Re: C PATCH to remove unused block of code

[Marek Polacek]

On Thu, Aug 17, 2017 at 02:17:31PM +, Joseph Myers wrote:
> On Thu, 17 Aug 2017, Marek Polacek wrote:
> 
> > I've been itching to remove this code for some time now.  The comment 
> > suggests
> > that the code is actually unused, so I replaced the body of that "else if" 
> > with
> > gcc_unreachable (); and ran regtest/bootstrap and nothing broke, so I 
> > propose
> > to do away with it.
> > 
> > Bootstrapped/regtested on x86_64-linux, ok for trunk?
> 
> OK, with an appropriate change to the comment above the 
> c_parser_postfix_expression function to say that compound literals are not 
> handled here and callers have to call 
> c_parser_postfix_expression_after_paren_type on encountering them.  (I've 
> reviewed all paths to c_parser_postfix_expression in the current parser 
> and don't think any of them can send compound literals to it, because 
> either they are parsing a specific more restricted syntax incompatible 
> with compond literals (OMP cases), are parsing an expression that might be 
> a cast expression so need to parse the (type) first to distinguish, or are 
> parsing a context such as sizeof where a parenthesized type name is 
> allowed as well as a postfix expression and again need to parse the (type) 
> first to distinguish.)

Thanks for verifying.  Here's what I'm going to install:

2017-08-18  Marek Polacek  

* c-parser.c (c_parser_postfix_expression): Remove unused code.  Update
commentary.

diff --git gcc/c/c-parser.c gcc/c/c-parser.c
index 1402ba67204..5c965d4420d 100644
--- gcc/c/c-parser.c
+++ gcc/c/c-parser.c
@@ -7752,7 +7752,8 @@ c_parser_generic_selection (c_parser *parser)
 }
 
 /* Parse a postfix expression (C90 6.3.1-6.3.2, C99 6.5.1-6.5.2,
-   C11 6.5.1-6.5.2).
+   C11 6.5.1-6.5.2).  Compound literals aren't handled here; callers have to
+   call c_parser_postfix_expression_after_paren_type on encountering them.
 
postfix-expression:
  primary-expression
@@ -7792,7 +7793,7 @@ c_parser_generic_selection (c_parser *parser)
  __builtin_types_compatible_p ( type-name , type-name )
  __builtin_complex ( assignment-expression , assignment-expression )
  __builtin_shuffle ( assignment-expression , assignment-expression )
- __builtin_shuffle ( assignment-expression , 
+ __builtin_shuffle ( assignment-expression ,
 assignment-expression ,
 assignment-expression, )
 
@@ -7943,28 +7944,6 @@ c_parser_postfix_expression (c_parser *parser)
  set_c_expr_source_range (, loc, close_loc);
  mark_exp_read (expr.value);
}
-  else if (c_token_starts_typename (c_parser_peek_2nd_token (parser)))
-   {
- /* A compound literal.  ??? Can we actually get here rather
-than going directly to
-c_parser_postfix_expression_after_paren_type from
-elsewhere?  */
- location_t loc;
- struct c_type_name *type_name;
- c_parser_consume_token (parser);
- loc = c_parser_peek_token (parser)->location;
- type_name = c_parser_type_name (parser);
- c_parser_skip_until_found (parser, CPP_CLOSE_PAREN,
-"expected %<)%>");
- if (type_name == NULL)
-   {
- expr.set_error ();
-   }
- else
-   expr = c_parser_postfix_expression_after_paren_type (parser,
-type_name,
-loc);
-   }
   else
{
  /* A parenthesized expression.  */

Marek

Re: C PATCH to remove unused block of code

[Joseph Myers]

On Thu, 17 Aug 2017, Marek Polacek wrote:

> I've been itching to remove this code for some time now.  The comment suggests
> that the code is actually unused, so I replaced the body of that "else if" 
> with
> gcc_unreachable (); and ran regtest/bootstrap and nothing broke, so I propose
> to do away with it.
> 
> Bootstrapped/regtested on x86_64-linux, ok for trunk?

OK, with an appropriate change to the comment above the 
c_parser_postfix_expression function to say that compound literals are not 
handled here and callers have to call 
c_parser_postfix_expression_after_paren_type on encountering them.  (I've 
reviewed all paths to c_parser_postfix_expression in the current parser 
and don't think any of them can send compound literals to it, because 
either they are parsing a specific more restricted syntax incompatible 
with compond literals (OMP cases), are parsing an expression that might be 
a cast expression so need to parse the (type) first to distinguish, or are 
parsing a context such as sizeof where a parenthesized type name is 
allowed as well as a postfix expression and again need to parse the (type) 
first to distinguish.)

-- 
Joseph S. Myers
jos...@codesourcery.com

C PATCH to remove unused block of code

[Marek Polacek]

I've been itching to remove this code for some time now.  The comment suggests
that the code is actually unused, so I replaced the body of that "else if" with
gcc_unreachable (); and ran regtest/bootstrap and nothing broke, so I propose
to do away with it.

Bootstrapped/regtested on x86_64-linux, ok for trunk?

2017-08-17  Marek Polacek  

* c-parser.c (c_parser_postfix_expression): Remove unused code.

diff --git gcc/c/c-parser.c gcc/c/c-parser.c
index 1402ba67204..8511a8b4fe7 100644
--- gcc/c/c-parser.c
+++ gcc/c/c-parser.c
@@ -7943,28 +7943,6 @@ c_parser_postfix_expression (c_parser *parser)
  set_c_expr_source_range (, loc, close_loc);
  mark_exp_read (expr.value);
}
-  else if (c_token_starts_typename (c_parser_peek_2nd_token (parser)))
-   {
- /* A compound literal.  ??? Can we actually get here rather
-than going directly to
-c_parser_postfix_expression_after_paren_type from
-elsewhere?  */
- location_t loc;
- struct c_type_name *type_name;
- c_parser_consume_token (parser);
- loc = c_parser_peek_token (parser)->location;
- type_name = c_parser_type_name (parser);
- c_parser_skip_until_found (parser, CPP_CLOSE_PAREN,
-"expected %<)%>");
- if (type_name == NULL)
-   {
- expr.set_error ();
-   }
- else
-   expr = c_parser_postfix_expression_after_paren_type (parser,
-type_name,
-loc);
-   }
   else
{
  /* A parenthesized expression.  */

Marek