From: Warren Young
Date: August 20 2007 7:09pm
Subject: Re: Eyeballs needed on new reference counted pointer template
List-Archive: http://lists.mysql.com/plusplus/6960
Message-Id: <46C9E6EC.6050409@etr-usa.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Jonathan Wakely wrote:
> 
> there's a race on the memory allocation in ResUse::field_types(int).

I'm pretty sure the whole create-on-demand strategy used with FieldNames 
and FieldTypes is a crock, but I haven't put in the time to prove it 
yet.  Aside from the thread safety implications, it means you have to 
test the pointer all the time to be sure it's allocated, and if not, 
repeat the allocation code inline.  Bleh.

At *minimum* we need to hide the pointer behind an accessor that wraps 
all this.  If we can just allocate it on object creation without 
horrible consequences, so much the better.