Flavio S. Glock wrote:
2006/5/18, Jerrad Pierce <[EMAIL PROTECTED]>:
I've run into an odd issue with DateTime and Memoization ...
DateTime objects can mutate, and this can potentially
place the memoization buffer in an invalid state, because the internal
data is shared.
One solution could be to deep clone every DT before modification,
but this can be as expensive as not memoizing at all.
I think this should be mentioned in the docs. A quick grep for
memoi[sz]e|mutate|share of the perldoc didn't show anything appropriate.
I think a new CAVEATS or WARNINGS section in the doc would be the
right place for this, as it should be kept seperate from the list of
KNOWN BUGS.
--
David Cantrell
