From db243b9794a83d238768eb6a60fa0aec88d5a694 Mon Sep 17 00:00:00 2001
From: Gregory Burd <greg@basho.com>
Date: Mon, 25 Jun 2012 11:54:38 +0530
Subject: [PATCH] Don't bother writing out the bloom filter when a level is
 empty, instead just write 0 for the bloom size and before that record the
 level's filter size so as to re-create a properly sized filter later when
 reading this file.

---
 src/hanoidb_reader.erl | 21 +++++++++++++++++----
 src/hanoidb_writer.erl | 18 ++++++++----------
 2 files changed, 25 insertions(+), 14 deletions(-)

diff --git a/src/hanoidb_reader.erl b/src/hanoidb_reader.erl
index 6ac867d..60c7b63 100644
--- a/src/hanoidb_reader.erl
+++ b/src/hanoidb_reader.erl
@@ -79,12 +79,25 @@ open(Name, Config) ->
             %% read root position
             {ok, <<RootPos:64/unsigned>>} = file:pread(File, FileInfo#file_info.size - 8, 8),
             {ok, <<BloomSize:32/unsigned>>} = file:pread(File, FileInfo#file_info.size - 12, 4),
-            {ok, BloomData} = file:pread(File, (FileInfo#file_info.size - 12 - BloomSize), BloomSize),
 
-            Bloom = hanoidb_util:decode_bloom(BloomData),
+            Bloom =
+                case BloomSize of
+                    0 ->
+                        {ok, <<FilterSize:32/unsigned>>} = file:pread(File, (FileInfo#file_info.size - 16), 4),
+                        bloom:new(FilterSize, 0.001);
+                    _ ->
+                        {ok, BloomData} = file:pread(File, (FileInfo#file_info.size - 12 - BloomSize), BloomSize),
+                        hanoidb_util:decode_bloom(BloomData)
+                end,
 
-            %% suck in the root
-            {ok, Root} = read_node(File, RootPos),
+            %% read in the root node
+            Root =
+                case read_node(File, RootPos) of
+                    {ok, Node} ->
+                        Node;
+                    eof ->
+                        none
+                end,
 
             {ok, #index{file=File, root=Root, bloom=Bloom, name=Name, config=Config}}
     end.
diff --git a/src/hanoidb_writer.erl b/src/hanoidb_writer.erl
index 12c9720..6345d99 100644
--- a/src/hanoidb_writer.erl
+++ b/src/hanoidb_writer.erl
@@ -186,27 +186,25 @@ do_open(Name, Options, OpenOpts) ->
 
 
 %% @doc flush pending nodes and write trailer
-flush_nodes(#state{ nodes=[], last_node_pos=LastNodePos, last_node_size=_LastNodeSize, bloom=Bloom }=State) ->
-    IdxFile = State#state.index_file,
+flush_nodes(#state{ nodes=[], last_node_pos=LastNodePos, last_node_size=_LastNodeSize, bloom=Bloom, index_file=IdxFile }=State) ->
 
-    RootPos =
-        case LastNodePos =:= undefined of
-            true ->
+    {BloomBin, BloomSize, RootPos} =
+        case LastNodePos of
+            undefined ->
                 %% store contains no entries
                 ok = file:write(IdxFile, <<0:32,0:16>>),
-                ?FIRST_BLOCK_POS;
+                FilterSize = bloom:filter_size(Bloom),
+                {<<FilterSize:32/unsigned>>, 0, ?FIRST_BLOCK_POS};
             _ ->
-                LastNodePos
+                EncodedBloom = hanoidb_util:encode_bloom(Bloom),
+                {EncodedBloom, byte_size(EncodedBloom), LastNodePos}
         end,
 
-    BloomBin = hanoidb_util:encode_bloom(Bloom),
-    BloomSize = byte_size(BloomBin),
     Trailer = << 0:32, BloomBin/binary, BloomSize:32/unsigned,  RootPos:64/unsigned >>,
 
     ok = file:write(IdxFile, Trailer),
     ok = file:datasync(IdxFile),
     ok = file:close(IdxFile),
-
     {ok, State#state{ index_file=undefined, index_file_pos=undefined }};
 
 flush_nodes(State=#state{ nodes=[#node{level=N, members=[{_,{Pos,_Len}}]}], last_node_pos=Pos })