Updated status pages

[Jim Dehnert]

> From nathan at codesourcery.com Mon Aug  7 05:10:19 2000
> 
> Jim Dehnert wrote:
> > Nathan, have you had a chance to look at the RTTI structure
> > definitions? And Alex was going to go over the mangling examples.
> finally had some time to do that. A few more nits -- pretty tiny in
> the big scheme, but we may as well try and get consistency!
> 
> 1) 2.9.5/6.c The definition of __vmi_class_type_info contains
> the base array definition as,
>         __base_type_info const base_info[1];
> That should be __base_class_type_info. Why is it const qualified?
> The other data members are not, and this is not a pointer to a const
> object, which would justify a qualification. I beleive that qualifier
> should be removed.

Got these.

> 2) 2.9.5/6.b has the base pointer called `base_type' 2.9.5/6.c has
> __base_class_info's base pointer called `base'. These should have the
> same name.

And these.

> 3) 2.9.5/6.c has a `flags' member of __vmi_class_type_info but
> an `offset_flags' member of __base_class_info. Again, these should have
> the same name, (and so should their associated enumerations).

I don't think so.  "flags" contains only flags, whereas "offset_flags"
contains a 56-bit offset as well.

> 4) 2.9.5/6.c. The enumeration should include the bit shift to obtain the
> offset. I recommend,
>         offset_shift = 8

Got it.

> 5) 2.9.5/7. The __pbase_type_info has a `qualifier_flags'
> member, which now contains more than just type qualifiers. In line
> with __vmi_class_type_info, I suggest renaming to simply `flags'.

OK.

> 6) 2.9.5/7 has a `pointee' member, but the text describing the qualifier
> flag meanings uses the word `target'. These should be consistent (I have
> no preference).

Got them.

> 7) 2.9.5/9 The `context_class' member should be a const qualified pointer
>         const __class_type_info *context_class;
> Maybe shortening the name to simply `context' would be less verbose too.

OK.  We also decided at the meeting that we needed to move all the
member and enum constant names to the reserved namespace (for macro
avoidance).  Take another look...

Jim

-	    Jim Dehnert		dehnert at sgi.com
				(650)933-4272