strings.subr revision 250702
1238438Sdteskeif [ ! "$_STRINGS_SUBR" ]; then _STRINGS_SUBR=1 2238438Sdteske# 3247280Sdteske# Copyright (c) 2006-2013 Devin Teske 4238438Sdteske# All Rights Reserved. 5238438Sdteske# 6238438Sdteske# Redistribution and use in source and binary forms, with or without 7238438Sdteske# modification, are permitted provided that the following conditions 8238438Sdteske# are met: 9238438Sdteske# 1. Redistributions of source code must retain the above copyright 10238438Sdteske# notice, this list of conditions and the following disclaimer. 11238438Sdteske# 2. Redistributions in binary form must reproduce the above copyright 12238438Sdteske# notice, this list of conditions and the following disclaimer in the 13238438Sdteske# documentation and/or other materials provided with the distribution. 14238438Sdteske# 15238438Sdteske# THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 16238438Sdteske# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING BUT NOT LIMITED TO, THE 17238438Sdteske# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 18238438Sdteske# ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 19238438Sdteske# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 20238438Sdteske# DAMAGES (INLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 21238438Sdteske# OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 22238438Sdteske# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 23238438Sdteske# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 24238438Sdteske# OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 25238438Sdteske# SUCH DAMAGE. 26238438Sdteske# 27238438Sdteske# $FreeBSD: head/usr.sbin/bsdconfig/share/strings.subr 250702 2013-05-16 16:51:52Z dteske $ 28249751Sdteske# 29249751Sdteske############################################################ GLOBALS 30238438Sdteske 31249751Sdteske# 32249751Sdteske# Valid characters that can appear in an sh(1) variable name 33249751Sdteske# 34249751Sdteske# Please note that the character ranges A-Z and a-z should be avoided because 35249751Sdteske# these can include accent characters (which are not valid in a variable name). 36249751Sdteske# For example, A-Z matches any character that sorts after A but before Z, 37249751Sdteske# including A and Z. Although ASCII order would make more sense, that is not 38249751Sdteske# how it works. 39249751Sdteske# 40249751SdteskeVALID_VARNAME_CHARS="0-9ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz_" 41249751Sdteske 42249751Sdteske############################################################ FUNCTIONS 43249751Sdteske 44238438Sdteske# f_substr "$string" $start [ $length ] 45238438Sdteske# 46238438Sdteske# Simple wrapper to awk(1)'s `substr' function. 47238438Sdteske# 48238438Sdteskef_substr() 49238438Sdteske{ 50238438Sdteske local string="$1" start="${2:-0}" len="${3:-0}" 51238438Sdteske echo "$string" | awk "{ print substr(\$0, $start, $len) }" 52238438Sdteske} 53238438Sdteske 54250701Sdteske# f_snprintf $var_to_set $size $format ... 55250701Sdteske# 56250701Sdteske# Similar to snprintf(3), write at most $size number of bytes into $var_to_set 57250701Sdteske# using printf(1) syntax (`$format ...'). The value of $var_to_set is NULL 58250701Sdteske# unless at-least one byte is stored from the output. 59250701Sdteske# 60250701Sdteskef_snprintf() 61250701Sdteske{ 62250701Sdteske local __var_to_set="$1" __size="$2" 63250701Sdteske shift 2 # var_to_set/size 64250701Sdteske eval "$__var_to_set"=\$\( printf \"\$@\" \| awk -v max=\"\$__size\" \'' 65250701Sdteske { 66250701Sdteske len = length($0) 67250701Sdteske max -= len 68250701Sdteske print substr($0,0,(max > 0 ? len : max + len)) 69250701Sdteske if ( max < 0 ) exit 70250701Sdteske max-- 71250701Sdteske }'\' \) 72250701Sdteske} 73250701Sdteske 74238438Sdteske# f_longest_line_length 75238438Sdteske# 76238438Sdteske# Simple wrapper to an awk(1) script to print the length of the longest line of 77238438Sdteske# input (read from stdin). Supports the newline escape-sequence `\n' for 78238438Sdteske# splitting a single line into multiple lines. 