master: improve accuracy of locations in error messages about nonterminals

[Akim Demaille]

commit 875ef1b90c4cbdc637b6944d4d6eb5b2b7bb7eef
Author: Akim Demaille <address@hidden>
Date:   Wed Jan 14 13:41:32 2015 +0100

    symbol: use the first occurrence as an LHS as defining location
    
    Currently on the following grammar:
    
        %type <foo> foo
        %%
        start: foo | bar | "baz"
        foo: foo
        bar: bar
    
    bison reports:
    
        warning: 2 nonterminals useless in grammar [-Wother]
        warning: 4 rules useless in grammar [-Wother]
        1.13-15: warning: nonterminal useless in grammar: foo [-Wother]
         %type <foo> foo
                     ^^^
        3.14-16: warning: nonterminal useless in grammar: bar [-Wother]
         start: foo | bar | "baz"
                      ^^^
        [...]
    
    i.e., the location of the first occurrence of a symbol is taken as its
    definition point.  In the case of nonterminals, the first occurrence
    as a left-hand side of a rule makes more sense:
    
        warning: 2 nonterminals useless in grammar [-Wother]
        warning: 4 rules useless in grammar [-Wother]
        4.1-3: warning: nonterminal useless in grammar: foo [-Wother]
         foo: foo
         ^^^
        5.1-3: warning: nonterminal useless in grammar: bar [-Wother]
         bar: bar
         ^^^
        [...]
    
    * src/symtab.h, src/symtab.c (symbol::location_of_lhs): New.
    (symbol_location_as_lhs_set): New.
    * src/parse-gram.y (current_lhs): Use it.
    * tests/reduce.at: Update locations.

diff --git a/NEWS b/NEWS
index e64c6e8..bba0f5f 100644
--- a/NEWS
+++ b/NEWS
@@ -33,13 +33,14 @@ GNU Bison NEWS
                                       ^^^^^^^^^^^
 
   Now, rules whose left-hand side symbol is useless are no longer reported
-  as useless.
+  as useless.  The locations of the errors have also been adjusted to point
+  to the first use of the nonterminal as a left-hand side of a rule:
 
     warning: 1 nonterminal useless in grammar [-Wother]
     warning: 4 rules useless in grammar [-Wother]
-    2.14-16: warning: nonterminal useless in grammar: exp [-Wother]
-     input: '0' | exp
-                  ^^^
+    3.1-3: warning: nonterminal useless in grammar: exp [-Wother]
+     exp: exp '+' exp | exp '-' exp | '(' exp ')'
+     ^^^
     2.14-16: warning: rule useless in grammar [-Wother]
      input: '0' | exp
                   ^^^
diff --git a/src/parse-gram.c b/src/parse-gram.c
index 825e2b0..43eeaeb 100644
--- a/src/parse-gram.c
+++ b/src/parse-gram.c
@@ -1,4 +1,4 @@
-/* A Bison parser, made by GNU Bison 3.0.2.13-975bb-dirty.  */
+/* A Bison parser, made by GNU Bison 3.0.2.29-9a91e.  */
 
 /* Bison implementation for Yacc-like parsers in C
 
@@ -44,7 +44,7 @@
 #define YYBISON 1
 
 /* Bison version.  */
-#define YYBISON_VERSION "3.0.2.13-975bb-dirty"
+#define YYBISON_VERSION "3.0.2.29-9a91e"
 
 /* Skeleton name.  */
 #define YYSKELETON_NAME "yacc.c"
@@ -3118,6 +3118,8 @@ current_lhs (symbol *sym, location loc, named_ref *ref)
 {
   current_lhs_symbol = sym;
   current_lhs_location = loc;
+  if (sym)
+    symbol_location_as_lhs_set (sym, loc);
   /* In order to simplify memory management, named references for lhs
      are always assigned by deep copy into the current symbol_list
      node.  This is because a single named-ref in the grammar may
diff --git a/src/parse-gram.h b/src/parse-gram.h
index c40151d..26bb3c5 100644
--- a/src/parse-gram.h
+++ b/src/parse-gram.h
@@ -1,4 +1,4 @@
-/* A Bison parser, made by GNU Bison 3.0.2.13-975bb-dirty.  */
+/* A Bison parser, made by GNU Bison 3.0.2.29-9a91e.  */
 
