readlink replacement

[feherke]

Hi

The [tt]readlink[/tt] command is usefull to find out the target of a symbolic link. Any idea for similar commands, in case [tt]readlink[/tt] is not available ?

Ok, I can write a function, thats not a problem. ( No error checking in below codes. )

readlink() { ls -ld "$1" | sed 's/.*-> //'; }
[gray]# or[/gray]
readlink() { find "$1" -mindepth 0 -maxdepth 0 -printf "%l\n"; }

But while a [tt]readlink()[/tt] system call is available, I hope there is a somehow standard and faster way to do this.

Thanks.

Feherke.

http://rootshell.be/~feherke/

 

[Annihilannic]

Under what OS's is readlink available as a command, out of curiosity?

I'm not aware of anything else to do that, I would have suggested your first function, or ls -L if you just needed to get the attributes of the targeted file, or check for its existence.

Annihilannic.

 

[feherke]

Hi

Is in the GNU coreutils package. In the ancestors of coreutils, the shellutils, fileutils and textutils packages was not.

Feherke.

http://rootshell.be/~feherke/

 

[mrn]

use -follow with the find command.

Mike

"A foolproof method for sculpting an elephant: first, get a huge block of marble, then you chip away everything that doesn't look like an elephant."

http://www.airport-parking-site.co.uk/

 

[feherke]

Hu

I do not want to follow the link. I want to get its target.


Feherke.

http://rootshell.be/~feherke/

 

[mrn]

PN=`basename "$0"` # Resolvelink


MAXHOPS=30

Usage () {
echo >&2 "$PN - resolve symbolic links.
usage: $PN [file | directory] [...]

Each line starts with an classification character:
# absolute path name of the starting link
> link to another link
+ link resolved (entry is listable)
- link resolved (entry is invalid)
! error, i.e. too many hops (more than $MAXHOPS)"
exit 1
}

Msg () {
for MsgLine
do echo >&2 "$PN: $MsgLine"
done
}

Fatal () { Msg "$@"; exit 1; }

RelToAbs () {
for RelPath
do
Dir=`dirname "$RelPath"`
Dir=`cd $Dir; pwd`
# Work around for SOLARIS 2.5 ksh93 bug ("cd" prints new directory)
Dir=`echo "$Dir" | sed q`
case "$RelPath" in
*/) Base=`basename \`echo "$RelPath"\` | sed 's:/[/]*$::'`;;
*) Base=`basename "$RelPath"`;;
esac
echo "$Dir/$Base"
done
}

Resolve () {
for SymLink
do
L=`RelToAbs "$SymLink"`
Indent=
HopCount=0
(
while [ -h "$L" ]
do
if [ -z "$Indent" ]
then
echo "# $L"
Indent="> "
else
echo "${Indent}$L `RelToAbs \"$L\"`"
fi

ParentDir=`dirname "$L"`
[ -n "$ParentDir" -a -d "$ParentDir" ] || {
Msg "could not determine parent directory for $L"
break
}
cd "$ParentDir"
BaseName=`basename "$L"`
[ -n "$BaseName" ] || {
Msg "could not determine basename for $L"
break
}
Ref=`ls -ld "$BaseName" | awk '{print $NF}'`
if [ -n "$Ref" ]
then
if [ "$L" != "$Ref" ]
then
L="$Ref"
else # direct loop (i.e. "a -> a")
echo "! loop detected - $Ref (`pwd`)"
break
fi
else
Msg "could not list link $L - should not occur"
break;
fi
HopCount=`expr $HopCount + 1`
if [ ${HopCount:=0} -gt $MAXHOPS ]
then
echo "! too many links (more than $MAXHOPS): $Ref (`pwd`)"
break
fi
done

# Ref contains the last (relative) path of the symbolical
# link chain

if [ X`ls -d "$Ref" 2>/dev/null` != X ]
then
echo "+ $Ref `RelToAbs $Ref`"
else
echo "- $Ref"
fi
)
done
}

verbose=no
set -- `getopt h "$@"`
while [ $# -gt 0 ]
do
case "$1" in
-v) verbose=yes;;
--) shift; break;;
-h) Usage;;
-*) Usage;;
*) break;; # First file name
esac
shift
done

[ $# -lt 1 ] && set -- * # Default: current directory

find "$@" -type l -print |
while read Link ignore
do
[ -h "$Link" ] || continue
Resolve "$Link"
done

Mike

"A foolproof method for sculpting an elephant: first, get a huge block of marble, then you chip away everything that doesn't look like an elephant."

http://www.airport-parking-site.co.uk/