At 13:05 03.10.2002, Derick Rethans wrote: >On Thu, 3 Oct 2002, Marcus Börger wrote: > > > You do not test implicit flush here. > > You only ensure by testing that implicit_flush > > has no impact on some internal functions. > >I know, that was exactly what I wanted to test. I'll change the >description. > > > One of Yasuos commit you removed was intended > > to ensure that buffering cannot be skipped in anyway > > when such internal functions are used. From my > > point of view Yasuos commit is not a real solution > > since it uses two more functions. > > > > In my outstanding patch mentioned already i am > > addressing this issue by adding new functions call > > php_ob_capture_start()/stop() and a better handling > > of the variables affecting buffering actions. > >I don't think these two new functions are needed, as ob capturing from >internal functions has been working for 10 months without problems. >However I might have 'broken' something again by reverting the patches >regarding to implicit flushing, but I think we should first have >somebody look at it with great knowlegde of all these things, and I'll >ask Zeev to do this, as he already offered his support.
The reason for the new functions is a) make the code more understandable: We use output buffering for buffering, transformation and capturing b) make changes easier since only code in output.c must be changed. See last commits... marcus -- PHP Development Mailing List <http://www.php.net/> To unsubscribe, visit: http://www.php.net/unsub.php
