我试图写一个bash脚本测试,需要一个参数,并通过curl发送到网站。我需要url编码的值,以确保特殊字符被正确处理。最好的方法是什么?
这是我到目前为止的基本脚本:
#!/bin/bash
host=${1:?'bad host'}
value=$2
shift
shift
curl -v -d "param=${value}" http://${host}/somepath $@
我试图写一个bash脚本测试,需要一个参数,并通过curl发送到网站。我需要url编码的值,以确保特殊字符被正确处理。最好的方法是什么?
这是我到目前为止的基本脚本:
#!/bin/bash
host=${1:?'bad host'}
value=$2
shift
shift
curl -v -d "param=${value}" http://${host}/somepath $@
当前回答
awk版本的直接链接:http://www.shelldorado.com/scripts/cmds/urlencode 我用了很多年了,效果很好
:
##########################################################################
# Title : urlencode - encode URL data
# Author : Heiner Steven (heiner.steven@odn.de)
# Date : 2000-03-15
# Requires : awk
# Categories : File Conversion, WWW, CGI
# SCCS-Id. : @(#) urlencode 1.4 06/10/29
##########################################################################
# Description
# Encode data according to
# RFC 1738: "Uniform Resource Locators (URL)" and
# RFC 1866: "Hypertext Markup Language - 2.0" (HTML)
#
# This encoding is used i.e. for the MIME type
# "application/x-www-form-urlencoded"
#
# Notes
# o The default behaviour is not to encode the line endings. This
# may not be what was intended, because the result will be
# multiple lines of output (which cannot be used in an URL or a
# HTTP "POST" request). If the desired output should be one
# line, use the "-l" option.
#
# o The "-l" option assumes, that the end-of-line is denoted by
# the character LF (ASCII 10). This is not true for Windows or
# Mac systems, where the end of a line is denoted by the two
# characters CR LF (ASCII 13 10).
# We use this for symmetry; data processed in the following way:
# cat | urlencode -l | urldecode -l
# should (and will) result in the original data
#
# o Large lines (or binary files) will break many AWK
# implementations. If you get the message
# awk: record `...' too long
# record number xxx
# consider using GNU AWK (gawk).
#
# o urlencode will always terminate it's output with an EOL
# character
#
# Thanks to Stefan Brozinski for pointing out a bug related to non-standard
# locales.
#
# See also
# urldecode
##########################################################################
PN=`basename "$0"` # Program name
VER='1.4'
: ${AWK=awk}
Usage () {
echo >&2 "$PN - encode URL data, $VER
usage: $PN [-l] [file ...]
-l: encode line endings (result will be one line of output)
The default is to encode each input line on its own."
exit 1
}
Msg () {
for MsgLine
do echo "$PN: $MsgLine" >&2
done
}
Fatal () { Msg "$@"; exit 1; }
set -- `getopt hl "$@" 2>/dev/null` || Usage
[ $# -lt 1 ] && Usage # "getopt" detected an error
EncodeEOL=no
while [ $# -gt 0 ]
do
case "$1" in
-l) EncodeEOL=yes;;
--) shift; break;;
-h) Usage;;
-*) Usage;;
*) break;; # First file name
esac
shift
done
LANG=C export LANG
$AWK '
