> 2b) offer a curl_free() (or similar) function
This seems to me to be the simplest, safest, most straight forward solution to
the issue. These seem to be good and desireable design goals. And it is pretty
trivial for existing applications of the library to be converted to it (change
the name of the routine that frees the storage that was actually returned by
curl). If the application can also call the curl_malloc() routine (whatever its
name is ;-), then the application could also use a single memory model for all
of its storage (in cases where it may be "confused" about who allocated which
memory). So, even the most complex cases seem to be converted in a
straightforward manner (by complete conversion to curl's memory management
routines).

Peace. -njg

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Received on 2002-09-20