diff mbox series

[C++] simplify deferred parsing lexer

Message ID 8529c108-d5de-1eac-fce5-696839ee1fd9@acm.org
State New
Headers show
Series [C++] simplify deferred parsing lexer | expand

Commit Message

Nathan Sidwell Oct. 28, 2019, 12:33 p.m. UTC
We use an eof_token global variable as a sentinel on a deferred parse 
(such as in-class function definitions, or default args).  This 
complicates retrieving the next token in certain places.

As such deferred parses always nest properly and completely before 
resuming the outer lexer, we can simply morph the token after the 
deferred buffer into a CPP_EOF token and restore it afterwards.  I 
finally got around to implementing it with this patch.

One complication is that we have to change the discriminator for when 
the token's value is a tree.  We can't look at the token's type because 
it might have been overwritten.  I add a bool flag to the token (there's 
several spare bits), and use that.  This does simplify the discriminator 
because we just check a single bit, rather than a set of token types.

I think this will allow us to report the location of EOF tokens -- 
currently we think they have no location, because they're (nearly) 
always the shared eof_token.  that's a later change, as it could change 
some diagnostic locations in the testsuite [to be better].

applying to trunk.

nathan
diff mbox series

Patch

2019-10-28  Nathan Sidwell  <nathan@acm.org>

	* parser.h (struct cp_token): Drop {ENUM,BOOL}_BITFIELD C-ism.
	Add tree_check_p flag, use as nested union discriminator.
	(struct cp_lexer): Add saved_type & saved_keyword fields.
	* parser.c (eof_token): Delete.
	(cp_lexer_new_main): Always init last_token to last token of
	buffer.
	(cp_lexer_new_from_tokens): Overlay EOF token at end of range.
	(cp_lexer_destroy): Restore token under the EOF.
	(cp_lexer_previous_token_position): No check for eof_token here.
	(cp_lexer_get_preprocessor_token): Clear tree_check_p.
	(cp_lexer_peek_nth_token): Check CPP_EOF not eof_token.
	(cp_lexer_consume_token): Assert not CPP_EOF, no check for
	eof_token.
	(cp_lexer_purge_token): Likewise.
	(cp_lexer_purge_tokens_after): No check for EOF token.
	(cp_parser_nested_name_specifier, cp_parser_decltype)
	(cp_parser_template_id): Set tree_check_p.

Index: gcc/cp/parser.c
===================================================================
--- gcc/cp/parser.c	(revision 277460)
+++ gcc/cp/parser.c	(working copy)
@@ -53,9 +53,4 @@  along with GCC; see the file COPYING3.
    and c-lex.c) and the C++ parser.  */
 
-static cp_token eof_token =
-{
-  CPP_EOF, RID_MAX, 0, false, false, false, 0, { NULL }
-};
-
 /* The various kinds of non integral constant we encounter. */
 enum non_integral_constant {
@@ -661,10 +656,8 @@  cp_lexer_new_main (void)
     }
 
-  lexer->last_token = lexer->buffer->address ()
+  lexer->next_token = lexer->buffer->address ();
+  lexer->last_token = lexer->next_token
                       + lexer->buffer->length ()
 		      - 1;
-  lexer->next_token = lexer->buffer->length ()
-		      ? lexer->buffer->address ()
-		      : &eof_token;
 
   /* Subsequent preprocessor diagnostics should use compiler
@@ -688,5 +681,12 @@  cp_lexer_new_from_tokens (cp_token_cache
   /* We do not own the buffer.  */
   lexer->buffer = NULL;
-  lexer->next_token = first == last ? &eof_token : first;
+
+  /* Insert an EOF token.  */
+  lexer->saved_type = last->type;
+  lexer->saved_keyword = last->keyword;
+  last->type = CPP_EOF;
+  last->keyword = RID_MAX;
+
+  lexer->next_token = first;
   lexer->last_token = last;
 
@@ -705,5 +705,12 @@  static void
 cp_lexer_destroy (cp_lexer *lexer)
 {
-  vec_free (lexer->buffer);
+  if (lexer->buffer)
+    vec_free (lexer->buffer);
+  else
+    {
+      /* Restore the token we overwrite with EOF.  */
+      lexer->last_token->type = lexer->saved_type;
+      lexer->last_token->keyword = lexer->saved_keyword;
+    }
   lexer->saved_tokens.release ();
   ggc_free (lexer);
@@ -732,6 +739,4 @@  static inline cp_token_position
 cp_lexer_token_position (cp_lexer *lexer, bool previous_p)
 {
-  gcc_assert (!previous_p || lexer->next_token != &eof_token);
-
   return lexer->next_token - previous_p;
 }
@@ -752,8 +757,5 @@  static inline cp_token_position
 cp_lexer_previous_token_position (cp_lexer *lexer)
 {
-  if (lexer->next_token == &eof_token)
-    return lexer->last_token - 1;
-  else
-    return cp_lexer_token_position (lexer, true);
+  return cp_lexer_token_position (lexer, true);
 }
 
@@ -808,4 +810,5 @@  cp_lexer_get_preprocessor_token (cp_lexe
   token->purged_p = false;
   token->error_reported = false;
+  token->tree_check_p = false;
 
   /* On some systems, some header files are surrounded by an
@@ -1083,14 +1086,7 @@  cp_lexer_peek_nth_token (cp_lexer* lexer
   --n;
   token = lexer->next_token;
-  gcc_assert (!n || token != &eof_token);
-  while (n != 0)
+  while (n && token->type != CPP_EOF)
     {
       ++token;
-      if (token == lexer->last_token)
-	{
-	  token = &eof_token;
-	  break;
-	}
-
       if (!token->purged_p)
 	--n;
@@ -1114,16 +1110,10 @@  cp_lexer_consume_token (cp_lexer* lexer)
   cp_token *token = lexer->next_token;
 
