[clisp-list] Re: Growing image problem.

[Kaz K. <ka...@as...>]

[I have to do more testing and provide a tiny test case that reproduces
this problem.]

Suppose you have a program that builds as a CLISP image, and you have a
feature in that program whereby it can regenerate itself using sources
and write a new image of itself.

When you repeatedly invoke this feature, you get a successively larger
and larger image, as reported by the figures printed by SAVEINITMEM.

This happens even though garbage collection is done just before calling
by calling (GC) just before (SAVEINITMEM).

In my program, I actually do a DELETE-PACKAGE before re-loading the
modules into the image, so that no symbolic references remain to any of
the top-level functions and variables. So the growth is not from things
like DEFVARS that still have old values, not replaced by the reloading.