Re: Warnings triggered by recent includefile cleanups

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
>> Does anyone object if I revert this code to the way it was?

> Considering that evidence shows that limits.h must have been available on
> all platforms at least since 6.5, in fact at least as long as the current
> regex engine has existed, values.h could not possibly have been included
> anywhere ever, so it's probably better to just remove it.

Hmm, it does look like regex has included <limits.h> unconditionally
since day 1, doesn't it? That sure suggests that this patch:

@@ -7,7 +7,7 @@
*
*
* IDENTIFICATION
- * $Header: /home/projects/pgsql/cvsroot/pgsql/src/backend/utils/adt/float.c,v 1.47 1999/07/17 20:17:55 momjian Exp $
+ * $Header: /home/projects/pgsql/cvsroot/pgsql/src/backend/utils/adt/float.c,v 1.48 1999/09/21 20:58:25 momjian Exp $
*
*-------------------------------------------------------------------------
*/
@@ -55,6 +55,13 @@
#include "postgres.h"
#ifdef HAVE_LIMITS_H
#include <limits.h>
+#ifndef MAXINT
+#define MAXINT INT_MAX
+#endif
+#else
+#ifdef HAVE_VALUES_H
+#include <values.h>
+#endif
#endif
#include "fmgr.h"
#include "utils/builtins.h"

was dead code when it was installed. The CVS log says
values.h patch from Alex Howansky
but I can't find anything from him in the mailing list archives.
(We seem to have lost the pgsql-patches archives, however, so if it
was just a patch that went by then there's no remaining doco.)

Bruce, does this ring a bell at all? Unless someone can remember
why this change seemed like a good idea, I think I will take Peter's
advice...

regards, tom lane