X-Git-Url: https://git.plutz.net/?p=cgilite;a=blobdiff_plain;f=cgilite.sh;h=2193e29ff90069b181db9475c439c89f34e5f529;hp=86ad28073018c9a2626877e74426c0b2d4a95914;hb=5038774f89b33ecd03e407898e9d57d38245b5bb;hpb=5b013b64b7bbc9a62775ed861e9c8b34ffb3dfaa diff --git a/cgilite.sh b/cgilite.sh index 86ad280..2193e29 100755 --- a/cgilite.sh +++ b/cgilite.sh @@ -1,6 +1,6 @@ #!/bin/sh -# Copyright 2017 - 2020 Paul Hänsch +# Copyright 2017 - 2021 Paul Hänsch # # This is CGIlite. # A collection of posix shell functions for writing CGI scripts. @@ -18,17 +18,48 @@ # You should have received a copy of the GNU Affero General Public License # along with CGIlite. If not, see . +[ -n "$include_cgilite" ] && return 0 +# guard set after webserver part + # ksh and zsh workaround # set -o posix # ksh, not portable setopt -o OCTAL_ZEROES 2>&- +# Integrated webserver request timeout +cgilite_timeout=2 + +# General environment variables +# $_EXEC - directory containing application itself +# $_DATA - direcotry where application data may be stored +# $_BASE - optional prefix for http path, e.g. "/myapp" +# +# Programmers should take care to use those variables throughout the +# application. +# Variables may be set via CLI argument, in environment, or left as default. + +for cgilite_arg in "$@"; do case $cgilite_arg in + --exec=*) _EXEC="${cgilite_arg#*=}";; + --data=*) _DATA="${cgilite_arg#*=}";; + --base=*) _BASE="${cgilite_arg#*=}";; +esac; done +unset cgilite_arg + +_EXEC="${_EXEC:-${0%/*}}" +_DATA="${_DATA:-.}" +_EXEC="${_EXEC%/}" _DATA="${_DATA%/}" _BASE="${_BASE%/}" + +export _EXEC _DATA _BASE + +# Carriage Return and Line Break characters for convenience CR=" " BR=' ' -cgilite_timeout=2 PATH(){ local str seg out + # normalize path + # read from stdin if no arguments are provided + [ $# -eq 0 ] && str="$(cat)" || str="$*" while [ "$str" ]; do seg=${str%%/*}; str="${str#*/}" @@ -42,18 +73,59 @@ PATH(){ [ "${str}" -a "${out}" ] && printf %s "$out" || printf %s/ "${out%/}" } -HEX_DECODE=' - s;\\;\\\\;g; :HEXDECODE_X; s;%([^0-9A-F]);\\045\1;g; tHEXDECODE_X; - # Hexadecimal { %00 - %FF } will be transformed to octal { \000 - \377 } for posix printf - s;%[0123].;&\\0;g; s;%[4567].;&\\1;g; s;%[89AB].;&\\2;g; s;%[CDEF].;&\\3;g; - s;%[048C][0-7]\\.;&0;g; s;%[048C][89A-F]\\.;&1;g; s;%[159D][0-7]\\.;&2;g; s;%[159D][89A-F]\\.;&3;g; - s;%[26AE][0-7]\\.;&4;g; s;%[26AE][89A-F]\\.;&5;g; s;%[37BF][0-7]\\.;&6;g; s;%[37BF][89A-F]\\.;&7;g; - s;%.[08](\\..);\10;g; s;%.[19](\\..);\11;g; s;%.[2A](\\..);\12;g; s;%.[3B](\\..);\13;g; - s;%.[4C](\\..);\14;g; s;%.[5D](\\..);\15;g; s;%.[6E](\\..);\16;g; s;%.[7F](\\..);\17;g; -' - HEX_DECODE(){ - printf -- "$(printf %s "$1" |sed -E "$HEX_DECODE")" + local pfx="$1" in="$2" out + # Print out Data encoded as Hex + # + # Arguments: + # pfx - required, prefix for a hex tupel, e.g. "\x", "%" "\", may be empty + # in - required, string to be decoded + # + # anything that does not constitute a tupel of valid Hex numerals + # will be copied to the output literally + + while [ "$in" ]; do + case $in in + "$pfx"[0-9a-fA-F][0-9a-fA-F]*) in="${in#${pfx}}";; + \\*) in="${in#?}"; out="${out}\\\\"; continue;; + %*) in="${in#?