Re: [both]: warshall-repair1.patch

[Akim Demaille]

| Hi all,
| 
| This patch should give an idea to solve the problem.
| Just an idea, beause I thing it is not enought.
| 
| Index: ChangeLog
| ===================================================================
| RCS file: /cvsroot/bison/bison/ChangeLog,v
| retrieving revision 1.173.2.223
| diff -u -r1.173.2.223 ChangeLog
| --- ChangeLog 2001/12/05 11:46:05     1.173.2.223
| +++ ChangeLog 2001/12/09 22:17:39
| @@ -1,3 +1,7 @@
| +2001-12-09  Marc Autret  <address@hidden>
| +
| +     * src/warshall.c (TC): Fix.
| +
|  2001-12-05  Akim Demaille  <address@hidden>
|  
|       Version 1.30f.
| Index: src/warshall.c
| ===================================================================
| RCS file: /cvsroot/bison/bison/src/warshall.c,v
| retrieving revision 1.6.2.1
| diff -u -r1.6.2.1 warshall.c
| --- src/warshall.c    2001/12/05 09:23:10     1.6.2.1
| +++ src/warshall.c    2001/12/09 22:17:42
| @@ -38,9 +38,9 @@
|    for (i = 0; i < n; ++i)
|      for (j = 0; j < n; ++j)
|        if (BITISSET (R (i), j))
| -     for (k = 0; k < rowsize; ++k)
| -       if (BITISSET (R (i), k))
| -         SETBIT (R (j), k);
| +     for (k = 0; k < n; ++k)
| +       if (BITISSET (R (j), k))
| +         SETBIT (R (i), k);
|  }

I don't think it is correct.  Another problem was the outer _has_ to
be the transition item, not one of the borders.  But you did spot the
bad guy, congrats!