 /* Bison interface for Yacc-like parsers in C
 
diff --git a/src/parse-gram.y b/src/parse-gram.y
index 9184d5f..5a0b200 100644
--- a/src/parse-gram.y
+++ b/src/parse-gram.y
@@ -865,6 +865,8 @@ current_lhs (symbol *sym, location loc, named_ref *ref)
 {
   current_lhs_symbol = sym;
   current_lhs_location = loc;
+  if (sym)
+    symbol_location_as_lhs_set (sym, loc);
   /* In order to simplify memory management, named references for lhs
      are always assigned by deep copy into the current symbol_list
      node.  This is because a single named-ref in the grammar may
diff --git a/src/symtab.c b/src/symtab.c
index 02b3eae..9524773 100644
--- a/src/symtab.c
+++ b/src/symtab.c
@@ -105,6 +105,7 @@ symbol_new (uniqstr tag, location loc)
 
   res->tag = tag;
   res->location = loc;
+  res->location_of_lhs = false;
   res->alias = NULL;
   res->content = sym_content_new (res);
   res->is_alias = false;
@@ -189,6 +190,7 @@ code_props_type_string (code_props_type kind)
   assert (0);
 }
 
+
 /*----------------------------------------.
 | Create a new semantic type, named TAG.  |
 `----------------------------------------*/
@@ -307,6 +309,13 @@ semantic_type_redeclaration (semantic_type *s, const char 
*what, location first,
 }
 
 
+void
+symbol_location_as_lhs_set (symbol *sym, location loc)
+{
+  if (!sym->location_of_lhs)
+    sym->location = loc;
+}
+
 
 /*-----------------------------------------------------------------.
 | Set the TYPE_NAME associated with SYM.  Does nothing if passed 0 |
diff --git a/src/symtab.h b/src/symtab.h
index 9719cd3..42d479c 100644
--- a/src/symtab.h
+++ b/src/symtab.h
@@ -87,10 +87,19 @@ struct symbol
 {
   /** The key, name of the symbol.  */
   uniqstr tag;
-  /** The location of its first occurrence.  */
+
+  /** The "defining" location.  */
   location location;
 
-  /* Points to the other in the symbol-string pair for an alias. */
+  /** Whether \a location is about the first uses as left-hand side
+      symbol of a rule (true), or simply the first occurrence (e.g.,
+      in a %type, or as a rhs symbol of a rule).  The former type of
+      location is more natural in error messages.  This Boolean helps
+      moving from location of the first occurrence to first use as
+      lhs. */
+  bool location_of_lhs;
+
+  /** Points to the other in the symbol-string pair for an alias. */
   symbol *alias;
 
   /** Whether this symbol is the alias of another or not. */
@@ -177,6 +186,12 @@ uniqstr symbol_id_get (symbol const *sym);
  */
 void symbol_make_alias (symbol *sym, symbol *str, location loc);
 
+/**
+ * This symbol is used as the lhs of a rule.  Record this location
+ * as definition point, if not already done.
+ */
+void symbol_location_as_lhs_set (symbol *sym, location loc);
+
 /** Set the \c type_name associated with \c sym.
 
     Do nothing if passed 0 as \c type_name.  */
diff --git a/tests/reduce.at b/tests/reduce.at
index c7edb10..d197dc6 100644
--- a/tests/reduce.at
+++ b/tests/reduce.at
@@ -254,9 +254,9 @@ not-reduced.y: warning: 3 rules useless in grammar [-Wother]
 not-reduced.y:14.1-13: warning: nonterminal useless in grammar: not_reachable 
[-Wother]
  not_reachable: useful  { /* A not reachable action. */ }
  ^^^^^^^^^^^^^
-not-reduced.y:11.6-19: warning: nonterminal useless in grammar: non_productive 
[-Wother]
-    | non_productive    { /* A non productive action. */ }
-      ^^^^^^^^^^^^^^
+not-reduced.y:17.1-14: warning: nonterminal useless in grammar: non_productive 
[-Wother]
+ non_productive: non_productive useless_token
+ ^^^^^^^^^^^^^^
 not-reduced.y:11.6-57: warning: rule useless in grammar [-Wother]
     | non_productive    { /* A non productive action. */ }
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
@@ -330,12 +330,12 @@ indirection: underivable;
 AT_BISON_CHECK([[-fcaret input.y]], 0, [],
 [[input.y: warning: 2 nonterminals useless in grammar [-Wother]
 input.y: warning: 3 rules useless in grammar [-Wother]
-input.y:5.15-25: warning: nonterminal useless in grammar: underivable [-Wother]
- exp: useful | underivable;
-               ^^^^^^^^^^^
-input.y:6.14-24: warning: nonterminal useless in grammar: indirection [-Wother]
+input.y:6.1-11: warning: nonterminal useless in grammar: underivable [-Wother]
  underivable: indirection;
-              ^^^^^^^^^^^
+ ^^^^^^^^^^^
+input.y:7.1-11: warning: nonterminal useless in grammar: indirection [-Wother]
+ indirection: underivable;
+ ^^^^^^^^^^^
 input.y:5.15-25: warning: rule useless in grammar [-Wother]
  exp: useful | underivable;
                ^^^^^^^^^^^