Other results from today's meeting (28 Sept 2000)

[Jim Dehnert]

Other discussions that happened today:

- RTTI field names:  Since those who cared most weren't in attendance,
  we didn't get far with this issue.  I will add the Standard-required
  function members to the std::type_info definition in the ABI, so it
  will be clear that they are there, and this is a virtual class, and
  I will also say so more clearly.

  Should the names have leading underscores?  There appears to be a
  strong desire for that, though I admit to agreeing that it should not
  really be necessary and would not in practice be a problem.  So I
  won't change this unless I see signs that those favoring them have
  been browbeaten into submission.

  Should the RTTI struct contents be normative?  I have heard a request
  for that, and not really much against it beyond the original
  suggestion that they not be.  Who cares?  Which do you prefer?  Why?

- Mark, would you please verify the changes in 3.3.3?  (If you already
  did, just tell me my memory's going, and that will suffice.)

- I will fix __cxa_throw_type_info to be a struct containing both a
  std::type_info and a pointer to a destructor.

- Christophe will look at my suggested solution of the G-4 issue.

- Jason is removing to Red Hat in Cambridge, and won't be at meetings
  in the future.  Wish him bon voyage...

Jim

-		Jim Dehnert  x3-4272