Bug # 21177 - main/main.c_r1.515 - Second call to sapi_activate?

Ref: http://bugs.php.net/21177 r1.515 seems to have added a second call to sapi_activate. In the case of this particular bug this is causing the following behavior: 1) Call sapi_active: initialize sapi globals 2) sapi/cgi sets SG(headers_sent) = 1 for options like -q & -f 3) New Code in patch r1.515 calls sapi_active again, reinitializing the sapi globals ( in this case SG(headers_sent) = 0 ) Can someone explain this second call to sapi_active? "gschlossnagle" perhaps? -Pollita