BEGIN {
# We assume an awk implementation that is just plain dumb.
# We will convert an character to its ASCII value with the
# table ord[], and produce two-digit hexadecimal output
# without the printf("%02X") feature.
EOL = "%0A" # "end of line" string (encoded)
split ("1 2 3 4 5 6 7 8 9 A B C D E F", hextab, " ")
hextab [0] = 0
for ( i=1; i<=255; ++i ) ord [ sprintf ("%c", i) "" ] = i + 0
if ("'"$EncodeEOL"'" == "yes") EncodeEOL = 1; else EncodeEOL = 0
}
{
encoded = ""
for ( i=1; i<=length ($0); ++i ) {
c = substr ($0, i, 1)
if ( c ~ /[a-zA-Z0-9.-]/ ) {
encoded = encoded c # safe character
} else if ( c == " " ) {
encoded = encoded "+" # special handling
} else {
# unsafe character, encode it as a two-digit hex-number
lo = ord [c] % 16
hi = int (ord [c] / 16);
encoded = encoded "%" hextab [hi] hextab [lo]
}
}
if ( EncodeEOL ) {
printf ("%s", encoded EOL)
} else {
print encoded
}
}
END {
#if ( EncodeEOL ) print ""
}
' "$@"
其他回答
这可能是最好的一个:
after=$(echo -e "$before" | od -An -tx1 | tr ' ' % | xargs printf "%s")
这是orwellophile回答的ksh版本,包含rawurlencode和rawurldecode函数(链接:如何为curl命令urlencode数据?)我没有足够的代表发表评论,因此有了新的帖子。
#!/bin/ksh93
function rawurlencode
{
typeset string="${1}"
typeset strlen=${#string}
typeset encoded=""
for (( pos=0 ; pos<strlen ; pos++ )); do
c=${string:$pos:1}
case "$c" in
[-_.~a-zA-Z0-9] ) o="${c}" ;;
* ) o=$(printf '%%%02x' "'$c")
esac
encoded+="${o}"
done
print "${encoded}"
}
function rawurldecode
{
printf $(printf '%b' "${1//%/\\x}")
}
print $(rawurlencode "C++") # --> C%2b%2b
print $(rawurldecode "C%2b%2b") # --> C++
url=$(echo "$1" | sed -e 's/%/%25/g' -e 's/ /%20/g' -e 's/!/%21/g' -e 's/"/%22/g' -e 's/#/%23/g' -e 's/\$/%24/g' -e 's/\&/%26/g' -e 's/'\''/%27/g' -e 's/(/%28/g' -e 's/)/%29/g' -e 's/\*/%2a/g' -e 's/+/%2b/g' -e 's/,/%2c/g' -e 's/-/%2d/g' -e 's/\./%2e/g' -e 's/\//%2f/g' -e 's/:/%3a/g' -e 's/;/%3b/g' -e 's//%3e/g' -e 's/?/%3f/g' -e 's/@/%40/g' -e 's/\[/%5b/g' -e 's/\\/%5c/g' -e 's/\]/%5d/g' -e 's/\^/%5e/g' -e 's/_/%5f/g' -e 's/`/%60/g' -e 's/{/%7b/g' -e 's/|/%7c/g' -e 's/}/%7d/g' -e 's/~/%7e/g')
这将对$1中的字符串进行编码,并将其输出到$url中。尽管你不需要把它放在var中。顺便说一句,没有包括sed for选项卡,认为它会把它变成空格
Orwellophile给出了一个很好的答案,它包含了一个纯bash选项(函数rawurlencode),我在我的网站上使用过(基于shell的CGI脚本,响应搜索请求的大量url)。唯一的缺点是高峰期间CPU过高。
我找到了一个改进的解决方案,利用bash的“全局替换”特性。有了这个解决方案,url编码的处理时间快了4倍。解决方案确定要转义的字符,并使用“全局替换”操作符(${var//source/replacement})来处理所有替换。这种速度的提高显然来自于使用bash内部循环,而不是显式循环。
性能:核心i3-8100 3.60Ghz。测试用例:来自堆栈溢出的1000个URL,类似于这个票据:“https://stackoverflow.com/questions/296536/how-to-urlencode-data-for-curl-command”。
现有解决方案:0.807秒 优化方案:0.162秒(5倍加速)
url_encode()
{
local key="${1}" varname="${2:-_rval}" prefix="${3:-_ENCKEY_}"
local unsafe=${key//[-_.~a-zA-Z0-9 ]/}
local -i key_len=${#unsafe}
local ch ch1 ch0
while [ "$unsafe" ] ;do
ch=${unsafe:0:1}
ch0="\\$ch"
printf -v ch1 '%%%02x' "'$ch'"
key=${key//$ch0/"$ch1"}
unsafe=${unsafe//"$ch0"}
done
key=${key// /+}
REPLY="$key"
# printf "%s" "$REPLY"
return 0
}
作为一个次要的额外字符,它使用'+'来编码空格。稍微紧凑的URL。
基准:
function t {
local key
for (( i=1 ; i<=$1 ; i++ )) do url_encode "$2" kkk2 ; done
echo "K=$REPLY"
}
t 1000 "https://stackoverflow.com/questions/296536/how-to-urlencode-data-for-curl-command"
有什么能比javascript更好地解析url呢?
node -p "encodeURIComponent('$url')"