The fastcgi++ library started out as a C++ alternative to the official FastCGI developers kit. It is released under the GNU Lesser General Public License. Although the official developers kit provided some degree of C++ interface, it was very limited. The goal of this project was to provide a framework that offered all the facilities that the C++ language has to offer. Over time the scope broadened to the point that it became more than just a simple protocol library, but a platform to develop web application under C++. To the dismay of many, this library has zero support for the old CGI protocol. The consensus was that if one were to be developing web applications under C++, efficient memory management and CPU usage would be a top priority, not CGI compatibility. Effective management of simultaneous requests without the need for multiple threads is something that fastcgi++ does best. Session data is organized into meaningful data types as opposed to a series of text strings. Internationalization and Unicode support is another top priority. The library is templated to allow internal wide character use for efficient text processing while code converting down to utf-8 upon transmission to the client.
- Updated boost.m4 macro to lastest version
- Fixed possible end iterator dereferences
- IPv6 address parser no longer accepted non hex characters
- Removed potential for read overflow
- Http::Environment::requestMethod now initialized
- Fcgistream::Encoder::m_state is now initialized
- Uses autotools
- Has session handling capabilities
- Has SQL query capabilities
- Fixed bug in parsing post data
- Fixed buffer flushing bug
- Removed dependence on Linux epoll socket polling. Switched to POSIX poll().
- Code will now compile on older GCC systems.
- Fixed issue where parameters would not be processed if bunched into a single record as mod_fcgid does.
- Refined error handling.
- Implemented an internal UTF-8 code conversion facet to remove dependence on functional UTF-8 locales.