Re: Bison 1.30d

[Akim Demaille]

Hi Hans,

Please, never drop the public lists.

| At 20:17 +0100 2001/11/23, Akim Demaille wrote:
| >I thought you had an access to the CVS repo.
| >
| >        http://www.lrde.epita.fr/~akim/download/bison-1.30e.tar.gz
| 
| Nope, only the daily CVS digests, and the ftp versions you create.
| 
| I got the following error:
| Error   : cannot convert
| 'unsigned long *' to
| 'int *'
| LR0.c line 97   kernel_size = XCALLOC (size_t, nsyms);
| 
| If I change to
|   static size_t *kernel_size = NULL;
| at the beginning of this file, the error goes away.

Nice shot, thanks, I'm addressing this now.

| As for the rule line numbers, I still get a difference between 1.30c and 
1.30e:
| 
| 1.30c:
| #if YYDEBUG != 0
| /* YYRLINE[YYN] -- source line where rule number YYN was defined. */
| static const short yyrline[] =
| {
|        0,    80,    82,    84,    86,    87,    88,    89,    90,    91,
|       97,   104,   107,   109,   114,   116,   118,   123,   128,   134,
|      136,   137,   144,   145,   152,   160,   162,   163,   172,   173,
|      176,   177,   187,   196,   197,   206,   207,   210,   213,   214,
|      217,   220,   222,   224,   233,   242,   245
| };
| #endif
| 
| 1.30e:
| #if YYDEBUG != 0
| /* YYRLINE[YYN] -- source line where rule number YYN was defined. */
| static const short yyrline[] =
| {
|        0,    80,    82,    84,    86,    87,    88,    89,    90,    91,
|       97,   104,   107,   109,   114,   116,   118,   123,   128,   134,
|      136,   137,   144,   145,   152,   162,   160,   172,   163,   176,
|      173,   177,   196,   187,   206,   197,   210,   213,   207,   217,
|      220,   214,   222,   233,   224,   245,   242
| };
| #endif
| 
| I do not know which one is the correct version. Funny things start to
| happen in the 1.30e version though at: 162, 160,... which seems wrong,
| followed by other strange numbers it seems. So 1.30e is probably still
| wrongo.

Hm... Indeed.  The culprits are the empty rules inserted by embedded
actions...  For the other rules, 1.30e is, IMO, ``more'' correct than
previous versions were.

I'll see if I can do something about these rules.