When is the next release of POI planned. I'd like to pick up this change.
Thanks,
Sanjiv
PS : I have build locally based on the CVS but since our app is going into
production in a month, I'd would like to be using an official release/patch
version.
On 3/25/05, Michael Zalewski <[EMAIL PROTECTED]> wrote:
>
> When I wrote this comment, I thought after I posted -- This is ridiculous.
> No sane person is going to try to rewrite BinaryTree like I suggested.
> Such
> a change must surely destroy every other dependant structure, which
> includes
> the file structure itself. Only a raving lunatic would actually spend time
> on this.
>
> So I tried :)
>
> Result was memory requirements decreased by about 1/3. (My file consists
> of
> only Strings, and the strings are short. I suspect most use cases would
> not
> see such an improvement in memory requirements).
>
> However, time to load my 65,000 unique string workbook decreased by a
> factor
> of of almost 10 (from over 5 minutes to about 30 sec). The strange
> phenomenon with the CPU going idle happened briefly for less than 3 sec,
> and
> only one time.
>
> Here is part of the code (to make my idea more clear)
>
> private static final class Node
> implements Map.Entry
> {
> private Comparable _dataKey; // instead of Comparable[] _data
> private Comparable _dataData;
> private Node _leftKey; // instead of Node[] _left
> private Node _leftData;
> private Node _rightKey; // instead of Node[] _right
> private Node _rightData;
> private Node _parentKey; // instead of Node[] _parent
> private Node _parentData;
> private boolean _blackKey; // instead of Boolean[] _black
> private boolean _blackData;
> private int _hashcode;
> private boolean _calculated_hashcode;
>
> /**
> * Make a new cell with given key and value, and with null
> * links, and black (true) colors.
> *
> * @param key
> * @param value
> */
>
> Node(final Comparable key, final Comparable value)
> {
> _dataKey = key; // much shorter ctor
> _dataData = value; // does not create any arrays
> }
>
> I'll put this into Bugzilla soon to start the discussion. But I should
> point
> out that I have run practically no tests. Gotta find the tests first.
>
> Are there any tests?
>
>
