Re: Extend readlink -e

[Eric Blake]

please keep the list in the loop

On 11/04/2011 10:42 AM, Peng Yu wrote:
> > I'm not sure I follow what you are asking for, then.

> I understand -f does what I need. I was saying that this fact is not clear
> just by reading the manual. I'm suggesting to reword the manual to
> something like this.
>
>         -f, --canonicalize
>                canonicalize by following every symlink in every component of
> the given name recursively; all but the last component must exist. If the
> argument is a real file or path, then the abspath of the argument is
> returned.

The current wording in the manual (that is, the info pages) is:

`Canonicalize mode'
     `readlink' outputs the absolute name of the given file which
     contains no `.', `..' components nor any repeated separators (`/')
     or symbolic links.

`-f'
`--canonicalize'
     Activate canonicalize mode.  If any component of the file name
     except the last one is missing or unavailable, `readlink' produces
     no output and exits with a nonzero exit code.  A trailing slash is
     ignored.

The information in 'readlink --help' (which also appears in 'man 
readlink') is intentionally shorter, but still gets across the 
difference in the three canonicalization modes:

  -f, --canonicalize            canonicalize by following every symlink in
                                every component of the given name 
recursively;
                                all but the last component must exist
  -e, --canonicalize-existing   canonicalize by following every symlink in
                                every component of the given name 
recursively,
                                all components must exist
  -m, --canonicalize-missing    canonicalize by following every symlink in
                                every component of the given name 
recursively,
                                without requirements on components 
existence


> I hope that it is now clear to you what I mean.

I'm not quite sure I see what you are asking to be changed.  It appears 
that you are complaining that the --help output doesn't mention anything 
about the canonicalization always producing an absolute name, unless 
part of the resolution process fails, and that the difference between 
the three canonicalization modes is how much of the user input name must 
already exist before considering resolution as failed.

At this point, patches might speak louder than words.  You are welcome 
to submit a patch against readlink.c that improves the --help wording, 
or against coreutils.texi to improve the info wording; remember that it 
is intentionally a goal of the --help output to be concise, while the 
info documentation should be complete.

--
Eric Blake   address@hidden    +1-801-349-2682
Libvirt virtualization library http://libvirt.org