79238438Sdteske# 80238438Sdteskef_longest_line_length_awk=' 81238438SdteskeBEGIN { longest = 0 } 82238438Sdteske{ 83238438Sdteske if (split($0, lines, /\\n/) > 1) 84238438Sdteske { 85238438Sdteske for (n in lines) 86238438Sdteske { 87238438Sdteske len = length(lines[n]) 88238438Sdteske longest = ( len > longest ? len : longest ) 89238438Sdteske } 90238438Sdteske } 91238438Sdteske else 92238438Sdteske { 93238438Sdteske len = length($0) 94238438Sdteske longest = ( len > longest ? len : longest ) 95238438Sdteske } 96238438Sdteske} 97238438SdteskeEND { print longest } 98238438Sdteske' 99238438Sdteskef_longest_line_length() 100238438Sdteske{ 101238438Sdteske awk "$f_longest_line_length_awk" 102238438Sdteske} 103238438Sdteske 104238438Sdteske# f_number_of_lines 105238438Sdteske# 106238438Sdteske# Simple wrapper to an awk(1) script to print the number of lines read from 107238438Sdteske# stdin. Supports newline escape-sequence `\n' for splitting a single line into 108238438Sdteske# multiple lines. 109238438Sdteske# 110238438Sdteskef_number_of_lines_awk=' 111238438SdteskeBEGIN { num_lines = 0 } 112238438Sdteske{ 113241700Sdteske num_lines += split(" "$0, unused, /\\n/) 114238438Sdteske} 115238438SdteskeEND { print num_lines } 116238438Sdteske' 117238438Sdteskef_number_of_lines() 118238438Sdteske{ 119238438Sdteske awk "$f_number_of_lines_awk" 120238438Sdteske} 121238438Sdteske 122238438Sdteske# f_isinteger $arg 123238438Sdteske# 124238438Sdteske# Returns true if argument is a positive/negative whole integer. 125238438Sdteske# 126238438Sdteskef_isinteger() 127238438Sdteske{ 128238438Sdteske local arg="$1" 129238438Sdteske 130238438Sdteske # Prevent division-by-zero 131238438Sdteske [ "$arg" = "0" ] && return $SUCCESS 132238438Sdteske 133238438Sdteske # Attempt to perform arithmetic divison (an operation which will exit 134238438Sdteske # with error unless arg is a valid positive/negative whole integer). 135238438Sdteske # 136240783Sdteske ( : $((0/$arg)) ) > /dev/null 2>&1 137238438Sdteske} 138238438Sdteske 139247280Sdteske# f_uriencode [$text] 140247280Sdteske# 141247280Sdteske# Encode $text for the purpose of embedding safely into a URL. Non-alphanumeric 142247280Sdteske# characters are converted to `%XX' sequence where XX represents the hexa- 143247280Sdteske# decimal ordinal of the non-alphanumeric character. If $text is missing, data 144247280Sdteske# is instead read from standard input. 145247280Sdteske# 146247280Sdteskef_uriencode_awk=' 147247280SdteskeBEGIN { 148247280Sdteske output = "" 149247280Sdteske for (n = 0; n < 256; n++) pack[sprintf("%c", n)] = sprintf("%%%02x", n) 150247280Sdteske} 151247280Sdteske{ 152247280Sdteske sline = "" 153247280Sdteske slen = length($0) 154247280Sdteske for (n = 1; n <= slen; n++) { 155247280Sdteske char = substr($0, n, 1) 156247280Sdteske if ( char !~ /^[[:alnum:]_]$/ ) char = pack[char] 157247280Sdteske sline = sline char 158247280Sdteske } 159247280Sdteske output = output ( output ? "%0a" : "" ) sline 160247280Sdteske} 161247280SdteskeEND { print output } 162247280Sdteske' 163247280Sdteskef_uriencode() 164247280Sdteske{ 165247280Sdteske if [ $# -gt 0 ]; then 166247280Sdteske echo "$1" | awk "$f_uriencode_awk" 167247280Sdteske else 168247280Sdteske awk "$f_uriencode_awk" 169247280Sdteske fi 170247280Sdteske} 171247280Sdteske 172247280Sdteske# f_uridecode [$text] 173247280Sdteske# 174247280Sdteske# Decode $text from a URI. Encoded characters are converted from their `%XX' 175247280Sdteske# sequence into original unencoded ASCII sequences. If $text is missing, data 176247280Sdteske# is instead read from standard input. 