如何递归解析符号链接而无需readlink或realpath? [英] How to recursively resolve symlinks without readlink or realpath?
问题描述
如果readlink
和realpath
不可用,编写脚本以找到链接目标的最佳便携式(POSIX?)方法是什么?
What's the best portable (POSIX?) way to script finding the target for a link if readlink
and realpath
are not available?
您是否要ls -l
,如果它以l
开头,请在->
之后以sed
开头并重复,直到不再以l
开头为止.
Would you ls -l
, and if it starts with l
take the text after the ->
with sed
and repeat until it no longer starts with l
?
推荐答案
每 BashFAQ#29 (它也支持@EugeniuRosca建议的建议的GNU查找方法):
Per BashFAQ #29 (which also endorses the GNU find approach suggested by @EugeniuRosca):
一个广泛可用的(虽然不是纯POSIX)选项是使用perl
:
One widely available (though not pure-POSIX) option is to use perl
:
target=/path/to/symlink-name perl -le 'print readlink $ENV{target}'
如果保证符号链接的名称不包含->
,则可以解析ls
的输出.
If your symlink's name is guaranteed not to contain ->
, you can parse the output of ls
.
以下代码结合了两种方法:
The below code combines both approaches:
# define the best readlink function available for this platform
if command -v readlink >/dev/null 2>/dev/null; then
# first choice: Use the real readlink command
readlink() {
command readlink -- "$@"
}
elif find . -maxdepth 0 -printf '%l' >/dev/null 2>/dev/null; then
# second choice: use GNU find
readlink() {
local ll candidate >/dev/null 2>&1 ||:
if candidate=$(find "$1" -maxdepth 0 -printf '%l') && [ "$candidate" ]; then
printf '%s\n' "$candidate"
else
printf '%s\n' "$1"
fi
}
elif command -v perl >/dev/null 2>/dev/null; then
# third choice: use perl
readlink() {
local candidate ||:
candidate=$(target=$1 perl -le 'print readlink $ENV{target}')
if [ "$candidate" ]; then
printf '%s\n' "$candidate"
else
printf '%s\n' "$1"
fi
}
else
# fourth choice: parse ls -ld
readlink() {
local ll candidate >/dev/null 2>&1 ||:
ll=$(LC_ALL=C ls -ld -- "$1" 2>/dev/null)
candidate=${ll#* -> }
if [ "$candidate" = "$ll" ]; then
printf '%s\n' "$1"
else
printf '%s\n' "$candidate"
fi
}
fi
readlink_recursive() {
local path prev_path oldwd found_recursion >/dev/null 2>&1 ||:
oldwd=$PWD; path=$1; found_recursion=0
while [ -L "$path" ] && [ "$found_recursion" = 0 ]; do
if [ "$path" != "${path%/*}" ]; then
cd -- "${path%/*}" || {
cd -- "$oldwd" ||:
echo "ERROR: Directory '${path%/*}' does not exist in '$PWD'" >&2
return 1
}
path=${PWD}/${path##*/}
fi
path=$(readlink "$path")
if [ -d "$path" ]; then
cd -- "$path"
path=$PWD
break
fi
if [ "$path" != "${path%/*}" ]; then
cd -- "${path%/*}" || {
echo "ERROR: Could not traverse from $PWD to ${path%/*}" >&2
return 1
}
path=${PWD}/${path##*/}
elif [ "$PWD" != "$oldwd" ]; then
path=${PWD}/$path
fi
for prev_path; do
if [ "$path" = "$prev_path" ]; then
found_recursion=1
break
fi
done
set -- "$path" "$@" # record path for recursion check
done
if [ "$path" != "${path%/../*}" ]; then
cd "${path%/*}" || {
echo "ERROR: Directory '${path%/*}' does not exist in $PWD" >&2
return 1
}
printf '%s\n' "$PWD/${path##*/}"
else
printf '%s\n' "$path"
fi
cd -- "$oldwd" ||:
}
这篇关于如何递归解析符号链接而无需readlink或realpath?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!