From: Warren Young Date: November 16 2004 5:40pm Subject: Re: RELEASE: v1.7.21 List-Archive: http://lists.mysql.com/plusplus/3647 Message-Id: <419A3B82.6040402@etr-usa.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Zahroof Mohamed wrote: > I've been using the GCC 3.4.2 for some time now with MingW Studio on > Windows 2000. Because of the circular dependancy between Connection > and Query, I could not use 1.7.21. Are you saying that an earlier version of MySQL++ worked with GCC 3.4.2, or that your switch to GCC 3.4.2 broke all MySQL++ versions? If the former, please be explicit about which version worked. > Any new ideas on how to fix this problem? My only idea is the brute force one: break the API and ABI. I'm waiting for someone to come up with a better idea. It's been a week since I solicited opinions, and nothing useful has come up yet. > In the header file compare1.hh, in the declaration of class > MysqlcprCStr, line 29, index is not defined. I added > protected: > unsigned int index; > to this class to solve the problem. Incorrect fix. index is a member element of the parent class. I don't know why GCC cannot see it. I've had to add explicit qualifiers to MySQL++ before to work around these things. This one looks like a GCC bug to me. > My compiler showed several warnings - one for all the #pragma stuff > that was added for the benefit of windows users using Visual C++. I > had to comment out all the #pragma stuff to get rid of the warnings. Instead of hacking those pragmas out -- which are necessary on some platforms -- why don't you send me a patch to platform.h (nee' mysql++-windows.hh) that will allow the MinGW GCC to skip those pragmas? MySQL++ doesn't move forward on platforms I don't use without patches. > With GCC 3.4.2, I also got the warning that had been > deprecated and to use instead. This has been fixed since 1.7.19.