Re: GC lossage

>>>>> "Eli" == Eli Naeher <[email protected]> writes:

    Eli> We have recently been seeing processes dropping into ldb
    Eli> under 19c with GC lossage errors (example errors with
    Eli> backtraces at the end of this post). I haven't yet tested

Read the messages closely. They all say 

    CMUCL has run out of dynamic heap space (512 MB).
     You can control heap size with the -dynamic-space-size commandline option.

It's a bug that you don't get the heap overflow restarts, but that's
really irrelevant here.  You have run out of heap.  You can use the
-dynamic-space-size option to increase your heap.

However, if you don't expect to be using 512 MB of heap, then it might
be a problem.

    Eli> I did see this in the 19d (pre-)release notes:

    Eli>       - Some issues with GC of weak pointers and weak hash-tables have
    Eli>         been fixed.  (CMUCL would sometimes crash to ldb about weird,
    Eli>         invalid objects.)  There are, however, still issues with weak
    Eli>         pointers.

    Eli> Is it this issue to which the notes are referring?

I think this is relevant only if you using weak pointers or weak
hash-tables.  Are you?

Ray