TL;DR: why [ symlink_to_file_a -ef file_a ] return true ?
I need to test if a file b (mostly symlink) is the same than a file a for trying to know if a and b are hard linked together, and hard link them if itsn't the case.
I use conditional expression -ef and bash manual says :
file1 -ef file2
True if file1 and file2 refer to the same device and inode numbers.
I compare just symlink/regular file and inodes of a and b are different but the result is True.
An answer of a similar question says :
If yes, then will the inode number be same for target and links?
No. Usually, the symlink is a file with its own inode, (with file-type, own data blocks, etc.)
I'm not sure of what I understood but I may found some explanations about it in ext4 spec :
The target of a symbolic link will be stored in this field if the target string is less than 60 bytes long. Otherwise, either extents or block maps will be used to allocate data blocks to store the link target.
I tried with target shorter/longer than 60B but there is no difference.
$ cat test.sh
#!/usr/bin/env bash
foo="foooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo"
mkdir /tmp/test
cd /tmp/test
touch a
ln -s /tmp/test/a b
ls -li
if [ a -ef b ]
then
echo b : same device or inode of a
else
echo b : different device or inode of a
fi
mkdir /tmp/test/${foo}
cd /tmp/test/${foo}
touch c
ln -s /tmp/test/${foo}/c d
ls -li
if [ c -ef d ]
then
echo d : same device or inode of c
else
echo d : different device or inode of c
fi
$ ./test.sh
156490 -rw-rw-r-- 1 msi msi 0 nov. 25 23:55 a
156491 lrwxrwxrwx 1 msi msi 11 nov. 25 23:55 b -> /tmp/test/a
b : same device or inode of a
total 4
156494 -rw-rw-r-- 1 msi msi 0 nov. 25 23:55 c
156495 lrwxrwxrwx 1 msi msi 155 nov. 25 23:55 d -> /tmp/test/foooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo/c
d : same device or inode of c
Inodes are different but the test succeed and I didn't see what is wrong.
