sorry but that i dont understand. The factory makes certain assumptions about what is possible and what not in creating an adapter. Changing the assumptions is a B/C break.
Of course its a bug that ZendX Firebird cannot be loaded from the previous assumptions but that is not a valid cause to overthrow them. There has to be a BC way to solve both cases. Additionally this is a change in a mini version, breaking code that was perfectly valid and should still be valid in all versions from 1.0 to 1.9.2 updates. I cant understand that reasoning. greets, Benjamin On Tue, 22 Sep 2009 11:46:27 -0400, Matthew Weier O'Phinney <matt...@zend.com> wrote: > -- till <klimp...@gmail.com> wrote > (on Tuesday, 22 September 2009, 05:09 PM +0200): >> On Tue, Sep 22, 2009 at 11:39 AM, Jonathan Maron >> <jonathan.a.ma...@gmail.com> wrote: >> > This modification is important to note in ZF 1.9.3: >> > >> > http://framework.zend.com/issues/browse/ZF-5606 >> > >> > The change will probably break some (older) applications. >> > >> > Correct: >> > >> > $db = Zend_Db::factory('Pdo_Mysql', $params); >> > >> > Incorrect (as of ZF 1.9.3) >> > >> > $db = Zend_Db::factory('PDO_MYSQL', $params); >> >> I'm just wondering why this was "fixed" now and not in 2.0? > > Because it was leading to other issues. > > BC breaks, while regrettable, are allowed if they fix a more fundamental > issue.