Project

General

Profile

Feature #13896

Find.find -> Use Dir.children instead of Dir.entries

Added by esparta (Espartaco Palma) 2 months ago. Updated 2 months ago.

Status:
Closed
Priority:
Normal
Assignee:
-
Target version:
-
[ruby-core:82786]

Description

Dir.children is available since Feature #11302. Find.find can
use of the new list (having no '.' neither '..' entries), making
now superflous an if statement.

This change can improve the performance of Find.find when the path
has lots of entries (thousands?).

Some profiling I did using 50,000 files on a given folder, using this code:

total_size = 0

Find.find(ENV["HOME"]) do |path|
  if FileTest.directory?(path)
   if File.basename(path)[0] == ?.
     Find.prune       # Don't look any further into this directory.
   else
     next
   end
     else
       total_size += FileTest.size(path)
     end
   end
end

Before the patch

 ~/ruby -rprofile before.rb
  %   cumulative   self              self     total
 time   seconds   seconds    calls  ms/call  ms/call  name
 48.37     9.24      9.24   100014     0.09     0.60  Find.find
 13.52    11.82      2.58    50005     0.05     0.07  nil#
  5.17    12.81      0.99    50005     0.02     0.35  Kernel#catch
  5.05    13.77      0.96    50006     0.02     0.04  Kernel#dup
  4.96    14.72      0.95    50006     0.02     0.02  Kernel#initialize_dup
  3.87    15.46      0.74        1   738.89  5952.54  Array#reverse_each
  2.39    15.92      0.46   100012     0.00     0.00  String#==
  1.98    16.29      0.38    50004     0.01     0.01  File.join
  1.93    16.66      0.37    50004     0.01     0.01  FileTest.size
  1.87    17.02      0.36    50005     0.01     0.01  File.lstat
  1.76    17.36      0.34    50005     0.01     0.01  FileTest.directory?
  1.34    17.61      0.26    50004     0.01     0.01  Integer#+
  1.30    17.86      0.25    50005     0.00     0.00  File::Stat#directory?
  1.29    18.11      0.25    50006     0.00     0.00  String#initialize_copy
  1.29    18.35      0.25    50004     0.00     0.00  Array#unshift
  1.28    18.60      0.24    50006     0.00     0.00  Array#shift
  1.25    18.84      0.24    50004     0.00     0.00  Kernel#untaint
  1.25    19.08      0.24    50005     0.00     0.00  Kernel#taint
  0.10    19.09      0.02        1    19.53    19.55  Dir.entries
  0.02    19.10      0.00        1     4.10     4.10  Array#sort!
  0.00    19.10      0.00        2     0.47     1.12  Kernel#require
  0.00    19.10      0.00        4     0.02     0.04  Gem.find_unresolved_default_spec
  0.00    19.10      0.00        2     0.04  9549.52  Array#each
  0.00    19.10      0.00        1     0.05     0.07  MonitorMixin#mon_enter
  0.00    19.10      0.00        1     0.04     0.07  MonitorMixin#mon_exit
  0.00    19.10      0.00        1     0.04     0.05  Module#module_function
  0.00    19.10      0.00        4     0.01     0.01  IO#set_encoding
  0.00    19.10      0.00        1     0.02     0.02  Dir.open
  0.00    19.10      0.00        1     0.02     0.11  Array#collect!
  0.00    19.10      0.00        1     0.02     0.02  MonitorMixin#mon_check_owner
  0.00    19.10      0.00        2     0.01     0.01  Module#method_added
  0.00    19.10      0.00        3     0.00     0.00  Thread.current
  0.00    19.10      0.00        2     0.01     0.01  String#encoding
  0.00    19.10      0.00        2     0.01     0.01  BasicObject#singleton_method_added
  0.00    19.10      0.00        2     0.00     0.00  Kernel#respond_to?
  0.00    19.10      0.00        1     0.01     0.01  TracePoint#enable
  0.00    19.10      0.00        1     0.01     0.01  Gem::Specification.unresolved_deps
  0.00    19.10      0.00        1     0.01     0.01  File.exist?
  0.00    19.10      0.00        1     0.01     0.01  File.basename
  0.00    19.10      0.00        1     0.01     0.01  Thread::Mutex#lock
  0.00    19.10      0.00        1     0.01     0.01  String#[]
  0.00    19.10      0.00        1     0.01     0.01  Gem.suffixes
  0.00    19.10      0.00        1     0.01     0.01  Thread::Mutex#unlock
  0.00    19.10      0.00        1     0.01     0.01  Encoding.find
  0.00    19.10      0.00        1     0.01     0.01  BasicObject#==
  0.00    19.10      0.00        1     0.00     0.00  TracePoint#disable
  0.00    19.10      0.00        1     0.00     0.00  Kernel#block_given?
  0.00    19.10      0.00        1     0.00 19100.95  #toplevel

