On 1/18/2011 7:34 AM, Jim Jagielski wrote: > > On Jan 18, 2011, at 8:26 AM, William A. Rowe Jr. wrote: > >> On 1/17/2011 3:38 PM, [email protected] wrote: >>> Author: jim >>> Date: Mon Jan 17 21:38:08 2011 >>> New Revision: 1060105 >>> >>> URL: http://svn.apache.org/viewvc?rev=1060105&view=rev >>> Log: >>> Fix cases where off_t (and APR_OFF_T_FMT) may be "larger" than >>> int64 (and APR_INT64_T_FMT). >> >> Simple question, does it makes sense to decouple this from APR_INT64_T >> entirely? >> > > I'm not sure... We certainly have more history with the logic > being tied to int64 and so I feel "safer" keeping it coupled > with that, but making the singular exception when off_t > int64. > Even if they are the same size, we use the older, proven code path.
I totally agree for 1.x, I was thinking strictly of 2.0. Like you say, it would be a change deserving of significant attention, like a major revision would attract.
