From: Warren Young
Date: August 23 2008 3:30pm
Subject: Re: Unable to build MySQL++ on Solaris
List-Archive: http://lists.mysql.com/plusplus/7914
Message-Id: <0E9C7692-ABB4-4818-8F02-6507BCB0FE1E@etr-usa.com>
MIME-Version: 1.0 (Apple Message framework v928.1)
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit

On Aug 23, 2008, at 1:05 AM, Kartik Mahajan wrote:

> "The compiler option -m64 will build a 64-bit lib."

You would add that with a command like this:

	./configure CXXFLAGS=-m64 --my-other-configure-flags

> /rusers/Charging/kartikm/RtConnectionPoolManager/mysql++-3.0.5>

You are aware that 3.0.6 is out and that it directly addresses the  
Solaris issues you reported here last week, right?  I spent last  
Saturday finishing it, directly as a result of our back-and-forth  
then.  I wouldn't have rushed it if I didn't think you had an  
immediate need for it.

Speaking of help offered and ignored, why are you continuing to try to  
build MySQL++ against a private copy of MySQL 5 when there's a  
perfectly good MySQL 4 install on the machine already?

> I don't know why it is giving the error " checking if we can link to  
> MySQL C
> API library directly... no"

libmysqlclient may or may not depend on libz, depending on how it was  
built.  Worse, sometimes linking to libz if it isn't actually needed  
gives a build error.  The configure script is trying to build a  
trivial MySQL C API program and link it to libmysqlclient (only) and  
failing.  It then figures that it has to try linking to libz in  
addition, and that doesn't fix it, so it gives up.  It probably  
doesn't really need libz, but is failing to link to libmysqlclient for  
some other reason.  You'll have to read config.log to find out why.

I guess the configure script should try to tease these two cases  
apart, but for now, it's enough that the real cause is recognizable.