Re: dired-make-relative-symlink

[Pierre Lorenzon]

From: Pierre Lorenzon <devel@pollock-nageoire.net>
Subject: Re: dired-make-relative-symlink
Date: Tue, 04 Jul 2017 17:21:33 +0200 (CEST)

> 
> Hi,
> 
> Here is a very compact and short implementation
> 
> (defun dired-make-relative-symlink (file1 file2 &optional 
> ok-if-already-exists)
>   "Make a symbolic link (pointing to FILE1) in FILE2.
> The link is relative (if possible), for example
> 
>     \"/vol/tex/bin/foo\" \"/vol/local/bin/foo\"
> 
> results in
> 
>     \"../../tex/bin/foo\" \"/vol/local/bin/foo\""
>   (interactive "FRelSymLink: \nFRelSymLink %s: \np")
>   (let ((file1 (expand-file-name file1))
>       (file2 (expand-file-name file2)))
>     (make-symbolic-link (directory-file-name
>                        (file-relative-name file1
>                                            (file-name-directory
>                                             file2)))
>                       file2 ok-if-already-exists)))

                        Notice that dired-make-relative could
                        probably be used here as well. Anyway
                        dired-make-relative could certainly be
                        implmented by file-relative-name. The
                        code is not duplicated but at least
                        written three times !

                        I think it is the case in most of codes
                        and even when someone discovers it
                        maintainers do not have time to clean
                        such situations and verify that
                        modification do no cause bugs.


                        Regard

                        Pierre


> 
> I did not yet notice bugs but maybe there are some !
> 
> If you think it is accurate to replace the code in dired-x.el
> fell free to do it !
> 
> Regards
> 
> Pierre
> 
> 
> From: Pierre Lorenzon <devel@pollock-nageoire.net>
> Subject: dired-make-relative-symlink
> Date: Tue, 04 Jul 2017 16:10:38 +0200 (CEST)
> 
>> 
>> 
>> Hi,
>> 
>> I wonder why the function dired-make-relative-symlink
>> recalculate the relative name of the file instead of using
>> file-relative-name of the files.el library.
>> 
>> Making a few tests it seems that both codes produce the same
>> result. Anyway maybe these tests are not torturing the code
>> enough.
>> 
>> files.el library seems to be very old and it looks strange to
>> me that when dired-x.el was developped files.el was not known.
>> 
>> I did not report that as a bug since it is not a bug but I
>> think that a system is more maintainable when code is not
>> duplicated.
>> 
>> Regards
>> 
>> PS/ Sory if I did not notice an evident reason for which the
>> code is so !
>> 
>