Project

General

Profile

Actions

Feature #7596

closed

Find::find should not silently ignores errors

Added by hoylen (Hoylen Sue) about 12 years ago. Updated almost 11 years ago.

Status:
Closed
Target version:
[ruby-core:51025]

Description

=begin
The current implementation of (({Find::find})) silently ignores errors. It deliberately catches a number of (({Errno::*})) errors and just continues processing. This can cause unexpected (and often unnoticed) results when, for example, unreadable directories are encountered. Find will not recurse into those directories, but does also not tell the user that it is skipping them and their contents. This happened to me when there was a directory owned by another user.

I suggest making the default behaviour to ((not)) ignore errors. But then provide an ((option)) for the caller to indicate that they want (({find})) to keep going if it encounters some types of errors. This way the caller has control and the default behaviour is the one with "least surprise" for the caller.

Either that, or at least change the documentation to point out that the current implementation silently ignores errors and unreadable directories. That way the caller will know the limitations of the method.

http://www.ruby-doc.org/stdlib-1.9.3/libdoc/find/rdoc/Find.html

When updating the documentation, it would also be useful for documentation for the "Find" module's "find" class method to also point out that:

"The associated block is never called with "." or "..", except when they are explicitly provided as one of the arguments."

You can check this behavour by trying out different values for "testdirname" in the following command

ruby -e 'require "find"; Find.find(ARGV[0]) { |d| puts "<#{d}>" }' testdirname

=end

Actions #1

Updated by naruse (Yui NARUSE) almost 12 years ago

  • Tracker changed from Feature to Bug
  • Project changed from 14 to Ruby master
Actions #2

Updated by naruse (Yui NARUSE) almost 12 years ago

  • Tracker changed from Bug to Feature

Updated by drbrain (Eric Hodel) almost 12 years ago

  • Category set to lib
  • Target version set to 2.6

Updated by ko1 (Koichi Sasada) almost 12 years ago

  • Assignee set to matz (Yukihiro Matsumoto)

I'm not sure who can grab this ticket.
So I assign it to matz.

Updated by matz (Yukihiro Matsumoto) almost 11 years ago

I agree with adding keyword argument, e.g. ignore_error: true.
Choose a good keyword for the option.

Matz.

Updated by nobu (Nobuyoshi Nakada) almost 11 years ago

What about an optional error handler?
Simple flag may not be enough, for example, to search unreadable directories.

Updated by matz (Yukihiro Matsumoto) almost 11 years ago

Nobu, let's not add new things. This time let us focus on stopping ignoring errors.
Probably we can add another keyword argument to specify exception handler.

Matz.

Updated by nobu (Nobuyoshi Nakada) almost 11 years ago

ignore_error: true?

diff --git i/lib/find.rb w/lib/find.rb
index 6f3e428..c5fd35b 100644
--- i/lib/find.rb
+++ w/lib/find.rb
@@ -34,7 +34,7 @@ module Find
   #
   # See the +Find+ module documentation for an example.
   #
-  def find(*paths) # :yield: path
+  def find(*paths, ignore_error: true) # :yield: path
     block_given? or return enum_for(__method__, *paths)
 
     fs_encoding = Encoding.find("filesystem")
@@ -48,12 +48,14 @@ module Find
           begin
             s = File.lstat(file)
           rescue Errno::ENOENT, Errno::EACCES, Errno::ENOTDIR, Errno::ELOOP, Errno::ENAMETOOLONG
+            raise unless ignore_error
             next
           end
           if s.directory? then
             begin
               fs = Dir.entries(file, encoding: enc)
             rescue Errno::ENOENT, Errno::EACCES, Errno::ENOTDIR, Errno::ELOOP, Errno::ENAMETOOLONG
+              raise unless ignore_error
               next
             end
             fs.sort!
diff --git i/test/test_find.rb w/test/test_find.rb
index b924651..8bd6782 100644
--- i/test/test_find.rb
+++ w/test/test_find.rb
@@ -100,6 +100,16 @@ class TestFind < Test::Unit::TestCase
         a = []
         Find.find(d) {|f| a << f }
         assert_equal([d, dir], a)
+
+        a = []
+        Find.find(d, ignore_error: true) {|f| a << f }
+        assert_equal([d, dir], a)
+
+        a = []
+        assert_raise_with_message(Errno::EACCES, /#{Regexp.quote(dir)}/) do
+          Find.find(d, ignore_error: false) {|f| a << f }
+        end
+        assert_equal([d, dir], a)
       ensure
         File.chmod(0700, dir)
       end
@@ -115,6 +125,17 @@ class TestFind < Test::Unit::TestCase
         a = []
         Find.find(d) {|f| a << f }
         assert_equal([d, dir, file], a)
+
+        a = []
+        Find.find(d, ignore_error: true) {|f| a << f }
+        assert_equal([d, dir, file], a)
+
+        a = []
+        assert_raise_with_message(Errno::EACCES, /#{Regexp.quote(file)}/) do
+          Find.find(d, ignore_error: false) {|f| a << f }
+        end
+        assert_equal([d, dir, file], a)
+
         skip "no meaning test on Windows" if /mswin|mingw/ =~ RUBY_PLATFORM
         assert_raise(Errno::EACCES) { File.lstat(file) }
       ensure

Updated by nobu (Nobuyoshi Nakada) almost 11 years ago

  • Status changed from Open to Closed
  • % Done changed from 0 to 100

Applied in changeset r45241.


find.rb: add ignore_error

  • lib/find.rb (Find#find): add "ignore_error" keyword argument
    defaulted to true. [ruby-core:51025] [Feature #7596]
Actions

Also available in: Atom PDF

Like0
Like0Like0Like0Like0Like0Like0Like0Like0Like0