Re: find -f -, please?

[James Youngman]

On Mon, Feb 15, 2016 at 9:44 PM, Marc Chantreux <address@hidden> wrote:
> I finally wrote
>
>     a-very-long-list-of-files | perl -lnE 'print if -f && /[.]scm$/'

The solution I'd suggest for this constrained problem is:

 sed -e 's_^\([^/]\)_./\1_' < mylist  |
    xargs -d '\n' sh -c 'find "$@" -regex ".*\.scm$" -type f' sh | sed
-e 's_./__'

I use -d in preference to -L because -L leaves quote-processing
enabled. If -d is not available, you could use -L but you will need to
worry about quote and EOF processing (since the xargs you're using may
default to having an EOF string, which is annoying but
POSIX-compliant).

I use -regex in preference to -name because there is no precomputation
for fnmatch patterns in GNU find (yet...) so -regex is slightly
faster.

If we can relax the constraints (in the way the problem is set up) a
bit, I'd suggest using a NUL-terminated list of file names, since file
names can contain newlines.   Sometimes it will also be a win to sort
the input (to better take advantage of directory entry caches).

Quite honestly though, I'd say that using Perl is a perfectly fine way
to solve this problem.

> this isn't a problem on a BSD system as find has the wonderfull -f.
> According to http://netbsd.gw.com/cgi-bin/man-cgi?find+1+NetBSD-current
>
>     -f  Specifies a file hierarchy for find to traverse.
>         File hierarchies may also be specified as the
>         operands immediately following the options.
>
> It would be nice to get this feature into the gnu tools. I hope you like
> the idea.

AFAICS though, this strategy only works well if you are willing to
invoke a separate instance of find for item in that very long list.
That seems inefficient.   This strategy also doesn't work for path
names containing newlines.

Thanks for the suggestions.
James.