From dda942d7f1e9a2c2b01fd072f2c59fc29d5c5c9b Mon Sep 17 00:00:00 2001
From: Johannes 'josch' Schauer <josch@mister-muffin.de>
Date: Sun, 22 Mar 2020 09:53:39 +0100
Subject: [PATCH] add realpath_m.sh as an all-in-one replacement for realpath
 -m

---
 README.md     |  2 +-
 realpath.sh   |  2 +-
 realpath_m.sh | 80 +++++++++++++++++++++++++++++++++++++++++++++++++++
 test.sh       | 42 ++++++++++++++++-----------
 4 files changed, 108 insertions(+), 18 deletions(-)
 create mode 100755 realpath_m.sh

diff --git a/README.md b/README.md
index 7688247..dee1921 100644
--- a/README.md
+++ b/README.md
@@ -35,4 +35,4 @@ To verify that the script works with busybox:
 
     $ mkdir busybox
     $ busybox --install -s busybox
-    $ PATH=$(pwd)/busybox ./test.sh
+    $ PATH=$(pwd)/busybox /bin/busybox sh ./test.sh
diff --git a/realpath.sh b/realpath.sh
index 855df84..157dcd1 100755
--- a/realpath.sh
+++ b/realpath.sh
@@ -1,4 +1,4 @@
-#!/usr/bin/env sh
+#!/bin/sh
 
 # Sets REPLY to the absolute, well-formed path of a physical directory that
 # contains (or would contain) the supplied path. (Equivalent to the dirname of
diff --git a/realpath_m.sh b/realpath_m.sh
new file mode 100755
index 0000000..5abdb72
--- /dev/null
+++ b/realpath_m.sh
@@ -0,0 +1,80 @@
+#!/bin/sh
+
+set -eu
+
+realpath_canonical() {
+	# resolve symlink (if there is one) and use $@ to detect symlink loops
+	while [ -L "$1" ] && target=$(readlink -- "$1"); do
+		REPLY=$(dirname "$1")
+		# Resolve relative to symlink's directory
+		if [ "$REPLY" = "." ]; then
+			REPLY=$target
+		else
+			case "$target" in
+				/*)
+					REPLY=$target;;
+				*)
+					REPLY=$REPLY/$target;;
+			esac
+		fi
+		# Break out if we found a symlink loop
+		for target in "$@"; do [ "$REPLY" = "$target" ] && break 2; done
+		# Add to the loop-detect list and tail-recurse
+		set -- "$REPLY" "$@"
+	done
+	# recurse unless root
+	REPLY=$(dirname "$1")
+	if [ "$REPLY" != "/" ] && [ "$REPLY" != "." ]; then
+		REPLY=$(realpath_canonical "$REPLY");
+	fi
+	# set $1 to dirname and $2 to basename
+	set -- "$REPLY" "$(basename "$1")"
+	# combine canon parent w/basename
+	REPLY=$PWD
+	while [ $# -gt 0 ]; do
+		# handle case of component starting with double slash and no
+		# other slashes separately because it cannot be solved by
+		# globbing
+		if echo "$1" | grep -E '^//[^/]*$' >/dev/null 2>&1; then
+			REPLY=//
+			first=$(echo "$1" | cut -c3-)
+			shift
+			set -- "$first" "$@"
+			continue
+		fi
+		# all other cases can be expressed by globbing
+		case $1 in
+			/*)
+				REPLY=/
+				first=$(echo "$1" | sed -e 's/^\/\+//')
+				shift
+				set -- "$first" "$@"
+				;;
+			*/*)
+				front=${1%%/*}
+				back=$(echo "$1" | sed -e "s/^$front\\/\\+//")
+				shift
+				set -- "$front" "$back" "$@"
+				;;
+			''|.)
+				shift
+				;;
+			..)
+				REPLY=$(dirname "$REPLY")
+				shift
+				;;
+			*)
+				REPLY="${REPLY%/}/$1"
+				shift
+				;;
+		esac
+	done
+	echo "$REPLY"
+}
+
+if [ $# -ne 1 ]; then
+	echo "usage: $0 path"
+	exit 1
+fi
+
+realpath_canonical "$1"
diff --git a/test.sh b/test.sh
index de6c589..b3b5ca3 100755
--- a/test.sh
+++ b/test.sh
@@ -1,6 +1,6 @@
 #!/bin/sh
 
