MySQL Lists are EOL. Please join:

List:Commits« Previous MessageNext Message »
From:pekka Date:August 29 2007 6:45pm
Subject:bk commit into 5.0 tree (pekka:1.2476) BUG#29102
View as plain text  
Below is the list of changes that have just been committed into a local
5.0 repository of pekka. When pekka does a push these changes will
be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://dev.mysql.com/doc/mysql/en/installing-source-tree.html

ChangeSet@stripped, 2007-08-29 21:45:53+03:00, pekka@stripped +1 -0
  ndb - bug#29102 : use locked read even for blob parts

  ndb/src/ndbapi/NdbBlob.cpp@stripped, 2007-08-29 21:43:41+03:00, pekka@stripped +6 -1
    race condition : s/committedRead/readTuple/ when reading parts since TUP commits tuples separately

# This is a BitKeeper patch.  What follows are the unified diffs for the
# set of deltas contained in the patch.  The rest of the patch, the part
# that BitKeeper cares about, is below these diffs.
# User:	pekka
# Host:	sama.ndb.mysql.com
# Root:	/export/space/pekka/ndb/version/my50-bug29102

--- 1.35/ndb/src/ndbapi/NdbBlob.cpp	2007-04-25 21:57:01 +03:00
+++ 1.36/ndb/src/ndbapi/NdbBlob.cpp	2007-08-29 21:43:41 +03:00
@@ -892,7 +892,12 @@
   while (n < count) {
     NdbOperation* tOp = theNdbCon->getNdbOperation(theBlobTable);
     if (tOp == NULL ||
-        tOp->committedRead() == -1 ||
+        /*
+         * This was committedRead() before.  However lock on main
+         * table tuple does not fully protect blob parts since DBTUP
+         * commits each tuple separately.
+         */
+        tOp->readTuple() == -1 ||
         setPartKeyValue(tOp, part + n) == -1 ||
         tOp->getValue((Uint32)3, buf) == NULL) {
       setErrorCode(tOp);
Thread
bk commit into 5.0 tree (pekka:1.2476) BUG#29102pekka29 Aug