177247280Sdteske# 178247280Sdteskef_uridecode_awk=' 179247280SdteskeBEGIN { for (n = 0; n < 256; n++) chr[n] = sprintf("%c", n) } 180247280Sdteske{ 181247280Sdteske sline = "" 182247280Sdteske slen = length($0) 183247280Sdteske for (n = 1; n <= slen; n++) 184247280Sdteske { 185247280Sdteske seq = substr($0, n, 3) 186247280Sdteske if ( seq ~ /^%[[:xdigit:]][[:xdigit:]]$/ ) { 187247280Sdteske hex = substr(seq, 2, 2) 188247280Sdteske sline = sline chr[sprintf("%u", "0x"hex)] 189247280Sdteske n += 2 190247280Sdteske } else 191247280Sdteske sline = sline substr(seq, 1, 1) 192247280Sdteske } 193247280Sdteske print sline 194247280Sdteske} 195247280Sdteske' 196247280Sdteskef_uridecode() 197247280Sdteske{ 198247280Sdteske if [ $# -gt 0 ]; then 199247280Sdteske echo "$1" | awk "$f_uridecode_awk" 200247280Sdteske else 201247280Sdteske awk "$f_uridecode_awk" 202247280Sdteske fi 203247280Sdteske} 204247280Sdteske 205249751Sdteske# f_replaceall $string $find $replace [$var_to_set] 206249751Sdteske# 207250702Sdteske# Replace all occurrences of $find in $string with $replace. If $var_to_set is 208249751Sdteske# either missing or NULL, the variable name is produced on standard out for 209249751Sdteske# capturing in a sub-shell (which is less recommended due to performance 210249751Sdteske# degradation). 211249751Sdteske# 212249751Sdteskef_replaceall() 213249751Sdteske{ 214249751Sdteske local __left="" __right="$1" 215249751Sdteske local __find="$2" __replace="$3" __var_to_set="$4" 216249751Sdteske while :; do 217249751Sdteske case "$__right" in *$__find*) 218249751Sdteske __left="$__left${__right%%$__find*}$__replace" 219249751Sdteske __right="${__right#*$__find}" 220249751Sdteske continue 221249751Sdteske esac 222249751Sdteske break 223249751Sdteske done 224249751Sdteske __left="$__left${__right#*$__find}" 225249751Sdteske if [ "$__var_to_set" ]; then 226249751Sdteske setvar "$__var_to_set" "$__left" 227249751Sdteske else 228249751Sdteske echo "$__left" 229249751Sdteske fi 230249751Sdteske} 231249751Sdteske 232249751Sdteske# f_str2varname $string [$var_to_set] 233249751Sdteske# 234249751Sdteske# Convert a string into a suitable value to be used as a variable name 235249751Sdteske# by converting unsuitable characters into the underscrore [_]. If $var_to_set 236249751Sdteske# is either missing or NULL, the variable name is produced on standard out for 237249751Sdteske# capturing in a sub-shell (which is less recommended due to performance 238249751Sdteske# degradation). 239249751Sdteske# 240249751Sdteskef_str2varname() 241249751Sdteske{ 242249751Sdteske local __string="$1" __var_to_set="$2" 243249751Sdteske f_replaceall "$__string" "[!$VALID_VARNAME_CHARS]" "_" "$__var_to_set" 244249751Sdteske} 245249751Sdteske 246249751Sdteske# f_shell_escape $string [$var_to_set] 247249751Sdteske# 248249751Sdteske# Escape $string for shell eval statement(s) by replacing all single-quotes 249249751Sdteske# with a special sequence that creates a compound string when interpolated 250249751Sdteske# by eval with surrounding single-quotes. 251249751Sdteske# 252249751Sdteske# For example: 253249751Sdteske# 254249751Sdteske# foo="abc'123" 255249751Sdteske# f_shell_escape "$foo" bar # bar=[abc'\''123] 256249751Sdteske# eval echo \'$foo\' # produces abc'123 257249751Sdteske# 258249751Sdteske# This is helpful when processing an argument list that has to retain its 259249751Sdteske# escaped structure for later evaluations. 260249751Sdteske# 261249751Sdteske# WARNING: Surrounding single-quotes are not added; this is the responsibility 262249751Sdteske# of the code passing the escaped values to eval (which also aids readability). 263249751Sdteske# 264249751Sdteskef_shell_escape() 265249751Sdteske{ 266249751Sdteske local __string="$1" __var_to_set="$2" 267249751Sdteske f_replaceall "$__string" "'" "'\\''" "$__var_to_set" 268249751Sdteske} 269249751Sdteske 270249751Sdteske# f_shell_unescape $string [$var_to_set] 271249751Sdteske# 272249751Sdteske# The antithesis of f_shell_escape(), this function takes an escaped $string 273249751Sdteske# and expands it. 274249751Sdteske# 275249751Sdteske# For example: 276249751Sdteske# 277249751Sdteske# foo="abc'123" 278249751Sdteske# f_shell_escape "$foo" bar # bar=[abc'\''123] 279249751Sdteske# f_shell_unescape "$bar" # produces abc'123 280249751Sdteske# 281249751Sdteskef_shell_unescape() 282249751Sdteske{ 283249751Sdteske local __string="$1" __var_to_set="$2" 284249751Sdteske f_replaceall "$__string" "'\\''" "'" "$__var_to_set" 285249751Sdteske} 286249751Sdteske 287249751Sdteske############################################################ MAIN 288249751Sdteske 289244675Sdteskef_dprintf "%s: Successfully loaded." strings.subr 290244675Sdteske 291238438Sdteskefi # ! $_STRINGS_SUBR 292