[news.software.nn] problems compiling nn 6.4Beta

molenda@cs.umn.edu (Jason "Hax R Us" Molenda) (03/04/90)

I encountered two problems compiling nn version 6.4beta both of which
are really easy to fix.

1) Lines 580 and 581 in term.c should be reveresed.  On the original 
   distribution, they read

     use_vararg;
     int k, l;

   But for systems that don't have vararg stuff, usr_vararg is defined
   to be null so after the cpp gets done we have

     ; int k, l;

   which ain't legal.  So if you switch the lines to

      int k, l;
      use_vararg;

   it will compile correctly.  The following machines will have this
   problem (if NO_VARARGS is defined in your m-<sys>.h then you'll have this
   problem):

            3b1g
            Convex
            Gould
            Pyramid
            Symmetry
            VAX

2)  Line 564 of answer.c reads

          static char prev_aux_opt[5] = "-----";

    cc will die with the error

          "answer.c", line 564: too many initializers

    because the array that is created is prev_aux_opt[0..4] and the
    string is actually 6 chars long (five dashes + NUL).  The fix is 
    to change line 564 to

          static char prev_aux_opt[6] = "-----";

    or reduce the number of dashes by one.  :-)


Both of these problems were encountered on a Sequenty Symmetry S27 running
Dynix v3.0.12. (BSD 4.2 similar).  I haven't actually tried running the
thing yet; I'll save that for tomorrow.  caveat patchor. :-)

----
Jason Molenda, University of Minnesota Minneapolis, Computer Science dept
molenda@cs.umn.edu  ||  rutgers!umn-cs!molenda