-  gcc_assert (token != &eof_token);
   gcc_assert (!lexer->in_pragma || token->type != CPP_PRAGMA_EOL);
 
   do
     {
+      gcc_assert (token->type != CPP_EOF);
       lexer->next_token++;
-      if (lexer->next_token == lexer->last_token)
-	{
-	  lexer->next_token = &eof_token;
-	  break;
-	}
-
     }
   while (lexer->next_token->purged_p);
@@ -1151,5 +1141,5 @@  cp_lexer_purge_token (cp_lexer *lexer)
   cp_token *tok = lexer->next_token;
 
-  gcc_assert (tok != &eof_token);
+  gcc_assert (tok->type != CPP_EOF);
   tok->purged_p = true;
   tok->location = UNKNOWN_LOCATION;
@@ -1158,12 +1148,5 @@  cp_lexer_purge_token (cp_lexer *lexer)
 
   do
-    {
-      tok++;
-      if (tok == lexer->last_token)
-	{
-	  tok = &eof_token;
-	  break;
-	}
-    }
+    tok++;
   while (tok->purged_p);
   lexer->next_token = tok;
@@ -1179,10 +1162,7 @@  cp_lexer_purge_tokens_after (cp_lexer *l
   cp_token *peek = lexer->next_token;
 
-  if (peek == &eof_token)
-    peek = lexer->last_token;
-
   gcc_assert (tok < peek);
 
-  for ( tok += 1; tok != peek; tok += 1)
+  for (tok++; tok != peek; tok++)
     {
       tok->purged_p = true;
@@ -6617,4 +6597,5 @@  cp_parser_nested_name_specifier_opt (cp_
 	 so the memory will not be reclaimed during token replacing below.  */
       token->u.tree_check_value = ggc_cleared_alloc<struct tree_check> ();
+      token->tree_check_p = true;
       token->u.tree_check_value->value = parser->scope;
       token->u.tree_check_value->checks = get_deferred_access_checks ();
@@ -14802,4 +14783,5 @@  cp_parser_decltype (cp_parser *parser)
   start_token->type = CPP_DECLTYPE;
   start_token->u.tree_check_value = ggc_cleared_alloc<struct tree_check> ();
+  start_token->tree_check_p = true;
   start_token->u.tree_check_value->value = expr;
   start_token->u.tree_check_value->checks = get_deferred_access_checks ();
@@ -16589,4 +16571,5 @@  cp_parser_template_id (cp_parser *parser
 	 so the memory will not be reclaimed during token replacing below.  */
       token->u.tree_check_value = ggc_cleared_alloc<struct tree_check> ();
+      token->tree_check_p = true;
       token->u.tree_check_value->value = template_id;
       token->u.tree_check_value->checks = get_deferred_access_checks ();
Index: gcc/cp/parser.h
===================================================================
--- gcc/cp/parser.h	(revision 277460)
+++ gcc/cp/parser.h	(working copy)
@@ -42,21 +42,23 @@  struct GTY(()) tree_check {
 struct GTY (()) cp_token {
   /* The kind of token.  */
-  ENUM_BITFIELD (cpp_ttype) type : 8;
+  enum cpp_ttype type : 8;
   /* If this token is a keyword, this value indicates which keyword.
      Otherwise, this value is RID_MAX.  */
-  ENUM_BITFIELD (rid) keyword : 8;
+  enum rid keyword : 8;
   /* Token flags.  */
   unsigned char flags;
   /* True if this token is from a context where it is implicitly extern "C" */
-  BOOL_BITFIELD implicit_extern_c : 1;
+  bool implicit_extern_c : 1;
   /* True if an error has already been reported for this token, such as a
      CPP_NAME token that is not a keyword (i.e., for which KEYWORD is
      RID_MAX) iff this name was looked up and found to be ambiguous.  */
-  BOOL_BITFIELD error_reported : 1;
+  bool error_reported : 1;
   /* True for a token that has been purged.  If a token is purged,
      it is no longer a valid token and it should be considered
      deleted.  */
-  BOOL_BITFIELD purged_p : 1;
-  /* 5 unused bits.  */
+  bool purged_p : 1;
+  bool tree_check_p : 1;
+  /* 4 unused bits.  */
+
   /* The location at which this token was found.  */
   location_t location;
@@ -64,10 +66,8 @@  struct GTY (()) cp_token {
   union cp_token_value {
     /* Used for compound tokens such as CPP_NESTED_NAME_SPECIFIER.  */
-    struct tree_check* GTY((tag ("1"))) tree_check_value;
+    struct tree_check* GTY((tag ("true"))) tree_check_value;
     /* Use for all other tokens.  */
-    tree GTY((tag ("0"))) value;
-  } GTY((desc ("(%1.type == CPP_TEMPLATE_ID)"
-	       "|| (%1.type == CPP_NESTED_NAME_SPECIFIER)"
-	       "|| (%1.type == CPP_DECLTYPE)"))) u;
+    tree GTY((tag ("false"))) value;
+  } GTY((desc ("%1.tree_check_p"))) u;
 };
 
@@ -100,4 +100,8 @@  struct GTY (()) cp_lexer {
   vec<cp_token_position> GTY ((skip)) saved_tokens;
 
+  /* Saved pieces of end token we replaced with the eof token.  */
+  enum cpp_ttype saved_type : 8;
+  enum rid saved_keyword : 8;
+
   /* The next lexer in a linked list of lexers.  */
   struct cp_lexer *next;