C++ demangler horrors

[Gabriel Dos Reis]

"H. J. Lu" <hjl@lucon.org> writes:

| On Fri, Jun 27, 2003 at 09:41:13PM +0200, Gabriel Dos Reis wrote:
| > "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes:
| > 
| > [...]
| > 
| > |  > It uses <vector> and <string>.
| > |  > H.J.
| > | 
| > | Can't you write mini replacements?
| > 
| > I think it would not be a good idea to take that road. 
| > I would not recommand libstdc++-v3 to include mini implementations of
| > this or that functionality already part of our standard distribution
| > each time a source code managment issue surfaces. 
| > 
| 
| It will be used for libiberty, not libstdc++-v3.

I'm not a big fan of code duplication.  Having to maintain two 
implementations of the demangler is not something I would recommand.
I really appreciate the effort you invested in this issue.
Still, I think going further proving mini versions of standard containers
is overkill and a recipe for long-term confusion.

This is a classic boostrap problem, if the only tools available for
boostrapping are bare metal and genuine C compiler, then, one can
just add a bootstrap stage -- building cc1plus and libstdc++.  But, I
suspect that most of the time, in addition of the C compiler, we also
have a C++ compiler.

-- Gaby