🚀 go-pugleaf

What has been happening to the Font Manager?

I have managed to survive the migration to 'proper' code utilizing
the fmfreefont calls AND the rotatepagematrix bug in the 4.0.5[a-f]
Font Manager as well upgrading to the new bug-fixed Font Manger
(post 4.0.5F release).

BUT, either I am going crazy or something is still terribly wrong
with the FM.

If I run the following code on a pre-4.0.5 FM system (3.3.x - 4.0.4)
it does not result in any memory loss.  Under 4.0.5a, mm, f, and
f with fixed FM it hums along slowely eating all of my RAM!  I start
up gr_osview and watch while it runs and it just sucks it all away.
I saw 0 memory loss after 1.7million iterations on a 70GT running
3.3.2 but ran well into swap after 60,000 iteration on a 4.0.5f
Entry Indigo.

#include <stdio.h>
#include <gl.h>
#include <device.h>
#include <fmclient.h>

main() {
    long its=0;
    fmfonthandle f,fsized;
    noport();
    winopen("fmfreefont_test");
    gconfig();
    fminit();
    while(!getbutton(LEFTMOUSE)) {
        ++its;
        f = fmfindfont("Times-Roman");
        fsized = fmscalefont(f,18.0);
        fmsetfont(fsized);
        fmfreefont(fsized);
        fmfreefont(f);
    }
    printf("AFTER %ld iterations\n",its);

}

This is a critical issue with me as my code is running in an interactive
exhibit on a Crimson VGX (4.0.5a - 32MB RAM) and crashes every 30min or
so as RAM runs out.  This obviously DOES NOT look good on me, my client,
and SGI (which is a sponsor!).

Can someone tell me what I may be doing incorrectly?  I have spent
considerable time on this issue dealing with the last 2 'items' (BUGS)

I am posting this here since I find this to be the best forum for
rapid knowledgeable responses.

	Rodney Hoinkes
	Head of Design Applications
	Centre for Landscape Research
	University of Toronto
	Tel:   (416) 978-7197
	Email: rodney@dgp.utoronto.ca
	~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	APIAWWGATI
	"Anything's possible IF and WHEN we get around to it."