Small fixes (no. 2)

ChangeLog

@@ -845,7 +845,7 @@ Version 4.6.5 (3 November 2009)
  * Fix: Take more care in distinguishing mass and count nouns.
  * Fix: Old bug w/relative clauses: Rw+ is optional, not mandatory.
  * Provide tags identifying relative, superlative adjectives.
- * Remove BioLG NUMBER-AND-UNIT handling, its been superseded.
+ * Remove BioLG NUMBER-AND-UNIT handling, it's been superseded.
  * Fix handling of parenthetical phrases/clauses.
  * Fix: handling of capitalized first words ending in letter "s".
  * Fix: support "filler-it" SF link for "It was reasoned that..."

autogen.sh

@@ -8,8 +8,6 @@ if [ ! -f "autogen.sh" ]; then
     exit 1
 fi
 
-rm -f autogen.err
-
 run_configure=true
 while [ -n "$1" ]
 do
@@ -28,7 +26,8 @@ do
             shift
             ;;
         *)
-            break 2
+            echo "$0: Error: Unknown flag \"$1\"."
+            exit 1
     esac
 done
 
@@ -40,6 +39,8 @@ if [ $? -ne 0 ]; then
     exit 1
 fi
 
+rm -f autogen.err
+
 # If there's a config.cache file, we may need to delete it.
 # If we have an existing configure script, save a copy for comparison.
 # (Based on Subversion's autogen.sh, see also the deletion code below.)

bindings/python-examples/tests.py

@@ -490,7 +490,7 @@ def test_21_set_error_handler_None(self):
         self.numerr = LG_Error.printall(self.error_handler_test, None)
         self.assertEqual(self.numerr, self.numerr)
 
-    def test_22_defaut_handler_param(self):
+    def test_22_default_handler_param(self):
         """Test bad data parameter to default error handler"""
         # (It should be an integer >=0 and <= lg_None.)
         # Here the error handler is still set to None.

configure.ac

@@ -156,7 +156,7 @@ AC_SUBST(HOST_OS)
 
 # ====================================================================
 # FreeBSD work-around. Apparently, the AX_PTHREAD autoconf macro
-# fails to include -lstdthreads in it's results. See bug report
+# fails to include -lstdthreads in its results. See bug report
 # https://github.com/opencog/link-grammar/issues/1355
 # So we hack, and explicitly set it here.
 
@@ -495,7 +495,7 @@ AC_SUBST(SQLITE3_CFLAGS)
 # The AtomSpace dictionary backend
 # TODO - use pkg_config for find the libs.
 # TODO - do NOT specify -lpersist-rocks -lpersist-cog  here,
-#        they need to be dynmically loaded instead (for proper
+#        they need to be dynamically loaded instead (for proper
 #        initialization of the shared library ctors these contain).
 #        Alternately, if they are found, then wrap them with
 #        "-Wl,--no-as-needed"
@@ -930,7 +930,7 @@ fi
 
 # ===================================================================
 # swig is needed for compiling the Perl and Python bindings ...
-# ... well, actually, no, its not. 'make dist' is currently set up to
+# ... well, actually, no, it's not. 'make dist' is currently set up to
 # package all of the files generated by swig, so the user does not need
 # to actually install it.  However, swig is needed to create the package,
 # and also needed to build from a GitHub pull.

link-grammar/README.md

@@ -31,7 +31,7 @@ Listed in rough processing order.
 Version 5.3.14 - Improved error notification facility
 =====================================================
 
-This code is still "experimental", so it's API may be changed.
+This code is still "experimental", so its API may be changed.
 
 It is intended to be mostly compatible. It supports multi-threading -
 all of its operations are local per-thread.

link-grammar/api-structures.h

@@ -24,7 +24,7 @@
  *
  * To make the API simpler, each of these is typedef'ed as a pointer
  * to a data structure.  If you're not used to this, some of the code
- * may look strange, since its not plain that these types are pointers.
+ * may look strange, since it's not plain that these types are pointers.
  *
  *********************************************************************/
 

link-grammar/dict-atomese/README.md

@@ -151,7 +151,7 @@ same way, differing only in how the AtomSpace is managed.
 
