The build dies in the same place it did for the old ssleay
package:

gcc -I.. -I../../include -DRSAref -DTERMIOS -D_ANSI_SOURCE -O
 -fno-thread-jumps -fno-defer pop -fno-omit-frame-pointer -DBN_ASM
 -Wall -DB_ENDIAN -c -fPIC -DPIC bn_div.c
/usr/src/gnu/usr.bin/egcs/common/../../../dist/gcc/reload.c:3531: 
 Internal compiler error in function find_reloads
*** Error code 1

Moreover, it responds to the same workaround that ssleay did. (See
port-m68k/7152, patch reproduced below.) I propose this is better than
compiling the package with -O0! It would be nice to know if other m68k
ports have the same problem. Anyone?


$NetBSD$

--- crypto/bn/bn_div.c.orig	Thu Apr  9 06:59:28 1998
+++ crypto/bn/bn_div.c	Sun Mar 28 17:04:28 1999
@@ -61,7 +61,7 @@
 #include "bn_lcl.h"
 
 /* The old slow way */
-#if 0
+#if defined (__NetBSD__) && defined(__m68k__) && (__GNUC_MINOR__ == 91)
 int BN_div(dv, rem, m, d,ctx)
 BIGNUM *dv;
 BIGNUM *rem;