Re: [Patch] Add TYPE_NAME_IDENTIFIER

[Paolo Carlini <paolo dot carlini at oracle dot com>]

Hi,

On 04/14/2014 06:20 AM, Jason Merrill wrote:
> On 04/13/2014 12:20 PM, Paolo Carlini wrote:
> > this is a small clean-up discussed some time ago with Jason. The basic
> > idea seems pretty straightforward to me. However, at variance with that
> > discussion, I'm not proposing replacing the C++ front-end
> > TYPE_IDENTIFIER with the new TYPE_NAME_IDENTIFIER because, as far as I
> > can see, that front-end really needs DECL_NAME (TYPE_NAME
> > unconditionally. Eg, code like (in error.c):
> >
> >      case TEMPLATE_TEMPLATE_PARM:
> >        /* For parameters inside template signature.  */
> >        if (TYPE_IDENTIFIER (t))
> >      pp_cxx_tree_identifier (pp, TYPE_IDENTIFIER (t));
> >        else
> >      pp_cxx_canonical_template_parameter (pp, t);
> >        break;
> >
> > assumes that TYPE_IDENTIFIER is indeed an identifier and that would not
> > be necessarily the case in C++ when TYPE_NAME_IDENTIFIER boils down to
> > TYPE_NAME itself (in practice, fails like eg, abi/mangle12.C, where
> > TYPE_NAME is a TEMPLATE_DECL).
>
> You could avoid that issue if TYPE_NAME_IDENTIFIER checked for 
> IDENTIFIER_NODE rather than TYPE_DECL.
I see, but then, if I understand correctly what you are saying, I'm 
afraid we could end up with a somewhat loose condition, which would make 
me a little nervous:

#define TYPE_IDENTIFIER(NODE) \
 (TYPE_NAME (NODE) && TREE_CODE (TYPE_NAME (NODE)) != IDENTIFIER_NODE \
 ? DECL_NAME (TYPE_NAME (NODE)) : TYPE_NAME (NODE))

are we sure this is Ok vs the other front-ends, errors, etc? I'm 
wondering if alternately we could extend the condition to encompass the 
C++-specific cases while keeping it precise. For example the attached 
appears to pass the testsuite...

Or, it occurs to me:

#define TYPE_IDENTIFIER(NODE) \
 (TYPE_NAME (NODE) && DECL_P (TYPE_NAME (NODE)) \
 ? DECL_NAME (TYPE_NAME (NODE)) : TYPE_NAME (NODE))

which also appears to work and has the advantage that at least it picks 
the right tree class without using TEMPLATE_DECL explicitly (which, 
admittedly is a bit weird outside /cp). Sort of a compromise.

Paolo.

/////////////////////

Attachment: patch_type_identifier
Description: Text document