On 5/15/2011 1:46 AM, Graham Leggett wrote: > On 15 May 2011, at 1:46 AM, William A. Rowe Jr. wrote: > >> No argument, but there are 1) minor quibbles with the apr-2 interface, and >> 2) some significant work to replace the original with the new interface, and >> not sure who has cycles to attack this in the near term. If it is fixed, >> re-adding mod_session during 2.4.x cycle would be relatively painless, no? > > The only module affected is mod_session_crypto.c, I'm not sure how the scope > has expanded > to cover mod_session*.
Ok, good... if we decouple mod_session_crypto (until apr-2/1.x api is released) and leave the rest, that would be just fine. Folks at the short idea session couldn't think, offhand, if it was strongly coupled to the whole mod_session framework. And it could be readded just as soon as apr releases (with an apr rev check to enable). > Taking out mod_session* would also mean taking out mod_auth_form, which in > turn would be a > lot of work, and would take out a major new feature of v2.4. It would be a > far better use > of our time just taking out mod_session_crypto.c if it had to come down to it. So I'll update status with a patch to 'uncouple' and svn rm mod_session_crypto.c after the branches/2.4.x/ exists, and we can probably extend such a script with other 'very last minute' go/nogo decisions. > Full docs are here: > > http://httpd.apache.org/docs/trunk/mod/mod_auth_form.html > http://httpd.apache.org/docs/trunk/mod/mod_session.html > http://httpd.apache.org/docs/trunk/mod/mod_session_cookie.html > http://httpd.apache.org/docs/trunk/mod/mod_session_dbd.html > http://httpd.apache.org/docs/trunk/mod/mod_session_crypto.html