ACHTUNG: THE CPDO WIKI IS NOW (AS OF 2011-May-17) MAINTAINED ON A DEDICATED WIKI SITE: http://whiki.wanderinghorse.net/wikis/cpdo/?page=News
New features and significant changes are briefly described below. The newest items are at the top...
- Started porting the wiki over to http://whiki.wanderinghorse.net/wikis/cpdo/. The main source repo's wiki will be kept for historical reasons (and because Fossil gives us no way to remove pages), but will no longer be maintained.
- The new C++ wrapper API seems to be working.
- Added C++ variant of the AmalgamationBuild.
- Several C-side fixes for compiling in C++ mode.
- Holy cow... repeated tests are showing me that cpdo tests running on a 32-bit virtual machine hosted on 64-bit hardware consistently out-perform the native 64-bit build running on 64-bit hardware. The 32-bits-in-a-vm builds are consistently more than 50% faster. That's just plain weird.
- Got named parameter support working in cpdo_mysql5 by working around MySQL and using a custom parameter name parser. It should work just fine on any reasonably legal SQL code, and can be disabled if it causes grief. (The only constructs i've gotten it to fail on are illegal SQL, so it wasn't actually creating a problem where there wasn't one already.)
- Added cpdo_skeleton, a non-functional (but compilable) cpdo_driver skeleton to bootstrap the process of creating new drivers.
- Added printf-like formatting variants for most of the string-taking functions, e.g. cpdo_exec_f(), cpdo_prepare_f(), and cpdo_bind_string_f().
- Added support for bound parameter name handling. Alas, the MySQL API does not allow us to do this, so that driver does not support this.
- Got the API mostly where i want it. Only missing a small number of features, e.g. binding parameters by name.
- MySQL and sqlite3 drivers both pass the sanity tests, but i need to add more stress tests.