}"; out="${out}%%"; continue;; + *) att="${in%%"${pfx}"*}"; att="${att%%%*}"; att="${att%%\\*}" + out="${out}${att}"; in="${in#"${att}"}"; continue;; + esac; + + # Hex escapes for printf (e.g. \x41) are not portable + # The portable way for Hex output is transforming Hex to Octal + # (e.g. \x41 = \101) + case $in in + [0123]?*) out="${out}\\0";; + [4567]?*) out="${out}\\1";; + [89aAbB]?*) out="${out}\\2";; + [c-fC-F]?*) out="${out}\\3";; + esac + case $in in + [048cC][0-7]*) out="${out}0";; + [048cC][89a-fA-F]*) out="${out}1";; + [159dD][0-7]*) out="${out}2";; + [159dD][89a-fA-F]*) out="${out}3";; + [26aAeE][0-7]*) out="${out}4";; + [26aAeE][89a-fA-F]*) out="${out}5";; + [37bBfF][0-7]*) out="${out}6";; + [37bBfF][89a-fA-F]*) out="${out}7";; + esac + case $in in + ?[08]*) out="${out}0";; + ?[19]*) out="${out}1";; + ?[2aA]*) out="${out}2";; + ?[3bB]*) out="${out}3";; + ?[4cC]*) out="${out}4";; + ?[5dD]*) out="${out}5";; + ?[6eE]*) out="${out}6";; + ?[7fF]*) out="${out}7";; + esac + in="${in#?}" + in="${in#?}" + done + printf -- "$out" } if [ -z "$REQUEST_METHOD" ]; then @@ -72,15 +144,17 @@ if [ -z "$REQUEST_METHOD" ]; then (sleep $cgilite_timeout && kill $$) & cgilite_watchdog=$! while read REQUEST_METHOD REQUEST_URI SERVER_PROTOCOL; do + unset PATH_INFO QUERY_STRING cgilite_headers CONTENT_LENGTH CONTENT_TYPE + [ "${SERVER_PROTOCOL#HTTP/1.[01]${CR}}" ] && break kill $cgilite_watchdog SERVER_PROTOCOL="${SERVER_PROTOCOL%${CR}}" - PATH_INFO="$(HEX_DECODE "${REQUEST_URI%\?*}" |PATH)" + PATH_INFO="$(HEX_DECODE % "${REQUEST_URI%\?*}" |PATH)" [ "${REQUEST_URI}" = "${REQUEST_URI#*\?}" ] \ && QUERY_STRING='' \ || QUERY_STRING="${REQUEST_URI#*\?}" - cgilite_headers=''; while read -r hl; do + while read -r hl; do hl="${hl%${CR}}"; [ "$hl" ] || break case $hl in 'Content-Length: '*) CONTENT_LENGTH="${hl#*: }";; @@ -90,7 +164,7 @@ if [ -z "$REQUEST_METHOD" ]; then done export REMOTE_ADDR SERVER_NAME SERVER_PORT REQUEST_METHOD REQUEST_URI SERVER_PROTOCOL \ - PATH_INFO QUERY_STRING CONTENT_TYPE CONTENT_LENGTH + PATH_INFO QUERY_STRING CONTENT_TYPE CONTENT_LENGTH cgilite_headers # Try to serve multiple requests, provided that script serves a # Content-Length header. @@ -119,13 +193,20 @@ if [ -z "$REQUEST_METHOD" ]; then exit 0 fi +include_cgilite="$0" + if [ "${REQUEST_METHOD}" = POST -a "${CONTENT_LENGTH:-0}" -gt 0 -a \ "${CONTENT_TYPE}" = "application/x-www-form-urlencoded" ]; then cgilite_post="$(head -c "$CONTENT_LENGTH")" fi +PATH_INFO="$(PATH "/${PATH_INFO#${_BASE}}")" + +debug(){ [ $# -gt 0 ] && printf '%s\n' "$@" >&2 || tee -a /dev/stderr; } [ "${DEBUG+x}" ] && env >&2 +# general helper functions, see GET, POST, and REF below + cgilite_count(){ printf %s "&$1" \ | grep -oE '&'"$2"'=[^&]*' \ @@ -139,7 +220,7 @@ cgilite_value(){ str="${str#*&${name}=}" cnt=$((cnt - 1)) done - printf -- "$(printf %s "${str%%&*}" |sed -E 's;\+; ;g;'"$HEX_DECODE")" + HEX_DECODE % "$(printf %s "${str%%&*}" |tr + \ )" } cgilite_keys(){ @@ -151,15 +232,26 @@ cgilite_keys(){ | sort -u } -GET(){ cgilite_value "${QUERY_STRING}" $@; } +# Read arguments from GET, POST, or the query string of the referrer (REF). +# Example: +# GET varname n +# +# where n is number for the Nth occurence of a variable and defaults to 1 +# +# *_COUNT varname +# -> returns number of ocurences +# *_KEYS +# -> returns list of available varnames + +GET(){ cgilite_value "${QUERY_STRING}" "$@"; } GET_COUNT(){ cgilite_count "${QUERY_STRING}" $1; } GET_KEYS(){ cgilite_keys "${QUERY_STRING}"; } -POST(){ cgilite_value "${cgilite_post}" $@; } +POST(){ cgilite_value "${cgilite_post}" "$@"; } POST_COUNT(){ cgilite_count "${cgilite_post}" $1; } POST_KEYS(){ cgilite_keys "${cgilite_post}"; } -REF(){ cgilite_value "${HTTP_REFERER#*\?