Re: [Xsb-development] Deleting tables with conditional answers and trie allocation.

[Terrance S. <ts...@cs...>]

The problem is that if you abolish a conditional answer A, there
may be pointers from various other delay lists or delay elements
to A (abolishing subgoals can work similarly) that remain
untouched.  These dangling pointers have always been a semantic
problem, but didnt always cause a core dump before 3.0 as we
didnt always reclaim space for completed tables (however there
were many other bugs in the pre-3.0 table abolishing routines).
With space reclamation, if simplification (or other operations)
occurs, then space reclamation can cause a crash.

Solving the problem in general took more time than I had
available, as the delay list code dependencies be a little
difficult to navigate.  The result is that I decided to put in
the warning that you saw as a termporary device.

I'll try to think of a reasonable implementational solution over
the weekend.  In the mean time, abolish_all_tables/0 is always
correct as this gets rid of the delay dependencies in the table
as well as the call and answer tries.

I'm not sure, but I don't believe that the new incremental
tabling handles conditional answers properly (?)

Terry


On Wed, 8 Nov 2006, [ISO-8859-1] Gon=E7alo Lopes wrote:

> Hi all,
>
> I'm currently in the process of porting a program from XSB 2.6 / 2.7.1
> to XSB 3.0.1 and I'm getting the following warning message when I call
> abolish_table_pred/1:
>
> ++Warning[XSB]:  [Runtime/C] abolish_table_pred/1 is deleting a table
> with conditional answers: delay dependencies may be corrupted.
>
> Shortly thereafter the program crashes with a memory violation after
> calling the abolished tabled predicates. I'm using single-threaded
> slg-wam engine.
>
> What has changed from XSB 2.6/2.7.1 that could account for these
> behaviours? And more to the point, how can I guarantee that my tables
> are abolished cleanly, or at least have their previous answers
> invalidated? This is all during a single evaluation of a complex
> predicate, which produces answers from the table, consumes them, but
> afterwards has to remove those answers because external conditions
> have changed and these may no longer be valid.
>
> Thank you for the assistance,
>
> Gon=E7alo
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job ea=
sier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronim=
o
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> _______________________________________________
> Xsb-development mailing list
> Xsb...@li...
> https://lists.sourceforge.net/lists/listinfo/xsb-development
>
>