Difference between revisions of "News"
(OpenDBX release 1.3.9) |
m (→2008-05-23 - OpenDBX 1.3.9 released) |
||
Line 6: | Line 6: | ||
The OpenDBX library is heading towards the next stable series and in this release the focus was mainly on fixing bugs in existing code. As a result, some patches were applied to the sqlite3 and odbc backends to make them work even better. | The OpenDBX library is heading towards the next stable series and in this release the focus was mainly on fixing bugs in existing code. As a result, some patches were applied to the sqlite3 and odbc backends to make them work even better. | ||
− | Additionally, as a convenience for developers, odbx_result_finish() makes sure that all rows are fetched and the parameter for Result::getResult() were made optional. The odbx-sql utility was also improved and is capable of auto-completing all ANSI SQL keywords. | + | Additionally, as a convenience for developers, odbx_result_finish() makes sure that all rows are fetched and the parameter for Result::getResult() were made optional. The odbx-sql utility was also improved and is now capable of auto-completing all ANSI SQL keywords. |
Also, man pages for the odbx_lo_*() and odbx_result_finish() functions as well as the odbx-sql application were written. | Also, man pages for the odbx_lo_*() and odbx_result_finish() functions as well as the odbx-sql application were written. |
Revision as of 01:08, 23 May 2008
2008-05-23 - OpenDBX 1.3.9 released
The OpenDBX library is heading towards the next stable series and in this release the focus was mainly on fixing bugs in existing code. As a result, some patches were applied to the sqlite3 and odbc backends to make them work even better.
Additionally, as a convenience for developers, odbx_result_finish() makes sure that all rows are fetched and the parameter for Result::getResult() were made optional. The odbx-sql utility was also improved and is now capable of auto-completing all ANSI SQL keywords.
Also, man pages for the odbx_lo_*() and odbx_result_finish() functions as well as the odbx-sql application were written.
2008-05-02 - OpenDBX 1.3.8 released
The ODBC backend is now completely functional and ready for extensive testing. The OpenDBX C++ library was extended by the Result::columnPos() method to be able to use column names instead of numbers and the methods in the Result class were generally shortend.
Also, odbx_escape() was fixed as it didn't work any more after the change in release 1.3.7. The sqlite3 and sqlite backends contained a few bugs in their result function which are corrected now.
2008-04-23 - OpenDBX 1.2.6 released
This is a security release! The default implementation of odbx_escape() didn't escape single backslashes which may make the backends providing no own implementation vulnerable to SQL injection attacks.
In addition, bugs in the sqlite3 and oracle backends were fixed.
2008-04-12 - OpenDBX 1.3.7 released
Many improvements and bugfixes are included in this release, both in the library itself as well as in the supplementary parts. It's recommended to all developers who are using the OpenDBX library to upgrade to the latest version.
The existing function odbx_result_free() is deprecated in favor of the new odbx_result_finish() which is able to return error codes. Additionally, the C++ Stmt::bind() method of API was removed until the prepared statement API will arrive and the Stmt::escape() method was moved to the connection object.
It's now also possible to write data into large objects in firebird and oracle backends but the oracle LOB implementation doesn't work at the moment. I would be glad if an oracle specialist would be able to have a look what could be wrong.
2008-03-28 - OpenDBX 1.2.5 released
We are continuing to fix some bugs in the stable branch of the OpenDBX library but this release also enables building the sybase backend using FreeTDS version 0.63.
Firebird 2.0 changed it's behavior and doesn't support soft commits any more if two connections operate on the same table. The firebird backend uses hard commits now and additionally does a rollback in odbx_unbind() instead of a commit.
The pgsql_odbx_result() function in the pgsql backend cleans up correctly in case of errors in this release and returns only fatal errors via odbx_error_type() if the connection itself is dead.
2008-02-25 - OpenDBX 1.3.6 released
The new release features a small application for executing SQL statements from a shell environment. It can be used either interactively by typing commands directly into the terminal or via batch mode feed by stdin. It's great for shell scripts working with databases or for testing during application development.
The name of the C++ header file changed to conform to the naming of the STL. Furthermore, the C++ API was enhanced and consequently accepts not only strings but also char pointers. While adding this functions, the parameter signature of the existing functions Conn::create() and Stmt::bind() was streamlined.
2008-02-04 - OpenDBX 1.3.5 released
Finally, this development release contains a prototype of an ODBC backend for the OpenDBX database access library. It's not fully functional yet and misses the code for data retrieval but data definition and manipulation statements work. Also, the complete code does compile cleanly if the latest GCC compiler suite is used.
The automake scripts were improved to add "libdl" only to the core library and to make more clear what's missing if header files or libraries are missing. Minor fixes to the documentation were applied too.
2008-01-16 - OpenDBX 1.2.4 released
This is a bugfix release for the current stable series. It contains a backport from the development branch which removes the automake dependency from the main OpenDBX header file (odbx.h) and fixes the SYBDATETIME4 type definition in the mssql backend. Furthermore it adds support for the SYBLONGCHAR and SYBLONGBINARY types in mssql backend.
It does also contain a fix in the regression test program so it will build on the latest Linux installations. The documentation was updated too.
2007-10-06 - New PHPLights release
There are some major changes in this version of PHPLights in the ways things are done. The factory is completely rewritten to create new instances without a getInstance() method and to handle variable argument lists.
Furthermore, some kind of aspect-oriented programming was introduced which allows extremely reusable components. Also, some parts of the code were streamlined and cleaned up.
2007-09-23 - New PHPLights release
The PHPLights framework makes some progress regarding documentation but there is still much to do. At least there is now rudimentary documentation for all interfaces and the complete class tree available.
As a side effect, some inconsistencies has been fixed and some cleanups has been done while writing the documentation. All changes are compatible to the last release.
2007-09-02 - Initial PHPLights release
PHPLights is a modern, lightweight and PHP5 based application development framework. It makes heavy use of PHPs object oriented features like interfaces, abstract classes and exceptions. Design pattern where used whenever they provide advantages for the implementation making the core library very lean, easy to understand and extensible.
2007-08-06 - OpenDBX 1.3.4 released
The current release adds the C++ interface for the large object functions of the OpenDBX C API, the missing Conn::getCapability() member function and the regression test code for libopendbxplus. For more control over loading the backend modules, it's now possible to supply the full path to the backend library is first parameter to odbx_init().
Also, there was a change of the parameter type of odbx_lo_open() as well as a change regarding the visibility of the C++ class variables. Finally, the pgsql backend returned partly wrong results for boolean values if odbx_field_value() was called twice. This is fixed now.
2007-07-08 - OpenDBX 1.3.3 released
I'm proud to announce the first release which contains the much wanted C++ interface to the OpenDBX library. Much work has been put into this interface to ease development for the C++ community without sacrificing the speed and lightweight approach of the C interface. It wouldn't have been available at this stage without the sponsorship of Metaways.
Integration of the C++ interface was a good start to use a separate directory (/usr/include/opendbx/) for the header files including the new opendbx/api.hpp. The old odbx.h is still available but only as legacy file which includes the new opendbx/api.h header.