 Private Mode
 ------------
-In **private mode**, the system creates and maintains it's own private
+In **private mode**, the system creates and maintains its own private
 AtomSpace, and relies on being able to access a `StorageNode` from
 which appropriate language data can be fetched. This `StorageNode`
 must be configured in the `storage.dict` file. This `StorageNode` can
@@ -202,7 +202,7 @@ Both modes work as follows:
   will be mapped, one-to-one, to LG disjuncts, and used for parsing.
   If parsing fails, and supplemental word-pairs is enabled, then
   disjuncts will be decorated with additional word-pairs, hoping to
-  obtain a parse. If word-pairs are not avilable, then disjuncts can
+  obtain a parse. If word-pairs are not available, then disjuncts can
   be supplemented with random `ANY` link types (which connect to any
   other `ANY` connector, i.e. randomly.)
 * If there are no `Sections`, or if these are disabled, then word-pairs

link-grammar/dict-atomese/lookup-atomese.cc

@@ -644,7 +644,7 @@ void and_enchain_right(Pool_desc* pool, Exp* &andhead, Exp* &andtail, Exp* item)
 static void report_dict_usage(Dictionary dict)
 {
 	// We could also print pool_num_elements_issued() but this is not
-	// interesting; its slightly less than pool_size().
+	// interesting; it's slightly less than pool_size().
 	logger().info("LG Dict: %lu entries; %lu Exp_pool elts; %lu MiBytes",
 		dict->num_entries,
 		pool_size(dict->Exp_pool),

link-grammar/dict-atomese/sections.cc

@@ -351,7 +351,7 @@ Dict_node * lookup_section(Dictionary dict, const Handle& germ)
 	Exp* andtail = nullptr;
 
 	// We really expect dn to not be null here, but ... perhaps it
-	// is, if its just not been observed before.
+	// is, if it's just not been observed before.
 	Exp* eee = nullptr;
 	if (dn)
 	{

link-grammar/dict-atomese/word-pairs.cc

@@ -260,7 +260,7 @@ static Exp* make_pair_exprs(Dictionary dict, const Handle& germ)
 		// Get the cached link-name for this pair.
 		const std::string& slnk = cached_linkname(local, rawpr);
 
-		// Direction is easy to determine: its either left or right.
+		// Direction is easy to determine: it's either left or right.
 		char cdir = '+';
 		if (rawpr->getOutgoingAtom(1) == germ) cdir  = '-';
 
@@ -446,7 +446,7 @@ static Exp* make_any_conns(Dictionary dict, Pool_desc* pool)
 /// and arity is 3, then this will return `(A+ or B- or C+ or ())
 /// and (A+ or B- or C+ or ()) and (A+ or B- or C+ or ())`. When
 /// this is exploded into disjuncts, any combination is possible,
-/// from size zero to three. That's why its a Cartesian product.
+/// from size zero to three. That's why it's a Cartesian product.
 ///
 /// FYI, this is a work-around for the lack of a commmutative
 /// multi-product. What we really want to do here is to have the
@@ -497,7 +497,7 @@ Exp* make_cart_pairs(Dictionary dict, const Handle& germ,
 
 	Exp* optex = make_optional_node(pool, epr);
 
-	// If its 1-dimensional, we are done.
+	// If it's 1-dimensional, we are done.
 	if (1 == arity) return optex;
 
 	Exp* andhead = nullptr;

link-grammar/dict-common/dict-common.h

@@ -151,7 +151,7 @@ struct Dictionary_s
 	/* Duplicate words are disallowed in 4.0.dict unless
 	 * allow_duplicate_words is defined to "true".
 	 * Duplicate idioms are allowed, unless the "test" parse option
-	 * is set to "disalow-dup-idioms" (listing them for debug).
+	 * is set to "disallow-dup-idioms" (listing them for debug).
 	 * If these variables are 0, they get their allow/disallow values
 	 * when the first duplicate word/idiom is encountered.
 	 * 0: not set; 1: allow; -1: disallow */

link-grammar/dict-common/dict-utils.c

@@ -36,7 +36,7 @@ void patch_subscript(char * s)
 	if (*de == '\0') return;
 	dp = (int) *de;
 
-	/* If it's followed by a UTF8 char, its NOT a subscript */
+	/* If it's followed by a UTF8 char, it's NOT a subscript */
 	if (127 < dp || dp < 0) return;
 	/* assert ((0 < dp) && (dp <= 127), "Bad dictionary entry!"); */
 	if (isdigit(dp)) return;

link-grammar/dict-common/idiom.c

@@ -223,7 +223,7 @@ void insert_idiom(Dictionary dict, Dict_node * dn)
 
