🚀 go-pugleaf

In a attempt to port gnuchess-4.0 (pl61) to my 386BSD system I got the
strong impression that this program does not get the same carefull
treatment as the other members of the GNU-family. Here is why I think so:
 - (cd src; make ; make install)
   inevitable generates `NO LANGFILE'. Somebody implemented a `-L language'
   option in main.c but forgot that InitConst(Lang) does not initiate the
   language file `Lang' but LANGFILE.
 - the option -DAG4 which is still present in the Makefile is never used.
 - the most ugly code I saw is in the function Initialize_ttable () from
   init.c (from the indentation one can see that the programmer was in a
   great hurry and certainly had memory enough on his computer - what
   happens if ttblsize <= (1<<13)?)

 (I think I can easily find more proves ...)

I sincerely hope that the next version will be a lot more reliable.

--Piet

--from init.c ----
Initialize_ttable ()
{
  int doit = true;
  if (rehash < 0)
    rehash = MAXrehash;
while(doit && ttblsize > (1<<13)){
  ttable[0] = (struct hashentry *)malloc((unsigned)(sizeof(struct hashentry))*(ttblsize+rehash));
  ttable[1] = (struct hashentry *)malloc((unsigned)(sizeof(struct hashentry))*(ttblsize+rehash));
  if(ttable[0] == NULL || ttable[1] == NULL){
  if(ttable[0] != NULL)free(ttable[0]);
  ttblsize = ttblsize>>1;
  } else doit = false;
}
  if(ttable[0] == NULL || ttable[1] == NULL){ perror("memory alloc");exit(1);}
  printf("transposition table is %d\n",ttblsize);
}