Merge pull request #1588 from github/cache-bustin

Bust that cache

README.md

@@ -34,7 +34,7 @@ The repository stats API, accessed through `#languages`, can be used on a direct
 
 ***API UPDATE***
 
-Since [Version 3.0.0](https://github.com/github/linguist/releases/tag/v3.0.0) Linguist requires a git repository (in the form of a [Rugged::Repository](https://github.com/libgit2/rugged#repositories)) to be passed when initializing `Linguist::Repository`.
+Since [Version 3.0.0](https://github.com/github/linguist/releases/tag/v3.0.0) Linguist expects a git repository (in the form of a [Rugged::Repository](https://github.com/libgit2/rugged#repositories)) to be passed when initializing `Linguist::Repository`.
 
 
 ```ruby

lib/linguist/languages.yml

@@ -958,7 +958,7 @@ HTML+Django:
 
 HTML+ERB:
   type: markup
-  tm_scope: text.html.erb
+  tm_scope: text.html.ruby
   group: HTML
   lexer: RHTML
   aliases:

lib/linguist/repository.rb

@@ -128,13 +128,20 @@ module Linguist
     protected
 
     def compute_stats(old_commit_oid, cache = nil)
-      file_map = cache ? cache.dup : {}
       old_tree = old_commit_oid && Rugged::Commit.lookup(repository, old_commit_oid).tree
 
       read_index
 
       diff = Rugged::Tree.diff(repository, old_tree, current_tree)
 
+      # Clear file map and fetch full diff if any .gitattributes files are changed
+      if cache && diff.each_delta.any? { |delta| File.basename(delta.new_file[:path]) == ".gitattributes" }
+        diff = Rugged::Tree.diff(repository, old_tree = nil, current_tree)
+        file_map = {}
+      else
+        file_map = cache ? cache.dup : {}
+      end
+
       diff.each_delta do |delta|
         old = delta.old_file[:path]
         new = delta.new_file[:path]

lib/linguist/version.rb

@@ -1,3 +1,3 @@
 module Linguist
-  VERSION = "3.3.1"
+  VERSION = "3.4.0"
 end

test/test_repository.rb

@@ -68,6 +68,19 @@ class TestRepository < Test::Unit::TestCase
     assert !repo.breakdown_by_file["Ruby"].empty?
   end
 
+  def test_commit_with_git_attributes_data
+    # Before we had any .gitattributes data
+    old_commit = '4a017d9033f91b2776eb85275463f9613cc371ef'
+    old_repo = linguist_repo(old_commit)
+
+    # With some .gitattributes data
+    attr_commit = '7ee006cbcb2d7261f9e648510a684ee9ac64126b'
+    # It's incremental but should bust the cache
+    new_repo = Linguist::Repository.incremental(rugged_repository, attr_commit, old_commit, old_repo.cache)
+
+    assert new_repo.breakdown_by_file["Java"].include?("lib/linguist.rb")
+  end
+
   def test_linguist_override_vendored?
     attr_commit = '351c1cc8fd57340839bdb400d7812332af80e9bd'
     repo = linguist_repo(attr_commit).read_index