 	/* ---- end of the code alluded to above ---- */
 
-	/* now its time to insert them into the dictionary */
+	/* now it's time to insert them into the dictionary */
 
 	dn_list = start_dn_list;
 

link-grammar/dict-ram/dict-ram.c

@@ -69,7 +69,7 @@ void free_dictionary_root(Dictionary dict)
 /**
  * dict_order_strict - order two dictionary words in proper sort order.
  * Return zero if the strings match, else return in a unique order.
- * The order is NOT (locale-dependent) UTF8 sort order; its ordered
+ * The order is NOT (locale-dependent) UTF8 sort order; it's ordered
  * based on numeric values of single bytes.  This will uniquely order
  * UTF8 strings, just not in a LANG-dependent (locale-dependent) order.
  * But we don't need/want locale-dependent ordering!
@@ -309,7 +309,7 @@ Dict_node * dict_node_wild_lookup(Dictionary dict, const char *s)
  * The in-RAM representation is NOT a binary tree; instead it is a tree
  * of lists. An Exp node contains a tag: `AND_type`, `OR_type`, etc.
  * The `operand_next` field is used to hold a linked list which is
- * joined up with the given Exp type. This is more efficent to traverse,
+ * joined up with the given Exp type. This is more efficient to traverse,
  * and also saves space, as compared to an ordinary binary tree.
  *
  * Example:  (A or B or C or D) becomes

link-grammar/dict-sql/read-sql.c

@@ -66,7 +66,7 @@ static const char * make_expression(Dictionary dict,
 	while (*p && (lg_isspace((unsigned char)*p))) p++;
 	if (0 == *p) return p;
 
-	/* If it's an open paren, assume its the beginning of a new list */
+	/* If it's an open paren, assume it's the beginning of a new list */
 	if ('(' == *p)
 	{
 		p = make_expression(dict, ++p, pex);

link-grammar/parse/count.c

@@ -66,7 +66,7 @@ struct Table_tracon_s
  * Each element of these arrays points to a vector, called here word-vector,
  * with a size equal to abs(nearest_word - farthest_word) + 1.
  *
- * Each element of this vector predicts, in it's status field, whether the
+ * Each element of this vector predicts, in its status field, whether the
  * expected count is zero for a given null-count (when the word it
  * represents is the end word of the range).
  * This prediction is valid for null-counts up to null_count (or for any
@@ -164,7 +164,7 @@ static size_t estimate_tracon_entries(Sentence sent)
 }
 
 #if HAVE_THREADS_H && !__EMSCRIPTEN__
-/* Each thread will get it's own version of the `kept_table`.
+/* Each thread will get its own version of the `kept_table`.
  * If the program creates zillions of threads, then there will
  * be a mem-leak if this table is not released when each thread
  * exits. This code arranges so that `free_tls_table` is called

link-grammar/parse/extract-links.c

@@ -851,7 +851,7 @@ static void issue_links_for_choice(Linkage lkg, Parse_choice *pc,
  *
  * How it works:
  *
- * Each linkage has the abstact form of a binary tree, with left and
+ * Each linkage has the abstract form of a binary tree, with left and
  * right subtrees.  The Parse_set is an encoding for all possible
  * trees. Selecting a linkage is then a matter of selecting  tree from
  * out of the parse-set.

link-grammar/parse/fast-match.h

@@ -20,7 +20,7 @@
 #include "link-includes.h"              // for Sentence
 #include "memory-pool.h"
 
-typedef struct match_list_cache_sruct
+typedef struct
 {
 	Disjunct *d;                 /* disjuncts with a jet linkage */
 	Count_bin count;             /* the counts for that linkage */

link-grammar/parse/parse.c

@@ -354,7 +354,7 @@ static int linkage_equiv_p(Linkage lpv, Linkage lnx)
 	// names were the same. The connector types might still differ,
 	// due to intersection. The multi-connector flag might differ.
 	// However, neither of these are likely. It is plausible to skip
-	// this check entirely, its mostly a CPU-time-waster that will
+	// this check entirely, it's mostly a CPU-time-waster that will
 	// never find any differences for the almost any situation.
 	for (uint32_t li=0; li<lpv->num_links; li++)
 	{
@@ -460,7 +460,7 @@ static void deduplicate_linkages(Sentence sent, int linkage_limit)
 	    !sent->overflowed && (sent->num_linkages_found <= linkage_limit)))
 		return;
 
-	// Deduplicate the valid linkages only; its not worth wasting
+	// Deduplicate the valid linkages only; it's not worth wasting
 	// CPU time on the rest.  Sorting guarantees that the valid
 	// linkages come first.
 	uint32_t nl = sent->num_valid_linkages;

link-grammar/parse/preparation.c

@@ -121,7 +121,7 @@ static void build_sentence_disjuncts(Sentence sent, float cost_cutoff,
 #ifdef DEBUG
 	unsigned int dcnt, ccnt;
 	count_disjuncts_and_connectors(sent, &dcnt, &ccnt);
-	lgdebug(+D_PREP, "%u disjucts, %u connectors (%zu allocated)\n",
+	lgdebug(+D_PREP, "%u disjuncts, %u connectors (%zu allocated)\n",
 	        dcnt, ccnt,
 	        pool_num_elements_issued(sent->Connector_pool) - num_con_alloced);
 #endif

