List:MySQL++« Previous MessageNext Message »
From:Warren Young Date:March 24 2008 11:31pm
Subject:RELEASE: v3.0.1
View as plain text  
Several post-3.0.0 fixes, plus minor enhancements.  A few of these do 
break the ABI again, which is against "the rules," but I'm allowing it 
this time since it's in the name of bug fixing.  I don't expect to do 
this again until 4.0.

     o String objects can now be compared to mysqlpp::null directly.

     o Added a template operator== and operator!= to String, syntactic
       sugar for the existing String::compare() methods.

     o String::compare() now returns 0 ("equal") when one of the
       strings is an uninitialized String() (no refcounted buffer)
       and the other is empty.  It used to consider any initialized
       string greater than an uninitted one.  An uninitialized String
       appears empty, though, so this was incorrect.

     o Made Connection::thread_aware(), thread_start() and thread_end()
       static methods, so they can be called before you create your
       first connection.  Ditto for DBDriver versions of these methods.

     o Calling Connection::thread_start() and thread_end() in
       examples/cpool.cpp, as appropriate.  Above changes were necessary
       to make this work sensibly.

     o Made ConnectionPool::release() virtual, so your pool subclass can
       override it.

     o Added ConnectionPool::size(), so a subclass can know the current
       number of extant connections.

     o No longer single-quoting NOW() call generated for default
       init of DateTime type when building queries from SSQLS objects
       in Query::insert(), replace() and update().  The template query
       and stream interfaces of Query treated NOW() correctly already.

     o Fixed a bug that left SSQLS::table_override_ uninitted if
       you used certain of the generated ctors or set() member
       functions instead of others used by the examples.  This could
       cause a crash any time you caused SSQLS.table() to be called,
       such as when passing the SSQLS to Query::insert().

     o Minor memset bug fixed in test/uds.cpp.  Patch by Dave Jones.

As always, it's available at
RELEASE: v3.0.1Warren Young25 Mar
  • Loss of float precisionJim Wallace1 Apr
    • Re: Loss of float precisionWarren Young2 Apr
      • Re: Loss of float precisionDave Bailey2 Apr
        • Re: Loss of float precisionWarren Young2 Apr
          • RE: Loss of float precisionJim Wallace2 Apr
            • Re: Loss of float precisionWarren Young2 Apr