Re: [PATCH 3/3] Add name canonicalization for C

[Tom Tromey <tromey@adacore.com>]

>>>>> "Andrew" == Andrew Burgess <aburgess@redhat.com> writes:

>> struct type *
>> lookup_signed_typename (const struct language_defn *language, const char *name)
>> {
>> -  struct type *t;
>> -  char *uns = (char *) alloca (strlen (name) + 8);
>> -
>> -  strcpy (uns, "signed ");
>> -  strcpy (uns + 7, name);
>> -  t = lookup_typename (language, uns, NULL, 1);
>> -  /* If we don't find "signed FOO" just try again with plain "FOO".  */
>> -  if (t != NULL)
>> -    return t;
>> +  /* In C and C++, "char" and "signed char" are distinct types.  */
>> +  if (streq (name, "char"))
>> +    name = "signed char";

Andrew> I wondered why this "char" -> "signed char" conversion is done
Andrew> unconditionally for all languages, when the comment hints that the
Andrew> conversion only applies for C/C++?  I guess I would have expected a
Andrew> language check here.

lookup_signed_typename is only used by C and C-like languages.  For
non-C-like languages, sticking a "signed" (or "unsigned", see
lookup_unsigned_typename) prefix on a type name doesn't really make
sense anyway.

The uses outside c-exp.y are, IMNSHO, just leftover code from the bad
old days.  Like, I suspect there's no reason to have binop_promote at
all, instead this could be an explicit node in the expression tree, and
rather than having a big 'switch' on the language, each language could
simply make a different node.

Tom