R. Clayton (rclayton@monmouth.edu)
(no date)
Can the database have duplicate entries?
Effectively no; I don't want to see readers returning duplicate entries.
However you achieve that is up to you.
You write "term() may return while previously issued query() and write()
calls are still in progress" Did you mean to write "term() may BE CALLED
while previously issued query() and write() calls are still in progress"?
This would make more sense to us as, when term() returns, the database should
be gone.
No: term() is non-blocking. When term() returns, the database may still be in
use. However, it won't be put to any new use after term() returns.
will your test harness construct a KWICDB object then call init() on the
main thread before spawning new threads to call the query() and write()
functions?
That's the idea.
We're still a little fuzzy on why you included the init() with the signature
you gave it.
So am I. I may have been intending a factory pattern, or maybe I just wanted
to provide init() for anybody who might want to use it (but that doesn't
explain the signature).
I noticed that the synchronized keyword is NOT a part of a function's
signature and it isn't automatically inherited by the child class (Lea,
p. 76). In class, we had assumed the opposite was true.
You're right (and you can't declare synchronized methods in interfaces);
thanks for pointing it out.
This archive was generated by hypermail 2.0b3 on Tue Aug 20 2002 - 12:45:05 EDT