public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "giovannibajo at libero dot it" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/13847] Additions to the changes.html file for 3.4.
Date: Sun, 25 Jan 2004 16:50:00 -0000	[thread overview]
Message-ID: <20040125165049.32298.qmail@sources.redhat.com> (raw)
In-Reply-To: <20040124154329.13847.mattyt-bugzilla@tpg.com.au>


------- Additional Comments From giovannibajo at libero dot it  2004-01-25 16:50 -------
Befriending and explicitly instantiating typedefs won't work anymore.
***********************************************************************

Yes, this could use a changes.html entry. In standardese, GCC 3.4.0 enforces 
the standard clause in [basic.lookup.elab]/2 and /3, which talks about 
elaborated-type-specifiers (ETS): "If the name lookup finds a typedef-name, the 
elaborated-type-specifier is ill-formed."

An ETS is something of the form "class/struct/enum X;". ETS are mainly used for:

- forward declarations (not affected by recent changes).

- friend declarations. Given the bugfix above, GCC 3.4.0 doesn't allow anymore 
to befried a class through a typedef name:

class A;
typedef A B;
class C { 
  friend class B;   // ERROR! It used to work.
};

- Explicit instantiations. Again, you can't use typedef names anymore:

template <int> class A {};
typedef A<0> B;
template class B;   // ERROR! It used to work.

- As an additional note, ETS are also useful as a workaround if you have a both 
a type and a non-type entity bound to the same name.

struct A {};
int A;

void foo(void)
{
  A = 0;   // OK
  A a;     // ERROR, lookup finds "int A"
  struct A a;  // OK, lookup of ETS always finds "struct A".
}



Need a "template <>" before templated static variable definitions.
******************************************************************
This is already documented in changes.html.


The available function overloadings from a call in a template is determined at
definition time not instantiation time, except for Koenig lookup.
******************************************************************************
Ok. I would just say that, in templates, all (modulo bugs) unqualified names 
are now looked up in at definition time. This can be added as a rework of the 
third C++ entry in changes.html. An example code could be:

void foo(int);

template <int> struct A {
   static void bar(void){
       foo('a');
   }
};

void foo(char);

int main()
{
  A<0>::bar();    // Calls foo(int), used to call foo(char).
}


You need to put "class" on explicit instantiations of classes, ie no "template
A<B>;"
******************************************************************************
I didn't even know this was allowed before, must have been a GCC extension. OK, 
we can add a entry for this as well.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13847


  parent reply	other threads:[~2004-01-25 16:50 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-01-24 15:43 [Bug c++/13847] New: " mattyt-bugzilla at tpg dot com dot au
2004-01-24 15:47 ` [Bug c++/13847] " pinskia at gcc dot gnu dot org
2004-01-25 16:50 ` giovannibajo at libero dot it [this message]
2004-01-25 16:52 ` giovannibajo at libero dot it
2004-01-26 14:34 ` mattyt-bugzilla at tpg dot com dot au
2004-02-05 14:53 ` giovannibajo at libero dot it
2004-02-05 20:04 ` reichelt at gcc dot gnu dot org
2004-02-08 20:46 ` giovannibajo at libero dot it
2004-02-09 14:06 ` giovannibajo at libero dot it

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20040125165049.32298.qmail@sources.redhat.com \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).