[bug #56106] make it easier to exclude a directory

[Ernesto Alfonso]

URL:
  <https://savannah.gnu.org/bugs/?56106>

                 Summary: make it easier to exclude a directory
                 Project: findutils
            Submitted by: erjoalgo
            Submitted on: Tue 09 Apr 2019 05:13:40 AM UTC
                Category: None
                Severity: 3 - Normal
              Item Group: None
                  Status: None
                 Privacy: Public
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: None
           Fixed Release: None

    _______________________________________________________

Details:

The current ways of excluding a directory in find are confusing or have
shortcomings that make them hard to use, and there seems to be no simple,
intuitive, easy to remember way to do this.

Several ways of doing this have been proposed in this stackoverflow
question[1]:

The first one suggests using "-path ./DIRECTORY -prune". When I used tried to
use this approach to exclude a .git directory and its contents:

1. I first got an unexpected result:

    $ find . -path .git -prune
    # nothing matches

2. Then I tried adding a leading "./" to the -path, which gave me a single
result, .git, the same directory I am trying to exclude:

    $ find . -path ./.git -prune
    ./.git


3. Then I tried adding a "-o -type f" condition:

    $ find . -path ./.git -prune -o -type f

This includes only regular files, so it's not the same as simply "excluding a
directory", since I have to explicitly say what I want to include. It also
seems to still include the './.git' directory I wanted to exclude

I looked into the next answer, which relies on passing a shell-escaped glob
star:

    $ find -not -path "./.git/*"

This seems better than #3, since it is closer to "find everything except ..."
without having to specify what to include. The problem with this is that it
hard to build on top of find in shell scripts using this glob approach without
running into many types of shell escaping issues. This also still includes
"./.git" itself.

The next answer suggests that it is easier to reason about this form:

5. find build -not \( -path build/external -prune \) -name \*.js

I find this double negation confusing and don't see anything intuitive about
it, and it involves a complicated explanation about how -prune and -not work
together.

Could we have a simpler, script-friendly way of excluding a directory tree
that is more intuitive to remember and understand?

Perhaps something like:

    $ find . -not -prefix .git
    # maybe this should also work
    $ find . -not -prefix ./.git

The expected behavior would be to exclude everything matching a given prefix
or directory, including the directory itself, from the set of files or
directories that would normally be returned without the added predicate.

Or is there already an intuitive way of doing this that is free from some of
the shortcomings I mentioned above?

[1]
https://stackoverflow.com/questions/4210042/how-to-exclude-a-directory-in-find-command





    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?56106>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/