From 3751b47c504fe80fff5acd51a58904b5975eff4c Mon Sep 17 00:00:00 2001
From: Jordon Bedwell <jordon@envygeeks.io>
Date: Sat, 30 Apr 2016 01:16:43 -0500
Subject: [PATCH] Cleanup EntryFilter and make it far more robust.

* Allow users to filter directories by ending their path with "/"
* Allow users to filter with a Regexp, some scenariors can really require it.
* Use Pathutil#in_path? for Symlink verification, it real/expand.

This also requires some downstream work in "jekyll-watch" which at this time is
not very robust, it doesn't recognize the difference either, and should probably
start doing so (what I mean is detecting "/" and using the full path.)
---
 jekyll.gemspec             |  1 +
 lib/jekyll.rb              |  1 +
 lib/jekyll/entry_filter.rb | 76 +++++++++++++++++++++++++++++---------
 test/test_entry_filter.rb  | 11 ++++++
 4 files changed, 72 insertions(+), 17 deletions(-)

diff --git a/jekyll.gemspec b/jekyll.gemspec
index 62c94414..3f32f0a9 100644
--- a/jekyll.gemspec
+++ b/jekyll.gemspec
@@ -36,4 +36,5 @@ Gem::Specification.new do |s|
   s.add_runtime_dependency('rouge', '~> 1.7')
   s.add_runtime_dependency('jekyll-sass-converter', '~> 1.0')
   s.add_runtime_dependency('jekyll-watch', '~> 1.1')
+  s.add_runtime_dependency("pathutil", "~> 0.9")
 end
diff --git a/lib/jekyll.rb b/lib/jekyll.rb
index 8afdf06a..ffc1539f 100644
--- a/lib/jekyll.rb
+++ b/lib/jekyll.rb
@@ -16,6 +16,7 @@ end
 require 'rubygems'
 
 # stdlib
+require "pathutil"
 require 'forwardable'
 require 'fileutils'
 require 'time'
diff --git a/lib/jekyll/entry_filter.rb b/lib/jekyll/entry_filter.rb
index cccd05f9..496bf56f 100644
--- a/lib/jekyll/entry_filter.rb
+++ b/lib/jekyll/entry_filter.rb
@@ -1,12 +1,15 @@
 module Jekyll
   class EntryFilter
-    SPECIAL_LEADING_CHARACTERS = ['.', '_', '#', '~'].freeze
-
     attr_reader :site
+    SPECIAL_LEADING_CHARACTERS = [
+      '.', '_', '#', '~'
+    ].freeze
 
     def initialize(site, base_directory = nil)
       @site = site
-      @base_directory = derive_base_directory(@site, base_directory.to_s.dup)
+      @base_directory = derive_base_directory(
+        @site, base_directory.to_s.dup
+      )
     end
 
     def base_directory
@@ -14,14 +17,14 @@ module Jekyll
     end
 
     def derive_base_directory(site, base_dir)
-      if base_dir.start_with?(site.source)
-        base_dir[site.source] = ""
-      end
+      base_dir[site.source] = "" if base_dir.start_with?(site.source)
       base_dir
     end
 
     def relative_to_source(entry)
-      File.join(base_directory, entry)
+      File.join(
+        base_directory, entry
+      )
     end
 
     def filter(entries)
@@ -33,7 +36,9 @@ module Jekyll
     end
 
     def included?(entry)
-      glob_include?(site.include, entry)
+      glob_include?(site.include,
+        entry
+      )
     end
 
     def special?(entry)
@@ -51,25 +56,62 @@ module Jekyll
       excluded
     end
 
+    # --
+    # Check if a file is a symlink.
+    # NOTE: This can be converted to allowing even in safe,
+    #   since we use Pathutil#in_path? now.
+    # --
     def symlink?(entry)
       site.safe && File.symlink?(entry) && symlink_outside_site_source?(entry)
     end
 
+    # --
+    # NOTE: Pathutil#in_path? gets the realpath.
+    # @param [<Anything>] entry the entry you want to validate.
+    # Check if a path is outside of our given root.
+    # --
     def symlink_outside_site_source?(entry)
-      ! File.realpath(entry).start_with?(File.realpath(@site.source))
-    end
-
-    def ensure_leading_slash(path)
-      path[0..0] == "/" ? path : "/#{path}"
+      !Pathutil.new(entry).in_path?(
+        site.in_source_dir
+      )
     end
 
+    # --
+    # Check if an entry matches a specific pattern and return true,false.
     # Returns true if path matches against any glob pattern.
-    # Look for more detail about glob pattern in method File::fnmatch.
+    # --
     def glob_include?(enum, e)
-      entry = ensure_leading_slash(e)
+      entry = Pathutil.new(site.in_source_dir).join(e)
       enum.any? do |exp|
-        item = ensure_leading_slash(exp)
-        File.fnmatch?(item, entry) || entry.start_with?(item)
+
+        # Users who send a Regexp knows what they want to
+        # exclude, so let them send a Regexp to exclude files,
+        # we will not bother caring if it works or not, it's
+        # on them at this point.
+
+        if exp.is_a?(Regexp)
+          entry =~ exp
+
+        else
+          item = Pathutil.new(site.in_source_dir).join(exp)
+
+          # If it's a directory they want to exclude, AKA
+          # ends with a "/" then we will go on to check and
+          # see if the entry falls within that path and
+          # exclude it if that's the case.
+
+          if e.end_with?("/")
+            entry.in_path?(
+              item
+            )
+
+          else
+            File.fnmatch?(item, entry) ||
+              entry.to_path.start_with?(
+                item
+              )
+          end
+        end
       end
     end
   end
diff --git a/test/test_entry_filter.rb b/test/test_entry_filter.rb
index d20c9be2..d9e38300 100644
--- a/test/test_entry_filter.rb
+++ b/test/test_entry_filter.rb
@@ -14,6 +14,17 @@ class TestEntryFilter < JekyllUnitTest
       assert_equal %w[foo.markdown bar.markdown baz.markdown .htaccess], entries
     end
 
+    should "allow regexp filtering" do
+      files = %w(README.md)
+      @site.exclude = excludes = [
+        /README/
+      ]
+
+      assert_empty @site.reader.filter_entries(
+        files
+      )
+    end
+
     should "filter entries with exclude" do
       excludes = %w[README TODO vendor/bundle]
       files = %w[index.html site.css .htaccess vendor]