From: Warren Young Date: November 6 2004 1:16am Subject: Re: RELEASE: v1.7.20 List-Archive: http://lists.mysql.com/plusplus/3598 Message-Id: <418C2602.2090308@etr-usa.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Darren McDonald wrote: > Ok, I've seen more updates in the last few months that there have been > in the last two years, Hmmm, more updates than in the last four years, actually. :) There was a 3.5 year dead space between 1.7.9 and 1.7.10. > 1: Right now, which release is stable? I don't consider that any new "features" have been added while I've been involved. It's all been maintenance and polishing. As long as this situation continues, the stable release is the current release. If the current release is ever discovered to have serious problems, I withdraw it and roll the site back to the previous version, or I replace it with a fixed version. When/if we start adding features, then we may split things into stable and development braches. But I really don't see that happening, because there are no big features under consideration. If you're asking when I'm going to stop breaking the API and ABI, I think I just finished with that. :) Nothing else on the wishlist would seem to impact these. > 2: Do you guys need a hand with anything? The Wishlist is about as long as ever. Here's the current version of the list, broken down into maintenance vs. new feature items: New Feature Requests -------------------- o Look through my own code for uses of custom SQL queries, and try to reduce them to API features. o Overload SQLQuery::insert() to accept a pair of STL iterators for mass insertion of a range within a container. Bug Fix/Maintenance Items ------------------------- o is_connected() may not be returning the right value. o Does it build under GCC 3.4 yet? o One report says it doesn't build under what will become GCC 4.0. o A report says Query objects cannot be reused. Verify. o VC++ supports 64-bit ints, but not as "long long"s as GCC does. Is the stuff in defs.h sufficient to fix up this difference? o Documentation needs a thorough reworking. o Factor tests out of configure.in into individual files in a config subdir. o Fix: "coldata4.hh:11: warning: unused parameter `int dummy'". o SQLQuery shouldn't derive from std::stringstream. It should have a private std::stringstream member, and add the function definitions it needs to look like a stringstream. The Standard C++ library classes were not designed to be subclassed. o Split RPM into lib, devel and examples sections. For the latter, make a simple non-autotools Makefile that looks in the same directory as the lib and devel RPMs unpack into for the library and headers. o Rename *.cc to *.cpp. > Something simple to start off with The first five items in the maintenance section are ripe candidates for you. They would each be easy enough to check and fix if necessary, I just haven't gotten around to them. The first item on the feature list would also be good for you to tackle. Basically, it involves finding any place in your code where you're stuffing custom SQL into a Query object via its stream interface, and seeing if there is a feature we could add to the library to make that custom SQL unnecessary.