}" $@; } +REF(){ cgilite_value "${HTTP_REFERER#*\?}" "$@"; } REF_COUNT(){ cgilite_count "${HTTP_REFERER#*\?}" $1; } REF_KEYS(){ cgilite_keys "${HTTP_REFERER#*\?}"; } @@ -179,7 +271,8 @@ HEADER(){ } COOKIE(){ - HEX_DECODE "$( + # Read value of cookie + HEX_DECODE % "$( HEADER Cookie \ | grep -oE '(^|; ?)'"$1"'=[^;]*' \ | sed -En "${2:-1}"'{s;^[^=]+=;;; s;\+; ;g; p;}' @@ -187,35 +280,45 @@ COOKIE(){ } HTML(){ - # HTML Entity Coding - # Prints UTF-8 string as decimal Unicode Code Points - # Useful for escaping user input for use in HTML text and attributes - { [ $# -eq 0 ] && cat || printf %s "$*"; } \ - | hexdump -ve '/1 "%03o\n"' \ - | while read n; do - case $n in - # bitbanging octal UTF-8 chains into singular 7 digit octal numbers - [01]??) printf '0000%s' $n;; # 7 bit ASCII character, nothing to do - 2??) printf '%s' ${n#2};; # tail fragment, append 6 bit - 3[0123]?) printf '000%s' ${n#3};; # 2 octet (11 bit) chain start - 34?) printf '00%s' ${n#34};; # 3 octet (16 bit) chain start - 35?) printf '01%s' ${n#35};; # 3 octet chain start, high - 36?) printf '%s' ${n#36};; # 4 octet (21 bit) chain start - esac - done \ - | sed -E 's;.{7};&\n;g;' \ - | while read n; do - printf '&#%d;' $((0$n)) - done + # Escape HTML cahracters + # Also escape [, ], and \n for use in html-sh + local str out + [ $# -eq 0 ] && str="$(cat)" || str="$*" + while [ "$str" ]; do case $str in + \&*) out="${out}&"; str="${str#?}";; + \<*) out="${out}<"; str="${str#?}";; + \>*) out="${out}>"; str="${str#?}";; + \"*) out="${out}""; str="${str#?}";; + \'*) out="${out}'"; str="${str#?}";; + \[*) out="${out}["; str="${str#?}";; + \]*) out="${out}]"; str="${str#?}";; + "${CR}"*) out="${out} "; str="${str#?}";; + "${BR}"*) out="${out} "; str="${str#?}";; + *) out="${out}${str%%[]&<>\"\'${CR}${BR}[]*}"; str="${str#"${str%%[]&<>\"\'${CR}${BR}[]*}"}";; + esac; done + printf %s "$out" } URL(){ - # Code every character in URL escape hex format - # except alphanumeric ascii - - { [ $# -eq 0 ] && cat || printf %s "$*"; } \ - | hexdump -v -e '/1 ",%02X"' \ - | sed 's;,;%;g; s;%2F;/;g;' + # Escape pathes, so they can be used in link tags and HTTP Headers + local str out + [ $# -eq 0 ] && str="$(cat)" || str="$*" + while [ "$str" ]; do case $str in + \&*) out="${out}%26"; str="${str#?}";; + \"*) out="${out}%22"; str="${str#?}";; + \'*) out="${out}%27"; str="${str#?}";; + \?*) out="${out}%3F"; str="${str#?}";; + \#*) out="${out}%23"; str="${str#?}";; + \[*) out="${out}%5B"; str="${str#?}";; + \]*) out="${out}%5D"; str="${str#?}";; + \ *) out="${out}%20"; str="${str#?}";; + " "*) out="${out}%09"; str="${str#?}";; + "${CR}"*) out="${out}%0D"; str="${str#?}";; + "${BR}"*) out="${out}%0A"; str="${str#?}";; + %*) out="${out}%25"; str="${str#?}";; + *) out="${out}${str%%[]&\"\'\?# ${CR}${BR}%[]*}"; str="${str#"${str%%[]&\"\'\?# ${CR}${BR}%[]*}"}";; + esac; done + printf %s "$out" } SET_COOKIE(){ @@ -237,6 +340,7 @@ SET_COOKIE(){ } REDIRECT(){ + # Trigger redirct and terminate script printf '%s: %s\r\n' \ Status "303 See Other" \ Content-Length 0 \