Thorsten Behrens wrote:
[EMAIL PROTECTED] writes:/** Add a new weak reference. The implementation keeps the amount of memory consumed linear in the number of living references added, not linear in the number of total references added. @param e a non-null reference. */Hi Stephan, you make my day (again) - I have a variant of this in slideshow. Small nit: what about doing the check for stale references only every so often? Otherwise, time complexity for clients might easily become quadratic...
But how often should it be done? If done on only every k-th invocation of add for some fix k then the complexity would still be quadratic.
x-post & fup2 discuss
[at least the x-post part seems to have not worked] -Stephan --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
