[HTML] Add new C++ access check rules to changes.html (PR 14949)

[HTML] Add new C++ access check rules to changes.html (PR 14949)

[Giovanni Bajo]

Hello,

this patch implements PR 14949, which is basically adding a brief documentation
about the change introduced by the fix for PR 11174 (made by Kriang for GCC
3.4.0). Since a fair number of persons submitted an invalid bug about this, I
think it is worth documenting it.

Tested with the validator, OK to commit?

Giovanni Bajo


Index: changes.html
===================================================================
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-3.4/changes.html,v
retrieving revision 1.116
diff -c -3 -p -r1.116 changes.html
*** changes.html 20 Apr 2004 18:39:01 -0000 1.116
--- changes.html 10 Jun 2004 12:03:03 -0000
***************
*** 574,582 ****
     foo(a1);        // OK, a1 is a lvalue
   }</pre>

!  This might be surprising at first sight, especially since most
   popular compilers do not correctly implement this rule
!  (<a href="../bugs.html#cxx_rvalbind">further details</a>).</li>
    </ul>

    <h4>Runtime Library (libstdc++)</h4>
--- 578,617 ----
     foo(a1);        // OK, a1 is a lvalue
   }</pre>

!  <p>This might be surprising at first sight, especially since most
   popular compilers do not correctly implement this rule
!  (<a href="../bugs.html#cxx_rvalbind">further details</a>).</p></li>
!
!  <li>When forming a pointer to member or a pointer to member function,
!  access checks for class visibility (public, protected, private)
!  are now performed using the qualifying scope of the name itself.
!  This is better explained with an example:
!
!  <pre>
!  class A
!  {
!  public:
!    void pub_func(void);
!  protected:
!    void prot_func(void);
!  private:
!    void priv_func(void);
!  };
!
!  class B : public A
!  {
!  public:
!    void foo(void)
!    {
!      &amp;A::pub_func;   // OK, pub_func is accessible through A
!      &amp;A::prot_func;  // error, cannot access prot_func through A
!      &amp;A::priv_func;  // error, cannot access prot_func through A
!
!      &amp;B::pub_func;   // OK, pub_func is accessible through B
!      &amp;B::prot_func;  // OK, can access prot_func through B (within B)
!      &amp;B::priv_func;  // error, cannot access prot_func through B
!    }
!  };</pre></li>
    </ul>

    <h4>Runtime Library (libstdc++)</h4>

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Jason Merrill]

OK.

Jason

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Gerald Pfeifer]

On Thu, 10 Jun 2004, Giovanni Bajo wrote:
> Tested with the validator, OK to commit?

Jason already approved, but I'm wondering about one detail:

> Index: changes.html
> ===================================================================
> !  public:
> !    void pub_func(void);
> !  protected:
> !    void prot_func(void);
> !  private:
> !    void priv_func(void);
> !  };
> !
> !  class B : public A
> !  {
> !  public:
> !    void foo(void)

Do we really need all these "(void)"s instead of just "()"?  Personally, I
prefer the latter.

Gerald

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Jason Merrill]

On Thu, 10 Jun 2004 16:50:22 +0200 (CEST), Gerald Pfeifer <gp@suse.de> wrote:

> Do we really need all these "(void)"s instead of just "()"?

Not in C++, no.

Jason

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Giovanni Bajo]

Gerald Pfeifer wrote:

> Do we really need all these "(void)"s instead of just "()"?
> Personally, I prefer the latter.

Not really, it's just a matter of style. I am committing the patch without the
voids.

Giovanni Bajo

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Vaclav Haisman]

On Thu, 10 Jun 2004, Giovanni Bajo wrote:

> Hello,
>
> this patch implements PR 14949, which is basically adding a brief documentation
> about the change introduced by the fix for PR 11174 (made by Kriang for GCC
> 3.4.0). Since a fair number of persons submitted an invalid bug about this, I
> think it is worth documenting it.
>
> Tested with the validator, OK to commit?
>
> Giovanni Bajo
>
>
> Index: changes.html
> ===================================================================
> RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-3.4/changes.html,v
> retrieving revision 1.116
> diff -c -3 -p -r1.116 changes.html
> *** changes.html 20 Apr 2004 18:39:01 -0000 1.116
> --- changes.html 10 Jun 2004 12:03:03 -0000
> ***************
> *** 574,582 ****
>      foo(a1);        // OK, a1 is a lvalue
>    }</pre>
>
> !  This might be surprising at first sight, especially since most
>    popular compilers do not correctly implement this rule
> !  (<a href="../bugs.html#cxx_rvalbind">further details</a>).</li>
>     </ul>
>
>     <h4>Runtime Library (libstdc++)</h4>
> --- 578,617 ----
>      foo(a1);        // OK, a1 is a lvalue
>    }</pre>
>
> !  <p>This might be surprising at first sight, especially since most
>    popular compilers do not correctly implement this rule
> !  (<a href="../bugs.html#cxx_rvalbind">further details</a>).</p></li>
> !
> !  <li>When forming a pointer to member or a pointer to member function,
> !  access checks for class visibility (public, protected, private)
> !  are now performed using the qualifying scope of the name itself.
> !  This is better explained with an example:
> !
> !  <pre>
> !  class A
> !  {
> !  public:
> !    void pub_func(void);
> !  protected:
> !    void prot_func(void);
> !  private:
> !    void priv_func(void);
> !  };
> !
> !  class B : public A
> !  {
> !  public:
> !    void foo(void)
> !    {
> !      &amp;A::pub_func;   // OK, pub_func is accessible through A
> !      &amp;A::prot_func;  // error, cannot access prot_func through A
> !      &amp;A::priv_func;  // error, cannot access prot_func through A
I am no expert but isn't there a typo ---------------^^^^ here?

> !
> !      &amp;B::pub_func;   // OK, pub_func is accessible through B
> !      &amp;B::prot_func;  // OK, can access prot_func through B (within B)
> !      &amp;B::priv_func;  // error, cannot access prot_func through B
> !    }
> !  };</pre></li>
>     </ul>
>
>     <h4>Runtime Library (libstdc++)</h4>


VH

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Paolo Bonzini]

>>!      &A::pub_func;   // OK, pub_func is accessible through A
>>!      &A::prot_func;  // error, cannot access prot_func through A
>>!      &A::priv_func;  // error, cannot access prot_func through A
> I am no expert but isn't there a typo -------------^^^^ here?

Yes, there is...

>>!
>>!      &B::pub_func;   // OK, pub_func is accessible through B
>>!      &B::prot_func;  // OK, can access prot_func through B (within B)
>>!      &B::priv_func;  // error, cannot access prot_func through B
                                  ...and here as well ^^^^

Ciao,

Paolo

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Paolo Bonzini]

>>!      &A::pub_func;   // OK, pub_func is accessible through A
>>!      &A::prot_func;  // error, cannot access prot_func through A
>>!      &A::priv_func;  // error, cannot access prot_func through A
> I am no expert but isn't there a typo -------------^^^^ here?

Yes, there is...

>>!
>>!      &B::pub_func;   // OK, pub_func is accessible through B
>>!      &B::prot_func;  // OK, can access prot_func through B (within B)
>>!      &B::priv_func;  // error, cannot access prot_func through B
                                  ...and here as well ^^^^

Ciao,

Paolo

Re: [HTML] Add new C++ access check rules to changes.html (PR 14949)

[Giovanni Bajo]

Vaclav Haisman wrote:

>> A !      &A::priv_func;  // error, cannot access prot_func
>> through A 
> I am no expert but isn't there a typo ---------------^^^^ here?

Good catch! Fixed, thanks.

Giovanni Bajo