Mailing-List: contact commits-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@couchdb.apache.org
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
From: kocolosk@apache.org
To: commits@couchdb.apache.org
Subject: git commit: Fix retrieval of headers larger than 4k
Message-Id: <20111030182833.C86C45488E@tyr.zones.apache.org>
Date: Sun, 30 Oct 2011 18:28:33 +0000 (UTC)

Updated Branches:
  refs/heads/master 5f906a39b -> 1dc866b3c


Fix retrieval of headers larger than 4k

Our headers start with a <<1>> and then four bytes indicating the length
of the header and its checksum. When the header is larger than 4090
bytes it will be split across multiple blocks in the file and will need
to be reassembled on read. The reassembly consists of stripping out
<<0>> from the beginning of each subsequent block in the
remove_block_prefixes/2 function. The bug here is that we tell
remove_block_prefixes that we're starting 1 byte into the current block
instead of 5, so it ends up removing one good byte from the header and
injecting one or more random <<0>>s.

Headers larger than 4k are very rare and generally require a view group
with a huge number of indexes or indexes with fairly large reductions,
which explains why this bug has gone undetected until now.

Closes COUCHDB-1319.


Project: http://git-wip-us.apache.org/repos/asf/couchdb/repo
Commit: http://git-wip-us.apache.org/repos/asf/couchdb/commit/1dc866b3
Tree: http://git-wip-us.apache.org/repos/asf/couchdb/tree/1dc866b3
Diff: http://git-wip-us.apache.org/repos/asf/couchdb/diff/1dc866b3

Branch: refs/heads/master
Commit: 1dc866b3c795d7d7515f732f444cd54e656390bc
Parents: 5f906a3
Author: Adam Kocoloski <kocolosk@apache.org>
Authored: Wed Oct 26 14:04:54 2011 -0400
Committer: Adam Kocoloski <kocolosk@apache.org>
Committed: Sun Oct 30 14:27:39 2011 -0400

----------------------------------------------------------------------
 src/couchdb/couch_file.erl   |    2 +-
 test/etap/011-file-headers.t |    9 ++++++++-
 2 files changed, 9 insertions(+), 2 deletions(-)
----------------------------------------------------------------------


http://git-wip-us.apache.org/repos/asf/couchdb/blob/1dc866b3/src/couchdb/couch_file.erl
----------------------------------------------------------------------
diff --git a/src/couchdb/couch_file.erl b/src/couchdb/couch_file.erl
index f1d6cc4..2c2f11a 100644
--- a/src/couchdb/couch_file.erl
+++ b/src/couchdb/couch_file.erl
@@ -433,7 +433,7 @@ load_header(Fd, Block) ->
         RawBin = <<RestBlock/binary, Missing/binary>>
     end,
     <<Md5Sig:16/binary, HeaderBin/binary>> =
-        iolist_to_binary(remove_block_prefixes(1, RawBin)),
+        iolist_to_binary(remove_block_prefixes(5, RawBin)),
     Md5Sig = couch_util:md5(HeaderBin),
     {ok, HeaderBin}.
 

http://git-wip-us.apache.org/repos/asf/couchdb/blob/1dc866b3/test/etap/011-file-headers.t
----------------------------------------------------------------------
diff --git a/test/etap/011-file-headers.t b/test/etap/011-file-headers.t
index 81ffdb2..a26b032 100755
--- a/test/etap/011-file-headers.t
+++ b/test/etap/011-file-headers.t
@@ -22,7 +22,7 @@ main(_) ->
     {S1, S2, S3} = now(),
     random:seed(S1, S2, S3),
 
-    etap:plan(17),
+    etap:plan(18),
     case (catch test()) of
         ok ->
             etap:end_tests();
@@ -68,6 +68,13 @@ test() ->
     etap:is({ok, Size2}, couch_file:bytes(Fd),
         "Rewriting the same second header returns the same second size."),
 
+    couch_file:write_header(Fd, erlang:make_tuple(5000, <<"CouchDB">>)),
+    etap:is(
+        couch_file:read_header(Fd),
+        {ok, erlang:make_tuple(5000, <<"CouchDB">>)},
+        "Headers larger than the block size can be saved (COUCHDB-1319)"
+    ),
+
     ok = couch_file:close(Fd),
 
     % Now for the fun stuff. Try corrupting the second header and see