lib/gitlab_git/blob.rb in gitlab_git-7.2.24 vs lib/gitlab_git/blob.rb in gitlab_git-8.0.0

@@ -5,10 +5,15 @@
   module Git
     class Blob
       include Linguist::BlobHelper
       include EncodingHelper
 
+      # This number needs to be large enough to allow reliable content /
+      # encoding detection (Linguist) and LFS pointer parsing. All other cases
+      # where we need full blob data should use load_all_data!.
+      DATA_FRAGMENT_SIZE = 1024
+
       attr_accessor :name, :path, :size, :data, :mode, :id, :commit_id
 
       class << self
         def find(repository, sha, path)
           commit = repository.lookup(sha)
@@ -26,11 +31,11 @@
             if blob
               Blob.new(
                 id: blob.oid,
                 name: blob_entry[:name],
                 size: blob.size,
-                data: blob.content,
+                data: blob.content(DATA_FRAGMENT_SIZE),
                 mode: blob_entry[:filemode].to_s(8),
                 path: path,
                 commit_id: sha,
               )
             end
@@ -41,11 +46,11 @@
           blob = repository.lookup(sha)
 
           Blob.new(
             id: blob.oid,
             size: blob.size,
-            data: blob.content,
+            data: blob.content(DATA_FRAGMENT_SIZE),
           )
         end
 
         # Recursive search of blob id by path
         #
@@ -213,9 +218,17 @@
         !data || data == ''
       end
 
       def data
         encode! @data
+      end
+
+      # Load all blob data (not just the first DATA_FRAGMENT_SIZE bytes) into
+      # memory as a Ruby string.
+      def load_all_data!(repository)
+        return if @data == '' # don't mess with submodule blobs
+
+        @data = repository.lookup(id).content
       end
 
       def name
         encode! @name
       end