Handle ignored files with quoted (non-ASCII) filenames

jcouball · jcouball · commit 6ce3d4df8847 · 2024-06-01T00:04:12.000-07:00
diff --git a/lib/git/base.rb b/lib/git/base.rb
@@ -309,6 +309,13 @@ def grep(string, path_limiter = nil, opts = {})
       self.object('HEAD').grep(string, path_limiter, opts)
     end
 
+    # List the files in the worktree that are ignored by git
+    # @return [Array<String>] the list of ignored files relative to teh root of the worktree
+    #
+    def ignored_files
+      self.lib.ignored_files
+    end
+
     # removes file(s) from the git repository
     def rm(path = '.', opts = {})
       self.lib.rm(path, opts)
diff --git a/lib/git/escaped_path.rb b/lib/git/escaped_path.rb
@@ -3,7 +3,7 @@
 module Git
   # Represents an escaped Git path string
   #
-  # Git commands that output paths (e.g. ls-files, diff), will escape usual
+  # Git commands that output paths (e.g. ls-files, diff), will escape unusual
   # characters in the path with backslashes in the same way C escapes control
   # characters (e.g. \t for TAB, \n for LF, \\ for backslash) or bytes with values
   # larger than 0x80 (e.g. octal \302\265 for "micro" in UTF-8).
diff --git a/lib/git/lib.rb b/lib/git/lib.rb
@@ -574,18 +574,52 @@ def diff_index(treeish)
       diff_as_hash('diff-index', treeish)
     end
 
+    # List all files that are in the index
+    #
+    # @param location [String] the location to list the files from
+    #
+    # @return [Hash<String, Hash>] a hash of files in the index
+    #   * key: file [String] the file path
+    #   * value: file_info [Hash] the file information containing the following keys:
+    #     * :path [String] the file path
+    #     * :mode_index [String] the file mode
+    #     * :sha_index [String] the file sha
+    #     * :stage [String] the file stage
+    #
     def ls_files(location=nil)
       location ||= '.'
-      hsh = {}
-      command_lines('ls-files', '--stage', location).each do |line|
-        (info, file) = line.split("\t")
-        (mode, sha, stage) = info.split
-        if file.start_with?('"') && file.end_with?('"')
-          file = Git::EscapedPath.new(file[1..-2]).unescape
+      {}.tap do |files|
+        command_lines('ls-files', '--stage', location).each do |line|
+          (info, file) = line.split("\t")
+          (mode, sha, stage) = info.split
+          files[unescape_quoted_path(file)] = {
+            :path => file, :mode_index => mode, :sha_index => sha, :stage => stage
+          }
         end
-        hsh[file] = {:path => file, :mode_index => mode, :sha_index => sha, :stage => stage}
       end
-      hsh
+    end
+
+    # Unescape a path if it is quoted
+    #
+    # Git commands that output paths (e.g. ls-files, diff), will escape unusual
+    # characters.
+    #
+    # @example
+    #   lib.unescape_if_quoted('"quoted_file_\\342\\230\\240"') # => 'quoted_file_☠'
+    #   lib.unescape_if_quoted('unquoted_file')   # => 'unquoted_file'
+    #
+    # @param path [String] the path to unescape if quoted
+    #
+    # @return [String] the unescaped path if quoted otherwise the original path
+    #
+    # @api private
+    #
+    def unescape_quoted_path(path)
+      if path.start_with?('"') && path.end_with?('"')
+        Git::EscapedPath.new(path[1..-2]).unescape
+      else
+        path
+      end
     end
 
     def ls_remote(location=nil, opts={})
@@ -606,7 +640,7 @@ def ls_remote(location=nil, opts={})
     end
 
     def ignored_files
-      command_lines('ls-files', '--others', '-i', '--exclude-standard')
+      command_lines('ls-files', '--others', '-i', '--exclude-standard').map { |f| unescape_quoted_path(f) }
     end
 
     def untracked_files
diff --git a/tests/units/test_ignored_files_with_escaped_path.rb b/tests/units/test_ignored_files_with_escaped_path.rb
@@ -0,0 +1,23 @@
+#!/usr/bin/env ruby
+# encoding: utf-8
+
+require 'test_helper'
+
+# Test diff when the file path has to be quoted according to core.quotePath
+# See https://git-scm.com/docs/git-config#Documentation/git-config.txt-corequotePath
+#
+class TestIgnoredFilesWithEscapedPath < Test::Unit::TestCase
+  def test_ignored_files_with_non_ascii_filename
+    in_temp_dir do |path|
+      create_file('README.md', '# My Project')
+      `git init`
+      `git add .`
+      `git config --local core.safecrlf false` if Gem.win_platform?
+      `git commit -m "First Commit"`
+      create_file('my_other_file_☠', "First Line\n")
+      create_file(".gitignore", "my_other_file_☠")
+      files = Git.open('.').ignored_files
+      assert_equal(['my_other_file_☠'].sort, files)
+    end
+  end
+end

Original file line number	Diff line number	Diff line change
`@@ -3,7 +3,7 @@`
`3`	`3`	`module Git`
`4`	`4`	`# Represents an escaped Git path string`
`5`	`5`	`#`
`6`		`- # Git commands that output paths (e.g. ls-files, diff), will escape usual`
	`6`	`+ # Git commands that output paths (e.g. ls-files, diff), will escape unusual`
`7`	`7`	`# characters in the path with backslashes in the same way C escapes control`
`8`	`8`	`# characters (e.g. \t for TAB, \n for LF, \\ for backslash) or bytes with values`
`9`	`9`	`# larger than 0x80 (e.g. octal \302\265 for "micro" in UTF-8).`