Canonicalize conversion operators

  Consider a conversion operator such as:

operator foo const* const* ();

There are two small parser problems, highlighted by this test:

(gdb) p operator foo const* const*
There is no field named operatorfoo const* const *

GDB is looking up the symbol "operatorfoo const* const*" -- it is missing a
space between the keyword "operator" and the type name "foo const* const*".
Pedro also discovered this problem and has a patch for it pending on his
cxx-breakpoint-improvements branch.

However, something not address by Pedro's patch is that this input of the
user-defined type needs to be canonicalized so that different "spellings" of
the type are recognized:

(gdb) p operator const foo* const *
There is no field named operator const foo* const *

gdb/ChangeLog:

	* c-exp.y (oper): Canonicalize conversion operators of user-defined
	types.
	Add whitespace to front of type name.

gdb/testsuite/ChangeLog:

	* gdb.cp/cpexprs.cc (base) <operator fluff const* const*>: New
	method.
	(main): Call it.
	* gdb.cp/cpexprs.exp: Add new conversion operator to test matrix.
	Add additional user-defined conversion operator tests.
---
 gdb/c-exp.y                      | 10 ++++++++--
 gdb/testsuite/gdb.cp/cpexprs.cc  |  2 ++
 gdb/testsuite/gdb.cp/cpexprs.exp | 16 ++++++++++++++++
 3 files changed, 26 insertions(+), 2 deletions(-)
  

On 10/17/2017 11:12 PM, Keith Seitz wrote:
> Consider a conversion operator such as:
> 
> operator foo const* const* ();
> 
> There are two small parser problems, highlighted by this test:
> 
> (gdb) p operator foo const* const*
> There is no field named operatorfoo const* const *
> 
> GDB is looking up the symbol "operatorfoo const* const*" -- it is missing a
> space between the keyword "operator" and the type name "foo const* const*".
> Pedro also discovered this problem and has a patch for it pending on his
> cxx-breakpoint-improvements branch.
> 
> However, something not address by Pedro's patch is that this input of the
> user-defined type needs to be canonicalized so that different "spellings" of
> the type are recognized:
> 
> (gdb) p operator const foo* const *
> There is no field named operator const foo* const *
> 

Indeed.


> @@ -1630,7 +1630,13 @@ oper:	OPERATOR NEW
>  
>  			  c_print_type ($2, NULL, &buf, -1, 0,
>  					&type_print_raw_options);
> -			  $$ = operator_stoken (buf.c_str ());
> +
> +			  /* This also needs canonicalization.  */
> +			  std::string canon
> +			    = " " + cp_canonicalize_string (buf.c_str ());
> +			  if (canon.length () == 1)
> +			    canon = " " + buf.string ();
> +			  $$ = operator_stoken (canon.c_str ());

The length() == 1 check gave me pause.  It's checking that
cp_canonicalize_string returned empty of course, but it
wasn't super clear on a quick skim.

I think you could write it like this, making that part clearer,
and also saving a few string dups and copies:

			  /* This also needs canonicalization.  */
			  std::string canon
			    = cp_canonicalize_string (buf.c_str ());
			  if (canon.empty ())
			    canon = std::move (buf.string ());
			  $$ = operator_stoken ((" " + canon).c_str ());


>  
> +# Make sure conversion operator names are canonicalized and properly
> +# "spelled."

Period should outside the quote, I think.

Otherwise OK.

Thanks,
Pedro Alves
  

On 10/18/2017 04:19 AM, Pedro Alves wrote:
>> @@ -1630,7 +1630,13 @@ oper:	OPERATOR NEW
>>  
>>  			  c_print_type ($2, NULL, &buf, -1, 0,
>>  					&type_print_raw_options);
>> -			  $$ = operator_stoken (buf.c_str ());
>> +
>> +			  /* This also needs canonicalization.  */
>> +			  std::string canon
>> +			    = " " + cp_canonicalize_string (buf.c_str ());
>> +			  if (canon.length () == 1)
>> +			    canon = " " + buf.string ();
>> +			  $$ = operator_stoken (canon.c_str ());
> 
> The length() == 1 check gave me pause.  It's checking that
> cp_canonicalize_string returned empty of course, but it
> wasn't super clear on a quick skim.
> 
> I think you could write it like this, making that part clearer,
> and also saving a few string dups and copies:
> 
> 			  /* This also needs canonicalization.  */
> 			  std::string canon
> 			    = cp_canonicalize_string (buf.c_str ());
> 			  if (canon.empty ())
> 			    canon = std::move (buf.string ());
> 			  $$ = operator_stoken ((" " + canon).c_str ());
> 
> 

Yes, indeed. ISTR having something similar to that at one point, but I don't know why I changed it. [Maybe that was during the C++03/C++11 shake out?]

In any case, committed with that change.

Thank you for the review.

Keith