Re: C++ attributes ?

[Jeffrey A Law <law at cygnus dot com>]

  In message <19990831231011.B19288@tetto.liafa.jussieu.fr>you write:
  > I don't want to be rude to *Jason*, who's made some astounding 
contributions
  > to the compiler.  Please take the above comment tongue-in-cheek.
  > 
  > On the other hand, I don't buy the `volunteer project' part, especially as
  > I do contribute.   Sorry, but I can't write documentation for stuff
  > I don't understand.
Right.

  > > Also, it is not unreasonable to get a feature working first, then
  > > after it works, to document it.  I suspect that com_interface is even
  > > newer than init_priority.
  > 
  > More often than not, when documentation is not written  at the same time
  > as the code, it never is... 
Right.  Very very true.  Mike, you should know that! :-)  You've been involved
in probably the best/worst case of this in gcc's history -- the EH stuff.

  > We know of some parts of the compiler which took forever to document, some
  > which are not documented, or understood by about four people on Earth.
  > I'd very much would prefer newer stuff *not* to be in the same boat.
Agreed.

Now, can someone please explain to Marc what that COM stuff does at a high
level since Marc has volunteered to write up some documentation.

jeff