SLASHES!!!

if [ $# -lt 1 -o "$1" = "--help" ]; then
    echo 1>&2 
    echo 1>&2 'USAGE: '"${0##*/}"' <grep-commands>'
    echo 1>&2 
    echo 1>&2 '  grep -l <grep-commands> -- <file>'
    echo 1>&2 '  will be executed for each *regular file* <file>'
    echo 1>&2 '  ATTENTION! If you override the -l switch with a switch that makes grep'
    echo 1>&2 '  output all individual matches rather than just the matching files,'
    echo 1>&2 '  then DO NOT redirect output to a file that is in a directory that will be'
    echo 1>&2 '  scanned, or you risk creating an endless loop that will cause your'
    echo 1>&2 '  output file to grow till your disk is full.'
    echo 1>&2 
    exit 1
fi

# suppress ugly debug output from shell
trap ':' SIGPIPE

# construct find commands that match the prune_prefixes. Each prefix will be
# matched as -path <prefix> -or -path <prefix>/*
# so that the directory itself and all subdirectories are matched.
y=(\( -false)
for ((i=0; $i<${#prune_prefixes[@]}; i=$i+1)) 
do
    y[${#y[@]}]='-or'
    y[${#y[@]}]=-path
    y[${#y[@]}]="${prune_prefixes[$i]}"
    y[${#y[@]}]='-or'
    y[${#y[@]}]=-path
    y[${#y[@]}]="${prune_prefixes[$i]}/*"
done
y[${#y[@]}]=')'

cmd_pre=(-type f -exec grep -l)
cmd_post=(-- {} \;)

# In the following find command, the part
# -not ( ( "${y[@]}" -prune ) -or "${y[@]}" )
# is responsible for preventing the files that match prune_prefixes from
# being processed. The 2nd "${y[@]}" may seem redundant, but it isn't, because
# -prune has no effect and is always false when -depth is used (which someone
# might do in the future).
# The -true before "$@" ensures that -depth can be passed as 1st parameter
# of $cmd_pre (should someone change it in the future).
find "${fs_to_scan[@]}" -xdev $NOLEAF \
    -not \( \( "${y[@]}" -prune \) -or "${y[@]}" \) \
    -and \( -true "${cmd_pre[@]}" "$@" "${cmd_post[@]}" \)