On Fri, Sep 02, 2005 at 03:29:22PM -0400, Drew Van Zandt wrote:
> I would suggest adding a short sentence in the -type explanation
> pointing the user to -xtype if -follow is used, e.g.: If using
> -follow, see also -xtype.
The Texinfo documentation currently says this:
The following differences in behavior occur when the @samp{-L} option
is used:
@itemize @bullet
@item
@code{find} follows symbolic links to directories when searching
directory trees.
@item
@samp{-lname} and @samp{-ilname} always return false (unless they
happen to match broken symbolic links).
@item
@samp{-type} reports the types of the files that symbolic links point
to.
@item
Implies @samp{-noleaf} (@pxref{Directories}).
@end itemize
I have enhanced the final item like so:
@item
@samp{-type} reports the types of the files that symbolic links point
to. This means that in combination with @samp{-L}, @samp{-type l}
will be true only for broken symbolic links. To check for symbolic
links when @samp{-L} has been specified, use @samp{-xtype}.
I have also modified the description of -type and -xtype:
@deffn Test -type c
True if the file is of type @var{c}:
@table @code
@item b
block (buffered) special
@item c
character (unbuffered) special
@item d
directory
@item p
named pipe (FIFO)
@item f
regular file
@item l
symbolic link; if @samp{-L} is in effect, this is true only for broken
symbolic links. If you want to search for symbolic links when
@samp{-L} is in effect, use @samp{-xtype} instead of @samp{-type}.
@item s
socket
@item D
door (Solaris)
@end table
@end deffn
@deffn Test -xtype c
This test behaves the same as @samp{-type} unless the file is a
symbolic link. If the file is a symbolic link, the result is as
follows (in the table below, @samp{X} should be understood to
represent any letter except @samp{l}):
@table @samp
@item @samp{-P -xtype l}
True if the symbolic link is broken
@item @samp{-P -xtype X}
True if the (ultimate) target file is of type @samp{X}.
@item @samp{-L -xtype l}
Always true
@item @samp{-L -xtype X}
False unless the symbolic link is broken
@end table
In other words, for symbolic links, @samp{-xtype} checks the type of
the file that @samp{-type} does not check.
The @samp{-H} option also affects the behaviour of @samp{-xtype}.
When @samp{-H} is in effect, @samp{-xtype} behaves as if @samp{-L} had
been specified when examining files listed on the command line, and as
if @samp{-P} had been specified otherwise. If neither @samp{-H} nor
@samp{-L} was specified, @samp{-xtype} behaves as if @samp{-P} had
been specified.
@xref{Symbolic Links}, for more information on @samp{-follow} and
@samp{-L}.
@end deffn
> It was not apparent until you pointed it out to me that -type l
> wouldn't count a dereferenced link as a link; would find -follow -type
> l EVER return anything? If it doesn't ever return anything, it
> seems... awkward to have it at all, except for regularity in the
> command-line options. Should it "complain" that -follow and -type l
> are incompatible? (Note that I'm not saying it should, just raising
> the question.)
I hope that the above documentation improvements make the situation a
bit clearer. I have also made a small change to the manpage:-
.IP "\-type \fIc\fR"
File is of type \fIc\fR:
.RS
.IP b
block (buffered) special
.IP c
character (unbuffered) special
.IP d
directory
.IP p
named pipe (FIFO)
.IP f
regular file
.IP l
symbolic link; this is never true if the \-L option or the \-follow
option is in effect, unless the symbolic link is broken. If you want
to search for symbolic links when \-L is in effect, use \-xtype.
If you have any other suggested changes, please let me know.
Regards,
James.
_______________________________________________
Bug-findutils mailing list
Bug-findutils@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-findutils
