Reduce the memory requirement for large ispell dictionaries.

This patch eliminates per-chunk palloc overhead for most small allocations
needed in the representation of an ispell dictionary.  This saves close to
a factor of 2 on the current Czech ispell data.  While it doesn't cover
every last small allocation in the ispell code, we are at the point of
diminishing returns, because about 95% of the allocations are covered
already.

Pavel Stehule, rather heavily revised by Tom

src/include/tsearch/dicts/spell.h

@@ -161,6 +161,10 @@ typedef struct
 	SPELL	  **Spell;
 	int			nspell;			/* number of valid entries in Spell array */
 	int			mspell;			/* allocated length of Spell array */
+
+	/* These are used to allocate "compact" data without palloc overhead */
+	char	   *firstfree;		/* first free address (always maxaligned) */
+	size_t		avail;			/* free space remaining at firstfree */
 } IspellDict;
 
 extern TSLexeme *NINormalizeWord(IspellDict *Conf, char *word);