-set -e
+set -eu
 
 # we use the full path to realpath so that the coreutils version of realpath is
 # used even when run with busybox in $PATH
@@ -8,6 +8,10 @@ realpath_m() {
 	REPLY=$(/usr/bin/realpath -m "$1")
 }
 
+realpath_m2() {
+	REPLY=$("$origpwd/realpath_m.sh" "$1")
+}
+
 assert() {
 	expected=$1
 	shift
@@ -25,6 +29,7 @@ compare() {
 	assert "$("$cmd" "$@")" realpath_"$cmd" "$@"
 }
 
+origpwd=$(pwd)
 numokay=0
 numfail=0
 
@@ -131,14 +136,14 @@ rm x y z; rmdir q
 
 # Current directory:
 
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD" "$cmd" .
 	assert "$PWD" "$cmd" "$PWD"
 done
 
 # Nonexistent file:
 
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD/x" "$cmd" x
 	assert "$PWD/x" "$cmd" "$PWD/x"
 done
@@ -146,7 +151,7 @@ done
 # Symlink to non-existent target:
 
 ln -s y x
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD/y" "$cmd" x
 	assert "$PWD/y" "$cmd" "$PWD/x"
 done
@@ -155,7 +160,7 @@ done
 # Symlink chain:
 
 ln -s z y
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD/z" "$cmd" x
 	assert "$PWD/z" "$cmd" "$PWD/x"
 done
@@ -167,6 +172,7 @@ ln -s y z
 # difference to realpath -m: returned path is the target
 assert "$PWD/z" realpath_canonical x
 assert "$PWD/y" realpath_m x
+assert "$PWD/z" realpath_m2 x
 rm x y z
 
 
@@ -177,19 +183,19 @@ mkdir z
 ln -s $PWD/y x
 ln -s $PWD/x z/x
 
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD/y" "$cmd" x
 	assert "$PWD/y" "$cmd" z/x
 done
 
 ln -sf ../x z/x
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD/y" "$cmd" x
 	assert "$PWD/y" "$cmd" z/x
 done
 
 ln -s z q
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD/z" "$cmd" q
 	assert "$PWD/y" "$cmd" q/../x
 	assert "$PWD/z/a" "$cmd" q/a
@@ -211,7 +217,7 @@ assert "$PWD/q/foo" realpath_location q/foo/bar
 
 ### Missing elements
 
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$PWD/z/foo/bar/baz" "$cmd"  q/foo/bar/baz
 	assert "$PWD/z/foo/bar"     "$cmd"  q/foo/bar
 	assert /non-existent        "$cmd" /non-existent
@@ -224,7 +230,7 @@ done
 
 mkdir foo
 cd foo
-for cmd in realpath_canonical realpath_m; do
+for cmd in realpath_canonical realpath_m realpath_m2; do
 	assert "$OLDPWD/z" "$cmd" "$OLDPWD/q"
 	assert "$OLDPWD/y" "$cmd" "$OLDPWD/q/x"
 	assert "$OLDPWD/y" "$cmd" "../q/x"
@@ -234,15 +240,19 @@ done
 # We can follow symlinks even if $PWD is inaccessible:
 
 rmdir ../foo
-assert "$OLDPWD/z" realpath_canonical "$OLDPWD/q"
-assert "$OLDPWD/y" realpath_canonical "../q/x"
+for cmd in realpath_canonical; do
+	assert "$OLDPWD/z" "$cmd" "$OLDPWD/q"
+	assert "$OLDPWD/y" "$cmd" "../q/x"
+done
 
 # And we can do non-symlink canonicalizing on the base dir of the results:
 
-assert "$OLDPWD/z" realpath_canonical "../z"
-assert "$OLDPWD/z/foo" realpath_canonical "../z/foo"
-assert "$OLDPWD/foo" realpath_canonical "../foo"
-assert "$OLDPWD/z" realpath_canonical "../q"
+for cmd in realpath_canonical; do
+	assert "$OLDPWD/z" "$cmd" "../z"
+	assert "$OLDPWD/z/foo" "$cmd" "../z/foo"
+	assert "$OLDPWD/foo" "$cmd" "../foo"
+	assert "$OLDPWD/z" "$cmd" "../q"
+done
 
 cd ..