MySQL Lists are EOL. Please join:

List:MySQL++« Previous MessageNext Message »
From:Warren Young Date:January 23 2007 9:50pm
Subject:RELEASE: v2.2.0
View as plain text  
At long last, v2.2.0 is here.  There is only one minor bug fix since the 
beta, but if you didn't see the previous announcement, here's the change 
list since v2.1.1:

     o ColData, const_string, and SQLString can now be constructed
       with an explicit length parameter.  Furthermore, Query
       class's execute(), store() and use() call chains terminate
       in a version taking an explicit length parameter, instead
       of one taking a simple C string.  Together, this means
       that it's now easier to handle data from the SQL server
       containing nulls.  The library is almost certainly not yet
       capable of handling embedded nulls in all cases, but this
       is a big first step towards that.

     o Can now construct a DateTime object from a time_t, and
       convert a DateTime back to a time_t.  Patch by Korolyov Ilya.

     o Changed the way we're handling exported functions in the
       Windows DLL case so that it works more reliably under MinGW.

     o Added proper copy semantics to Connection, so that you get a
       new connection with the same parameters, not just a bitwise
       copy of the object.

     o Using an explicitly thread-safe variant of localtime() for
       time conversions where one is available.

     o Removed ListInsert template from myset.h.  This wasn't used
       within the library, and was never documented, so I'm betting
       that no one actually uses it.

     o Result::copy() was not copying the exception flag in
       all cases.  Fix by Steven Van Ingelgem.

     o Added exrun shell script and exrun.bat files to distribution,
       to avoid linkage errors when running the examples while
       you still have an older version of MySQL++ installed.

       what it really encodes is the version number in the mysql++.h
       file you're using, not the actual library version number.

     o Added mysqlpp::get_library_version(), which returns the
       library version number at build time.  Between this and
       the header version constant, you can check that you're not
       mixing MySQL++ header and library versions.

     o resetdb example uses these new version number affordances to
       double-check that you're not mixing libraries and headers
       from different versions.  This happens easily unless you
       take care of it (such as by using exrun) when you have one
       version of MySQL++ installed and you're trying to build and
       test a new version without blowing away the old one first
       or overwriting it.

     o No longer using recursive Makefiles on Unixy platforms
       or split lib + examples project files on VC++.  Everything is
       handled by a single top-level Makefile or project file, which
       is simpler for the end user, and makes better dependency
       management possible.

     o When looking for the MySQL C library on systems using
       autoconf, looking in .../lib64 wherever we are also looking
       in .../lib.

     o RPM build process no longer depends on Bakefile.  It means
       you have to build the examples when building an RPM even
       though they're never used within the RPM, but it's a better
       tradeoff in my opinion.

     o Updated include and library paths on Windows to reflect
       changes in the most recent MySQL installers.

     o Merged lib/defs.h and lib/platform.h into new file,
       lib/common.h.  Just cleans up the library internals.

     o Fixed build errors on Windows due to recent changes in MySQL.

     o Fixed a few memory leaks and double-deletes in Query class.

     o Fixed compatibility with STLPort's string implementation.
       Patch by dengxy at

     o Fixed a compatibility problem between Set<> template and
       SSQLS.  Patch by Korolyov Ilya.

     o Fixed build bug in SQLQueryParms due to a character
       signedness issue on PowerPC with GCC.  Patch by Michael

     o ~Transaction() can no longer throw exceptions.  It'll just
       quietly eat them, to avoid program termination.  Fix
       suggested by Alex Burton.

     o Fixed thread safety testing in autoconf case, accidentally
       broken during v2.1.0 development cycle.

     o Using Doxygen 1.5.1 to generate documentation.

As always, you can get it from
RELEASE: v2.2.0Warren Young23 Jan