link-grammar/parse/prune.c

@@ -44,7 +44,7 @@
 #define PRx(x) fprintf(stderr, ""#x)
 #define PR(...) true
 
-/* Indicator that this connector cannot be used -- that its "obsolete".  */
+/* Indicator that this connector cannot be used -- that it's "obsolete".  */
 #define BAD_WORD (MAX_SENTENCE+1)
 
 typedef uint8_t WordIdx_m;     /* Storage representation of word index */
@@ -1321,9 +1321,9 @@ static unsigned int cms_hash(const char *s)
 	return (i & (CMS_SIZE-1));
 }
 
-static void reset_last_criterion(multiset_table *cmt, const char *ctiterion)
+static void reset_last_criterion(multiset_table *cmt, const char *criterion)
 {
-	unsigned int h = cms_hash(ctiterion);
+	unsigned int h = cms_hash(criterion);
 
 	for (Cms *cms = cmt->cms_table[h]; cms != NULL; cms = cms->next)
 		cms->last_criterion = false;

link-grammar/prepare/exprune.c

@@ -185,7 +185,7 @@ static inline bool matches_S(connector_table **ct, int w, condesc_t * c)
  * doesn't match anything in the set S.
  *
  * If an OR or AND type expression node has one child, we can replace it
- * by it's child.  This, of course, is not really necessary, except for
+ * by its child.  This, of course, is not really necessary, except for
  * performance.
  */
 

link-grammar/tokenize/tokenize.c

@@ -1444,7 +1444,7 @@ static bool mprefix_split(Sentence sent, Gword *unsplit_word, const char *word)
 	mprefix_list = AFCLASS(dict->affix_table, AFDICT_MPRE);
 	mp_strippable = mprefix_list->length;
 	if (0 == mp_strippable) return false;
-	/* The mprefix list is revered-sorted according to prefix length.
+	/* The mprefix list is reversed-sorted according to prefix length.
 	 * The code here depends on that. */
 	mprefix = mprefix_list->string;
 
@@ -1539,11 +1539,11 @@ static bool mprefix_split(Sentence sent, Gword *unsplit_word, const char *word)
 }
 
 /* Return true if the word might be capitalized by convention:
- * -- if its the first word of a sentence
- * -- if its the first word following a colon, a period, a question mark,
+ * -- if it's the first word of a sentence
+ * -- if it's the first word following a colon, a period, a question mark,
  *    or any bullet (For example:  VII. Ancient Rome)
- * -- if its the first word following an ellipsis
- * -- if its the first word of a quote
+ * -- if it's the first word following an ellipsis
+ * -- if it's the first word of a quote
  *
  * XXX FIXME: These rules are rather English-centric.  Someone should
  * do something about this someday.
@@ -2270,7 +2270,7 @@ static void issue_r_stripped(Sentence sent,
 		if (NULL != r_stripped[1][i])
 		{
 			/* We are going to issue a subscripted word which is not a
-			 * substring of it's unsplit_word. For now, the token position
+			 * substring of its unsplit_word. For now, the token position
 			 * computation code needs an indication for that. */
 			replabel = strdupa(label);
 			replabel[0] = REPLACEMENT_MARK[0];

link-grammar/tracon-set.c

@@ -64,7 +64,7 @@ static tid_hash_t hash_connectors(const Connector *c, unsigned int shallow)
  * cache/swap trashing if the table temporary grows very big. However, it
  * had a bug, and it is not clear when to shrink the table - shrinking it
  * unnecessarily can cause an overhead of a table growth. Keep for
- * possible reimlementation of a similar idea.
+ * possible reimplementation of a similar idea.
  */
 static unsigned int find_prime_for(size_t count)
 {

link-parser/lg_readline.c

@@ -49,7 +49,7 @@ static wchar_t * prompt(EditLine *el)
 	return wc_prompt;
 }
 
-// lg_readline()is called via a chain of functions:
+// lg_readline() is called via a chain of functions:
 // fget_input_string -> get_line -> get_terminal_line -> lg_readline.
 // To avoid changing all of them, this variable is static for now.
 // FIXME: Move the call of find_history_filepath() to lg_readline(), and
@@ -82,7 +82,7 @@ void find_history_filepath(const char *dictname, const char *argv0,
 	{
 		prt_error("Warning: xdg_get_home(XDG_BD_STATE) failed; "
 		          "input history will not be supported.\n");
-		history_file = strdup("dev/null");
+		history_file = strdup("/dev/null");
 	}
 
 	if (get_verbosity() == D_USER_FILES)