After the patch

  %   cumulative   self              self     total
 time   seconds   seconds    calls  ms/call  ms/call  name
 45.15     7.70      7.70   100012     0.08     0.52  Find.find
 15.12    10.27      2.58    50005     0.05     0.07  nil#
  5.76    11.25      0.98    50005     0.02     0.31  Kernel#catch
  5.66    12.22      0.96    50006     0.02     0.04  Kernel#dup
  5.59    13.17      0.95    50006     0.02     0.02  Kernel#initialize_dup
  4.30    13.91      0.73        1   733.77  3952.79  Array#reverse_each
  2.12    14.27      0.36    50005     0.01     0.01  File.lstat
  2.08    14.62      0.35    50004     0.01     0.01  FileTest.size
  2.06    14.97      0.35    50004     0.01     0.01  File.join
  1.93    15.30      0.33    50005     0.01     0.01  FileTest.directory?
  1.47    15.55      0.25    50005     0.00     0.00  File::Stat#directory?
  1.46    15.80      0.25    50006     0.00     0.00  String#initialize_copy
  1.46    16.05      0.25    50004     0.00     0.00  Integer#+
  1.44    16.30      0.25    50004     0.00     0.00  Array#unshift
  1.44    16.54      0.24    50006     0.00     0.00  Array#shift
  1.41    16.78      0.24    50004     0.00     0.00  Kernel#untaint
  1.40    17.02      0.24    50005     0.00     0.00  Kernel#taint
  0.11    17.04      0.02        1    19.16    19.18  Dir.children
  0.02    17.04      0.00        1     4.14     4.14  Array#sort!
  0.00    17.04      0.00        1     0.61     0.69  Kernel#require_relative
  0.00    17.04      0.00        1     0.04     0.05  Module#module_function
  0.00    17.04      0.00        4     0.01     0.01  IO#set_encoding
  0.00    17.04      0.00        1     0.02 17043.77  Array#each
  0.00    17.04      0.00        1     0.02     0.11  Array#collect!
  0.00    17.04      0.00        1     0.02     0.02  Dir.open
  0.00    17.04      0.00        2     0.01     0.01  Module#method_added
  0.00    17.04      0.00        1     0.01     0.01  TracePoint#enable
  0.00    17.04      0.00        2     0.00     0.00  String#encoding
  0.00    17.04      0.00        2     0.00     0.00  BasicObject#singleton_method_added
  0.00    17.04      0.00        1     0.01     0.01  File.basename
  0.00    17.04      0.00        1     0.01     0.01  File.exist?
  0.00    17.04      0.00        1     0.01     0.01  String#[]
  0.00    17.04      0.00        1     0.01     0.01  Encoding.find
  0.00    17.04      0.00        1     0.01     0.01  String#==
  0.00    17.04      0.00        1     0.01     0.01  BasicObject#==
  0.00    17.04      0.00        1     0.00     0.00  TracePoint#disable
  0.00    17.04      0.00        1     0.00     0.00  Kernel#block_given?
  0.00    17.04      0.00        1     0.00     0.00  Kernel#respond_to?
  0.00    17.05      0.00        1     0.00 17045.11  #toplevel

https://github.com/ruby/ruby/pull/1697

Associated revisions

Revision 59926
Added by naruse (Yui NARUSE) 2 months ago

Find.find -> Use Dir.children instead of Dir.entries

Dir.children is available since Feature #11302.
Find.find can use of the new list (having no '.' neither '..' entries),
making now superflous an if statement.

This change can improve the performance of Find.find when the path
has lots of entries (thousands?).

https://bugs.ruby-lang.org/issues/11302
patched by Espartaco Palma esparta@gmail.com
https://github.com/ruby/ruby/pull/1697 fix GH-1697
[Feature #13896]

History

#1 Updated by esparta (Espartaco Palma) 2 months ago

  • Backport deleted (2.2: UNKNOWN, 2.3: UNKNOWN, 2.4: UNKNOWN)
  • Tracker changed from Bug to Misc

#2 Updated by naruse (Yui NARUSE) 2 months ago

  • Tracker changed from Misc to Feature

#3 Updated by naruse (Yui NARUSE) 2 months ago

  • Status changed from Open to Closed

Applied in changeset trunk|r59926.


Find.find -> Use Dir.children instead of Dir.entries

Dir.children is available since Feature #11302.
Find.find can use of the new list (having no '.' neither '..' entries),
making now superflous an if statement.

This change can improve the performance of Find.find when the path
has lots of entries (thousands?).

https://bugs.ruby-lang.org/issues/11302
patched by Espartaco Palma esparta@gmail.com
https://github.com/ruby/ruby/pull/1697 fix GH-1697
[Feature #13896]

Also available in: Atom PDF