On 2/7/2012 7:15 AM, Jim Jagielski wrote: > > On Feb 7, 2012, at 12:51 AM, William A. Rowe Jr. wrote: > >> On 2/6/2012 5:19 PM, Daniel Ruggeri wrote: >>> On 2/5/2012 1:10 PM, William A. Rowe Jr. wrote: >>>> How about >>>> >>>> 1. deprecating --with-included-apr syntax for... >>>> >>>> 2. --build-apr=/path syntax, default of '--build-apr=srcdir/apr' where >>>> and when srclib/apr exists. If they would rather check out httpd, apr >>>> and apr-util in parallel, it will be easy enough to choose something >>>> like --build-apr=../apr-1.4.5 >>> >>> Can we be sure that the configure script will continue to pass >>> parameters as it does today to these "sub" configure scripts, too? >> >> I thought that would be implicit ;-) > > As soon as Bill completes the above, I think we can progress > with discussions re: T&R 2.4.1.
What was consensus? I'm not really sure I read one; could you tally the discussion Jim? --build-apr only makes sense if we strip out 'builtin' srclib/apr. I was making a suggestion to Kaspar, though :) Spending my time in the morass of 2.0.65 so that we can T&R the vulnerable code and call it 'done, final, gone' (or nearly so). That, and getting patches into mod_ftp and tagging both the long neglected ftp and fcgid modules for votes. fcgid has some substantial fixes already in tree. My 2.4.x time is given over to solving disconnect/reuse socket/ ssl timeout flukes on windows right now, because I really want to fix any remaining edge cases. I do love hacking at autoconf magic (heh heh), but can't get to it this moment. If anyone else wants to play in the .m4 code, please 'steal this idea' :) Otherwise I am happy to do it, more on a timetable of early next week though.
