On Wed, 20 Aug 1997, Daniel Koepke wrote:
>>From what I can tell (and it may not be accurate, since I've not bothered
>to look at the patch), the patch only rounds up to 128 when you request
>a buffer smaller than 128. Otherwise, it uses: first, any buffer that
>is exactly the right size; second, the smallest existing buffer > the
>requested size; finally, a newly created buffer of the exact size
>requested.
Daniel is accurate even when he doesn't look at what he's talking about. ;)
>Although, rounding-up might not be a bad idea. It'd certainly reduce
>the number of malloc() calls (eg., 300 would use a 512 buffer, hence
>no malloc() of a 300 buffer; later requests for a 512 buffer will get
>that 512 buffer, thus knocking off another malloc() call).
Right, although if someone requests a 32 byte buffer and everything else is
filled, it will create a 32 byte buffer for it. It will probably time out
later if not used or you can forcibly kill it with
'buffer delete 32 temporary' (another new feature).
--
greerga@muohio.edu me@null.net | Genius may have its limitations, but stupidity
http://www.muohio.edu/~greerga | is not thus handicapped. -- Elbert Hubbard
+------------------------------------------------------------+
| Ensure that you have read the CircleMUD Mailing List FAQ: |
| http://democracy.queensu.ca/~fletcher/Circle/list-faq.html |
+------------------------------------------------------------+
This archive was generated by hypermail 2b30 : 12/08/00 PST