On Aug 23, 2007, at 4:05 PM, William A. Rowe, Jr. wrote:
Ruediger Pluem wrote:
But I admit that this is harder to audit and is more likely to
change at some
point of time to the usage of a pool.
More to the point, implementation of apr_ctime. The alternative of
no error
at all or no timestamp seemed worse, to me. Maybe an XXX comment
on trunk
to that effect? (I don't so much care about 0.9/1.2 which aren't
moving
targets, like trunk).
Yeah, the conditions and assumptions on which this
is based warrant some comments in the code :)
