[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Infinite loop bug in cache stuff

** Reply to rlp@mitre-bedford.arpa or rich@mitre.org **


I have just received some mail about this from the commonloops list.
The problem is that I reported this bug back in January.  No one
from Xerox (or anybody else) has ever responded to me about this.  Therefore,
I have had to continue use AAAI PCL.  Besides just being angry, I have a few

1.  If a bug report is made to the list, someone should a least acknowledge
    that it has been put on a bug list, and that bug could be given a number.

2.  When fixes are posted to the net, people could reference a number.

3.  It seems to me, from various postings, that there is a bug list at
    Xerox.  If that list was on arisia, people could see if their bug has 
    been reported, and if there is any fix.  This would be particularly useful
    for people who just started reading the mailing list.  That could prevent 
    people from wasting there time (like <waldemar@hx.LCS.MIT.EDU> did) 
    trying to find out what is causing their bug.  My bug report in January
    essentially gave the same info as in the 2 posting from 
    <waldemar@hx.LCS.MIT.EDU>.  If there was a bug list, he could have looked 
    at it and found the same info without having to spend how many hours 
    (days?) wading thru the PCL code.

Now onto the bug. As I and <waldemar@hx.LCS.MIT.EDU> have reported, it
is difficult to repeat this bug on a small example.  But it is NOT
intermittent.  I get it everytime I use my STRIPS-like planner on a specific
problem.  However, if you just look at the cache code, it is pretty obvious 
that something is wrong.

I realize that PCL is not a product, but if bug reports were a bit more 
organized, it would contribute to the development of a more efficient and
usable PCL.

	Rich Piazza